|Home | About | Journals | Submit | Contact Us | Français|
To (i) identify chronic wound bacteria and to test their ability to produce acyl-homoserine-lactones (AHLs) and autoinducer-2 (AI-2) cell-cell signaling molecules and (ii) determine if chronic wound debridement samples might contain these molecules.
Partial 16S rRNA gene sequencing revealed the identity of 46 chronic wound strains as belonging to nine genera. Using bio-reporter assays, 69.6% of the chronic wound strains were inferred to produce AI-2 while 19.6% were inferred to produced AHL molecules. At-least one strain from every genus, except those belonging to the genera Acinetobacter and Pseudomonas, were indicated to produce AI-2. Production of AI-2 in batch-cultures was growth-phase-dependent. Cross-feeding assays demonstrated that AHLs were produced by Acinetobacter spp., Pseudomonas aeruginosa and Serratia marcescens. Independent from studies of the bacterial species isolated from wounds, AHL and/or AI-2 signaling molecules were detected in 21 of 30 debridement samples of unknown microbial composition.
Chronic wound bacteria produce cell-cell signaling molecules. Resident species generally produce AI-2 molecules and aggressive transient species associated with chronic wounds typically produce AHLs. Both these classes of cell-cell signals are present in human chronic wounds.
Inter-bacterial cell-cell signaling may be an important factor influencing wound development and the presence of AHLs and AI-2 could be used as a predictor of wound severity. Manipulation of cell −cell signaling may provide a novel strategy for improving wound healing.
Chronic wounds currently affect 2.5 million to 4.5 million people in the US (Jones et al., 2007). This is a major health-problem as chronic wounds are unresponsive to typical treatment strategies and persist even after the implementation of normal wound debridement, wound irrigation, and antimicrobial-based treatment strategies (Dow et al., 1999; Harding et al., 2002). Three common types of chronic wounds are described by their etiology and include those related to neuropathic disorders (e.g. diabetic ulcers and diabetic foot ulcers or DUs), vascular disorders (e.g. venous leg ulcers or VLUs) and environmentally induced disorders (e.g. pressure ulcers or PUs, non-healing surgical wounds or NHSWs and general chronic non-healing wounds) (Mustoe et al., 2006; Wolcott et al., 2008). While their etiology is distinct, a commonality is the presence of high concentrations of bacteria within the wounds (Wolcott et al., 2008).
Recently, James et al. (James et al., 2008) demonstrated that many species of bacteria exist within chronic wounds as part of complex multi-species biofilms. This is a significant finding as bacteria within biofilms exhibit properties that are distinct from planktonic cells (Donlan & Costerton, 2002; Fux et al., 2005; Spormann, 2008). In particular, biofilm bacteria are more resistant to antimicrobials than their planktonic counterparts (Gilbert et al., 2002) and are also resilient to attack from the human immune system (Donlan & Costerton, 2002). In these resilient chronic wound biofilms, two distinct bacterial populations exist. One population can be described as the resident bacterial population, which multiplies and persists on the skin of healthy individuals, and the other is the transient bacterial population, which colonizes and multiplies in unhealthy human skin and is typically found in high concentrations in chronic wound biofilms (Price, 1938; Ruocco et al., 2007). Biofilm community analysis of chronic wounds have shown not only the presence of typically resident species of the normal cutaneous microbial flora, such as Streptococcus spp., Proteus mirabilis, Bacillus spp. and Staphylocccus spp. (Dekio et al., 2005; Dowd et al., 2008a; Gjodsbol et al., 2006; James et al., 2008), but also the presence of high cell numbers transient species which are typically Gram-negative rod-shaped bacteria (Lowbury, 1969) such as Pseudomonas aeruginosa (Gjodsbol et al., 2006) and Acinetobacter spp. (Bowler & Davies, 1999; Hill et al., 2003). The transient species are considered to be highly aggressive chronic wound biofilm species as they are able divide rapidly to numerically dominate within chronic wound multi-species biofilms. Indeed, a recent study by Gjodsbol and coworkers (Gjodsbol et al., 2006) demonstrated that chronic wounds that harbored P. aeruginosa within the biofilms were significantly larger than those without P. aeruginosa. An ability to control the development of multi-species biofilms and specifically inhibit pathogenic species within a biofilm would be beneficial in the treatment of chronic wounds.
