Silver resistance in MRSA isolated from wound and nasal sources in humans and animals

Abstract