The production and detection of bacterial cell-cell signaling molecules by species have been repeatedly linked to the enhanced development of single and multi-species biofilms (Irie & Parsek, 2008). A variety of structurally different bacterial cell-cell signaling molecules have been shown to mediate cell-cell communication and include acylated homoserine lactones (AHLs) and autoinducer-2 molecules. AHLs are produced solely by Gram negative bacteria (Williams, 2007) although different species often produced one or more different forms of AHLs. These different AHLs all consist of a homoserine lactone ring moiety but differ with respect to the length, degree of saturation and specific substitutions within an attached acyl side-chain. Because of the heterogeneity of AHL structure(s), AHLs have been proposed to mediate intra-species bacterial communication; different species typically only recognize AHLs produced from closely related species (Miller & Bassler, 2001). However, there are examples where bacterial species can detect AHLs from bacteria belonging to distantly related genera (Bernier et al., 2008; Stickler et al., 1998). Conversely, another class of cell-cell signaling molecule, called autoinducer-2 (AI-2), has been shown to mediate inter-species signaling (Rickard et al., 2006; Surette et al., 1999; Yoshida et al., 2005). AI-2 is an umbrella term used to describe a family of inter-convertible molecules that is derived from the same precursor molecule; 4,5-dihydroxy-2,3-pentanedione (DPD) (Semmelhack et al., 2005). DPD is produced or detected by many Gram positive and Gram negative bacteria (Sun et al., 2004) and has been shown to contribute to single- and dual-species biofilm development (Gonzalez Barrios et al., 2006; Hardie & Heurlier, 2008; Rickard et al., 2006).
Recently, work by Nakagami and coworkers (Nakagami et al., 2008) has shown that AHLs can be detected in pressure-induced infected ischemic wounds on rats. Of particular interest, was the finding that wounds that were infected with P. aeruginosa contained AHLs at concentrations up to 0.49 pmol/g. The amount of AHLs increased linearly as the cell density of P. aeruginosa in the wound also increased. No AHLs were detected in wounds that were not infected with P. aeruginosa. The amount of other cell signaling molecules, such as AI-2, was not determined.
In order to increase understanding of the role of bacterial cell-cell signaling in the development of human chronic wounds, it was the aim of this study to determine if bacteria that are often isolated from typical chronic wound biofilms produce cell-cell signaling molecules. It was also the aim of this work to determine if these signal molecules could be detected in etiologically distinct wounds. We discovered that many different species of chronic wound bacteria produce cell-cell signaling molecules and most resident species produce AI-2 while Gram-negative pathogens produce only AHLs. A study of chronic wound debridement samples of unknown microbial composition, but from different wound types (DUs, VLUs, NHWs and NHSWs) also indicates the presence of possible AHLs or AI-2 molecules in chronic wounds.
Forty-six strains were isolated from chronic wound debridement specimens from 32 patients undergoing standard sharp debridements as part of the normal course of their wound care management. These specimens were collected with written informed consent under a protocol approved by the Western Institutional Review Board (Olympia, Washington, USA). Strains were randomly chosen from culturable bacterial panels generated from debridement samples taken from patients suffering with diabetic ulcers (DUs), pressure ulcers (PUs), non-healing surgical wounds (NHSWs), non-healing wounds (NHWs) or venous leg ulcers (VLUs). The source of each strain is shown in Table 1. Strains were cultured in Schaedler broth (BD, Franklin Lakes, NJ, USA) supplemented with 1 mM boric acid. Liquid cultures were grown aerobically at 37°C with shaking at 200 rpm.
Vibrio harveyi strains BB170, BB152, BB886 and BB120 (Surette & Bassler, 1998) were used to detect AI-2 (4,5-dihydroxy-2,3-pentanedione) and the AHL N-(3-hydroxybutanoyl) homoserine lactone (HBHL). Vibrio harveyi BB170 detects AI-2 while Vibrio harveyi BB152 is a producer of AI-2 and was used as a positive control. Vibrio harveyi BB886 detects HBHL (also referred to as autoinducer-1) while Vibrio harveyi BB120 is a producer of HBHL and was used as a positive control. All V. harveyi strains were cultured in AB medium (Greenberg et al., 1979) at 30°C. Liquid cultures were grown aerobically at 30°C with shaking at 200 rpm.
Agrobacterium tumefaciens strains were used to detect other forms of AHL which include N-3-(oxo-octanoyl) homoserine lactone (referred to as OOHL) as well as a wide-range of structurally similar AHLs (Stickler et al., 1998). A. tumefaciens KYC6 (Stickler et al., 1998), a producer of OOHL and other AHL molecules, was grown on LB agar at 30°C. A. tumefaciens A136 (Stickler et al., 1998), which detects OOHL and other AHLs and responds by producing β-galactosidase, was grown in LB supplemented with spectinomycin (50 μg/ml) and tetracycline (4.5 g/m) at 30°C. All strains were stored at −70°C in 50% glycerol.
Separate to the collection of samples for the isolation of bacterial strains for this study, thirty human debridement samples were collected and frozen at −70°C until extraction of cell-free supernatant was required. Prior to extraction, the weight of each thawed sample was recorded and 500 μL of deionized water was added to the debridement sample. Samples were shaken vigorously using a Qiagen TissueLyser (Qiagen, Valencia, CA) at 30 Hz for 2 × 30 s. Subsequently, samples were centrifuged at 13,000 × g for 120 s. The supernatant was collected and filtered through a 0.45 μm Millex® HA filter unit (Millipore, Bedford, MA). Cell-free debridement samples were frozen immediately at −70°C for future use.
To determine the production of a broad range of AHLs from chronic wound strains, the AHL detector strain A. tumefaciens A136 was used in a cross-feeding assay developed by Stickler et al. (Stickler et al., 1998). LB agar was prepared by covering with 100 μl of 20 mg/ml X-Gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside). A. tumefaciens A136 was streaked onto the prepared LB agar and chronic wound strains were streaked in parallel approximately 1cm away. Plates were then incubated at 30°C for 48 h. If OOHL, or one of a variety of other AHLs with different acyl side-chain lengths and substitutions (6–12 carbon atoms in length), were produced by a test strain then A. tumefaciens A136 reported the presence of AHL molecules by activating a traI-lacZ fusion. Activation of the reporter fusion results in the cleavage of X-Gal which turns A. tumefaciens A136 colonies blue. Positive and negative controls consisted of culturing the reporter strain with the AHL over-producer strain A. tumefaciens KYC6 and with itself (an AHL non-producer).
To determine the presence of a broad range of AHLs in chronic wounds, the AHL detector strain A. tumefaciens A136 was also used. To detect AHL cell-cell signaling molecules, 6.5 mm diameter discs of Whatman filter paper (Number One, Whatman, Hillsboro, OR) were impregnated with 5 μl of cell-free supernatant from chronic wounds. These discs were placed onto the surface of LB agar plates supplemented with 50 μg/ml spectinomycin and 4.5 μg/ml tetracycline and pre-coated with 100 μl of X-Gal (20-mg/ml stock solution in dimethyl formamide) and 100 μl of a cell-suspension (OD 1.0 at 610 nm) of A. tumefaciens A136. Plates were incubated for 72 h at 30°C and 5 μl cell-free chronic wound supernatant were added to each respective disc every 12 h. The presence of AHLs in wound samples was inferred if the A. tumefaciens A136 colonies in contact with the disc or surrounding the disc became blue after 72 h incubation.
HBHL (an AHL that is also referred to as autoinducer-1) and AI-2 activity was determined by testing batch-culture cell-free supernatants using a modification of a bioluminescence assay by Frias et al. (Frias et al., 2001). Cells of V. harveyi BB170 (AI-2 reporter strain) or BB886 (HBHL reporter strain) were grown at 30°C in batch cultures of autoinducer bioassay (AB) medium according to the method of Surette and Bassler (Surette and Bassler, 1998). After 14 h of growth, V. harveyi BB170 or V. harveyi BB886 cultures were diluted 1:500 in fresh AB medium (Rickard et al., 2006). The diluted V. harveyi BB170 or V. harveyi BB886 cultures were stored at −70°C until required.
To analyze the amount of HBHL or AI-2 produced by chronic wound strains, each strain was grown individually in batch-cultures containing Schaedler medium supplemented with 1 mM boric acid. Following 3 h, 6.5 h, 8.5h and 24h batch-culture growth, 1 ml of cultures were collected and passed through a 0.2 μm filter unit (Fisher Scientific, Suwanee, GA). From these cell-free supernatants, 10 μl was added to 90 μl of thawed diluted V. harveyi BB170 culture (to detect AI-2) or to 90 μl of thawed diluted V. harveyi BB886 culture (to detect V. harveyi HBHL) in a 96-well microplate and analyzed in a Victor 3 multi-label counter (Perkin Elmer, Waltham, MA, USA). Bioluminescence values that were relative to un-inoculated medium were calculated as fold inductions (Blehert et al., 2003). For batch cultures studies a positive result was recorded for fold inductions of ≥30 over the media control. A 30-fold induction represents a fold-induction that is 1 % of the V. harveyi HBHL positive control (V. harveyi BB120) and 5% of the V. harveyi AI-2 positive control (V. harveyi BB152) (Table 2). The higher 5% cut-off point for determining if strains produce AI-2 was chosen because the addition to 1 mM boric acid to growth media not only strengthens the signal from AI-2 producing strains but also enhances background signal (DeKeersmaecker & Vanderleyden, 2003). A comparison to E. coli DH5α, that is known to be unable to produce AI-2 in boric acid supplemented Schaedler medium, shows that 5% was an appropriate cut-off (Table 2).
To determine the presence of HBHL or AI-2 in chronic wounds, V. harveyi BB886 (HBHL detector) and V. harveyi BB170 (AI-2 detector) were used. To detect the presence of HBHL or AI-2 molecules, 10 μl of wound sample was added to 90 μl of either thawed diluted V. harveyi BB170 culture or V. harveyi BB886 culture in a 96-well microtitre plate and analyzed in a Victor 3 Victor 3 multi-label counter (Perkin Elmer, Waltham, MA, USA). Bioluminescence relative to PBS (pH 7.4) was calculated as fold inductions (Blehert et al., 2003). Because wound samples were not supplemented with boric acid, fold inductions >10 (1% of typical bioluminescence from cell-free supernatants of V. harveyi BB152) were considered positive for AI-2 signal production. Samples that generated a fold-induction of greater than >10 bioluminesence from V. harveyi BB886 were also considered positive for HBHL.
Following a modified method of Rickard et al. (Rickard et al., 2002), strains were identified by polymerase chain reaction (PCR) amplification and partial sequencing of the 16S rRNA gene fragment. Amplification of 16S rRNA was performed by taking one 72 h-old colony of each strain grown on Schaedler agar. The colonies were boiled separately in 100 μl of double-autoclaved nanopure water for 7 min and 10 μl of each boiled suspension were used as template DNA for PCR amplifications. Degenerate primers 8FPL (Wilson et al., 1990) and 806R (Weisburg et al., 1991) were used to amplify fragments of 16S rDNA which correspond to nucleotides 8–806 in the Escherichia coli K12 16S rRNA gene sequence. PCR reactions were carried out in PCR Red-Taq reaction buffer (Sigma, St. Louis, MO). PCR consisted of 35 cycles at 94°C (1 min), 53.5°C (1 min) and 72°C (1 min), plus a final cycle with a 15 min chain elongation step at 72°C. Amplified products were purified (QIA-quick™ PCR purification kit, Qiagen, Warrington, UK) and purified PCR products were sequenced using the primers 806R and 8FPL. Sequencing reactions consisted of 20–80 ng of PCR product, 10 ng of primer and 2 μl of Big Dye™ (PE Applied Biosystems, Foster City, CA, USA) in a total volume of 20 μl. Samples were incubated at 94°C (4 min) followed by 25 cycles of 96°C (30 s), 50°C (15 s) and 60°C (4 min). Sequencing was performed in a Perkin-Elmer ABI 310 sequencer (Perkin-Elmer, Foster City, CA). Unambiguous sequences of 650–750 bases in length were obtained from each strain and compared to known sequences in the EMBL database using BLAST (Altschul et al., 1997). Based on the criteria described by Stackebrandt and Goebel (Stackebrandt & Goebel, 1994), 16S rRNA gene sequences that were 97–100 % identical to speciated strains in the EMBL database were assigned the genus and species name. Sequences that possessed a sequence identity of < 97 % to speciated strains in the EMBL database were only assigned the genus name.
The closest relative species or genera were assigned based upon compiled partial 16S rRNA gene sequences. Representative sequences of closely related strains were also aligned using CLUSTALX ver. 1.83 (Thompson et al., 1997). Neighbor-joining analysis was conducted with the correction of Jukes and Cantor (Jukes & Cantor, 1969) using TREECON ver. 1.3b (Van de Peer & De Wachter, 1993) with Thermus thermophilus (X07998) as the outgroup.
All strains were identified to the genus level and, based upon the criteria of Stackebrandt and Goebel (Stackebrandt & Goebel, 1994), all but two strains (Klebsiella sp. CWS47 and Staphylococcus sp. CWS31) were speciated (Table 1). The strains belonged to 9 genera and 33 of 46 strains shared ≥99 % identity to known 16S rRNA gene sequences from previously identified strains that are commonly isolated from chronic wounds. These include Serratia marcescens, Escherichia coli, Enterococcus faecalis, Pseudomonas aeruginosa and Staphylococcus aureus. The unspeciated strains, Klebsiella sp. CWS47 and Staphylococcus sp. CWS31, were not given species epithets as they shared an identity of ≤97 % to the nearest speciated 16S rRNA gene sequence in the EMBL database.
Partial 16S rRNA gene sequence alignment, nearest-neighbor pair-wise comparison and construction of a phylogram (Fig. 1) confirmed that the assigned identity of each chronic wound strain was appropriate. Additionally, the tree revealed that some strains of the same genera shared identical partial 16S rRNA gene sequences and some were less similar. For example, the partial 16S rRNA from E. faecalis CWS36 was identical to E. faecalis CWS17, while E. faecalis CWS15 was <98% identical to both E. faecalis CWS36 and E. faecalis CWS17. None of the strains with identical 16S rRNA gene sequences were isolated from the same individual (data not presented). Similar patterns of sequence identities were observed for clusters of strains belonging to Proteus mirabilis, S. marcescens, P. aeruginosa and S. aureus. Different patients with different wounds could harbor closely related or less-related strains of the same genera.
Two approaches were used to determine if the chronic wound bacteria produce AHLs. Both approaches relied upon the growth of the organisms in laboratory media. The first approach used a bioluminescence assay developed using V. harveyi BB886 to detect N-(3-hydroxybutanoyl) homoserine lactone (HBHL) (Surette & Bassler, 1998). The second approach used a cross-feeding assay, developed by Stickler et al. (Stickler et al., 1998) and relies upon the reporter strain A. tumefaciens A136 to detect the presence of a broad range of different forms of AHL molecules (homoserine lactone acyl side-chain lengths of C6–C12).
According to the V. harveyi BB886 bioluminescence assay, none of the chronic wound bacteria induced bioluminescence that was consistently >30.0 fold over the negative control and only generated signals that were ≤0.4 % of that from the positive control, V. harveyi BB120 (Table 2). Thus, all the chronic wound strains were deemed not to produce HBHL signal molecules. Cross-feeding assays, using A. tumefaciens A136, demonstrated that 19.6 % of the strains (9 of 46) produced AHL molecules (other than HBHL) (Table 2). Only 50 % (9 of 18) of the strains belonging to S. marcescens, P. aeruginosa and Acinetobacter spp. were inferred to produce AHL signal molecules (Fig. 1a and b). The reason why AHL cell-cell signal molecules were not detected for all strains belonging to these genera may be due to the absence of detectable quantities of AHL being produced by the seemingly non-AHL-producing strains. A negative response does not necessarily mean that these AHL-negative strains do not produce AHL, and instead may mean the amount produced is below the threshold for detection in this reporter system. Acinetobacter baumannii CWS18 elicited the strongest reaction from the A. tumefaciens A136 reporter strain. Such a reaction, inferred by the induction of β-galactosidase production and development of a blue colony type, may indicate that this strain produced a collection of different detectable AHL signal molecules or one type of AHL signal molecule at high concentration.
With the exception of strains belonging to the genera Pseudomonas and Acinetobacter, at least one strain belonging to each of the other 7 genera were inferred to produce AI-2 (Fig. 1, Table 2). The three highest AI-2 fold-induction values were obtained from batch cultures of P. mirabilis CWS11 (382.5 fold induction), E. faecalis CWS36 (295.6 fold induction) and S. marcescens CWS5 (271.6 fold induction).
The amount of AI-2 detected, from AI-2 producing strains, varied during batch-culture growth (four sample points in different growth-phases in batch culture) and changes in amounts of AI-2 was different for each species. For example, differences in AI-2 concentration (expressed as fold-induction) occurred over different time points and growth phases in batch-cultures of E. faecalis CWS8 and in batch-cultures of S. aureus CWS41 (Fig. 2). S. aureus CWS41 produced AI-2 in exponential phase and, upon entry into stationary phase, either halted production of AI-2 or sequestered (through cellular-uptake) AI-2 at the same rate at which it was being produced. Conversely, E. faecalis CWS8 produced AI-2 in exponential phase and early stationary phase but the amount of AI-2 that was present in later-stationary phase cultures was reduced by 74%, indicating that AI-2 was being internalized or degraded extracellulary.
Thirty chronic wound debridement samples were analyzed to determine if AHL and/or AI-2 molecules could be detected. Samples were taken from four etiologically different types of chronic wounds (Table 3). The V. harveyi BB886 bioluminescence assay, which detects HBHL, showed that none of the samples likely contained detectable quantities of HBHL as bioluminescence was <9.42 fold-induction over bioluminescence induced by phosphate buffered saline (Table 3). Use of A. tumefaciens A136 as a bio-reporter for other AHL molecules indicated that 15 of 30 of the samples contained detectable concentrations of AHL molecules other than HBHL (Table 3). The presence of these AHL molecules was inferred by a change in colour (from off-white to blue) of A. tumefaciens A136 colonies in-contact and around cell-free wound debridement sample impregnated discs. In all positive cases, colonies of A. tumefaciens A136 growing on or in close proximity to the discs impregnated with chronic wound debridement sample turned a shade of blue and colonies that were further away from the discs were a often lighter blue or remained white (Fig. 3). Of interest too, was that for some wound samples, a zone of inhibition was evident within 1–5 mm around the disc and was probably due to the presence of human-cell-produced inhibitory components or antibiotics from treatment regimes. Of the entire wound debridement samples tested, those from DUs most often contained AHLs (61.5%, Table 3). Debridement samples from non-healing wounds (NHSWs and NHWs) and VLUs least-often contained AHLs (57.1% and 30% respectively, Table 3).
Analysis of the cell-free extracts of debridement samples, using the V. harveyi BB170 bioluminescence assay, indicated that 8 of 30 of the samples contained AI-2. Of the entire wound debridement samples tested, those from DUs most often contained AI-2 (38.5%, Table 3). Debridement samples from VLUs and non-healing wounds (NHSWs and NHWs) least-often contained AI-2 (20% and 14.3% respectively, Table 3). Only two of the 30 debridement samples contained AI-2 and AHLs and these were from two debridement samples from DUs (Table 2).
This work demonstrates that clinical strains of bacteria, that were isolated from different types of chronic wounds, are able to produce cell-cell signaling molecules. Not only can the strains produce cell-cell signaling molecules, when grown under laboratory conditions, but these molecules are likely produced by the bacteria within chronic wounds. Thus, it is possible that these cell-cell signaling molecules mediate inter-bacterial communication in chronic wound biofilms.
The strains were isolated and randomly chosen for this study from 32 patients suffering from NHSWs, NHWs, DUs, PUs or VLUs. Identification by partial 16S rRNA gene sequencing showed that the strains were from genera that are typically isolated from human chronic wounds (Dowd et al., 2008a; James et al., 2008). Phylogenetic analysis revealed that clusters of very closely related strains belonging to the same species were isolated and identified from different patients (Table 1 and Fig. 1). Such a finding indicates that specific species are associated with chronic wounds, regardless of their etiology, and this has been shown by recent ecological studies by other researchers (Davies et al., 2004; Dowd et al., 2008a; Dowd et al., 2008b; Hill et al., 2003). A. iwoffii, S. marcescens and E. faecalis strains were isolated from chronic wounds and strains belonging to these species have been previously isolated from healthy and diseased skin (Berlau et al., 1999; Gao et al., 2007; Roth & James, 1988). S. aureus are also commonly isolated from both healthy and diseased skin (Gjodsbol et al., 2006), although S. aureus is often found at higher cell densities in wounds (Kirketerp-Moller et al., 2008). Strains belonging to A. baumannii and P. aeruginosa are commonly isolated from aggressive, rapidly expanding, chronic wounds and evidence suggests that they numerically dominate within most aggressive chronic wounds (Davies et al., 2004; Gjodsbol et al., 2006). Indeed, A. baumannii and P. aeruginosa are considered important wound pathogens and general nosocomial Gram-negative pathogens as they are extremely resistant to antimicrobials (Gootz & Marra, 2008; Hernandez, 2006; Howell-Jones et al., 2005; Navon-Venezia et al., 2005) and have a repertoire of mechanisms, such as the expression of surface-attached fimbriae and/or capsular polysaccharide (Navon-Venezia et al., 2005; Peleg et al., 2008; Pollack, 1984), to maintain their presence in chronic wounds. Consequently, these cell populations expand to the detriment of typical resident species within the population (Davies et al., 2004).
A comparison of the phylogenetic relatedness of the identified chronic wound strains to their ability to produce cell-cell signaling molecules indicated that members of all of the nine genera, identified in this study, produced either AHL or AI-2 cell-cell signaling molecules. Eight of the 46 strains were not indicated to produce cell-cell signaling molecules. However, for each of these strains, closely related strains belonging to the same genera were shown to produce at-last one class or type of signaling molecule (Fig. 1, Table 2). It is possible that observed differences were due to strain-specific timing of cell-cell signaling molecule production or strain-specific differences in the amounts of signaling molecule produced (which is below the level of detection for the observed negative strains). Every genera inferred to produce AHLs and/or autoinducer-2 cell-cell signaling molecules in this research have previously been inferred to be able to produce cell-signal molecules by other research groups. However, the majority of the strains tested by other researchers are strains from culture-collections or isolates from environments other than from chronic wounds. Clearly, clinical chronic wound strains share an ability to produce specific cell-cell signal molecules with other strains from the same genera. Also, when considering which stains produced which AHLs or AI-2, S. marcescens, albeit 3 of 7 of the strains belonging to this genera, produced detectable amounts of both classes of cell-cell signaling molecules. The ability to produce both classes of signal molecules has been shown before for this species, and AHLs and AI-2 are known to coordinate biofilm maturation and the production of various extra-cellular enzymes (Van Houdt et al., 2007). The only genera not able to produce AI-2 molecules were those associated with highly aggressive chronic wounds. Specifically, P. aeruginosa and members of the genera Acinetobacter did not produce AI-2 but did produce AHL cell-cell signaling molecules. AHL molecules are known to enhance the biofilm-forming ability of P. aeruginosa (Juhas et al., 2005) and Acinetobacter spp. (Niu et al., 2008). As such, AHLs may exclusively mediate inter-bacterial communication system between pathogenic chronic-wound species and facilitate the coordination of activities by groups of pathogenic chronic wound species within healthy (typically resident) multi-species biofilm communities. By coordinating their activities, pathogens will have an improved ability to expand within a biofilm at the expense of the resident community and consequently a chronic wound develops. When considering that A. tumefaciens A136 was used as a reporter system in this work, it should be remembered that it detects a broad range of AHL signal molecules. Specifically, it is most sensitive to OOHL, N-(3-oxo-octanoyl) homoserine lactone, and N-(3-oxodecanoyl) homoserine lactone. It can also detect longer chains but is insensitive to N-butanoyl-L-homoserine lactone, which is one of many AHL molecules of different chain lengths that is also produced by P. aeruginosa (Williams, 2007). A. tumefaciens A136 is generally considered adequate to study the presence of AHLs in biological samples, because bacteria often produce a broad range of AHLs and many produce long chain AHLs (Erickson et al., 2002; Williams, 2007). However the inclusion of a short chain AHL detector (C4-HSL to C6-HSL) such as Chromobacterium violaceum CV026 (McClean et al., 1997) could show that samples that were reported negative for AHLs by A. tumefaciens A136 actually contain short-chain AHLs.
AHLs and AI-2 were inferred, through bioluminescence and cross-feeding bioassays, to be present in chronic wounds (Table 2). Either of the two classes of molecules were indicated to be present within 21 of 30 (70%) of the chronic wound debridement samples. The method to detect these molecules relied upon the direct sampling of cell-free chronic wound debridement samples and this contains a complex mixture of human and microbial molecules (James et al., 2000; Trengove et al., 1996). While not performed in the work reported here, another confirmatory approach to demonstrate that the debridement samples contain bacterial signaling molecules could utilize a solvent extraction method and/or chromatographic techniques to separate and purify bacterial cell-cell signaling molecules from chronic wound debridement samples (Brelles-Marino & Bedmar, 2001; Chambers et al., 2005; Middleton et al., 2002; Nakagami et al., 2008). While the use of such protocols have not been successfully used for the purification of AI-2, solvent extraction and chromatographic approaches have demonstrated that AHLs are present in the sputum of cystic fibrosis patients (Middleton et al., 2002), in the ruminal contents of cows (Erickson et al., 2002), in human food (meat and fish) extracts (Medina-Martinez et al., 2007) as well as from pressure-induced ischemic wounds on rats that were infected with P. aeruginosa (Nakagami et al., 2008). In this latter study, Nakagami and coworkers were the first to quantify AHLs from wound samples. Interestingly, they did not detect AHLs in rats with low ( < 2.0×106 CFU/g) or high ( > 1.1×108 CFU/g) bacterial counts. As the authors pointed-out, the absence of AHLs at high bacterial densities could be attributed to the degradation (quenching) of AHLs by N-acyl-homoserine lactone acylases from P. aeruginosa (Sio et al., 2006), the enhanced production paraoxonase-like enymes by human cells that hydrolyze the lactone ring of AHLs (Yang et al., 2005) and opening of AHL ring in high pH conditions and changes of temperature (Yates et al., 2002). Coincidently, some species of bacteria are also known to internalize AI-2 under specific environmental conditions (Shao et al., 2007; Taga, 2007; Xavier & Bassler, 2005). A similar phenomenon may have occurred during batch culture of E. faecalis CWS8, where AI-2 concentration apparently dropped in later stationary phase (Fig. 2). Thus, a variety of factors may contribute to changes in cell-cell signal concentration in chronic wounds.
Even though the number of wound debridement samples studied in this work is fairly small (especially when considering the different types of wounds that were sampled), AHLs and AI-2 were only detected twice in the same sample (2 of 30 samples). Such a finding is intriguing and may indicate that either class of molecule could be present at high concentration when specific bacterial species are present in chronic wound biofilms. For example a shift in chronic wound community composition from an AI-2 producing resident community to that which contains predominantly AHL-producing pathogenic P. aeruginosa could be accompanied with a change in the dominant type of signal molecule in the wound. With less resident species producing AI-2 and numerically dominant P. aeruginosa cells producing AHLs, the only detectable class of molecule would be AHLs. This hypothesis must be tempered with the knowledge that, as described earlier, AHL and AI-2 can be degraded by bacterial and human cells and under certain environmental conditions. Indeed, while not studied here, an analysis of bacterial composition and species abundance in chronic wounds and comparison to the presence or absence of specific signal molecules could prove to be particularly enlightening and aid in determining a role for signal molecules in wounds. However, at present, available evidence does suggest that cell-cell signaling contributes to the expansion of P. aeruginosa in burn wounds (Bielecki et al., 2008; Rumbaugh et al., 1999), although parallel studies of changes in AHL and AI-2 concentrations has yet to be studied in chronic wounds. From a therapeutic standpoint, if cell-cell signaling could be interrupted, for example by QS interference (Morohoshi et al., 2007), then the chronicity of wounds could be reduced.
In conclusion, this study indicates that different clinically relevant chronic wound species produce different classes of cell-cell signal molecules and, for AI-2, is in a growth-phase-dependent manner. Evidence presented here also indicates that these molecules are produced in different types of chronic wounds. Changes in types of cell-cell signal molecule or signal concentration may occur, in part, as a consequence of shifts from an AI-2 producing resident community to that which contain AHL-producing Gram-negative pathogens that hinder wound healing.
This research was supported, in part, by grant number 5 P20 GM078445 from the National Institute of General Medical Sciences (NIGMS). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the NIGMS. The authors would like to thank Mr. Robert Lobe (University of Duisburg-Essen, Germany) for assistance in screening chronic wound samples for cell-cell signal molecules.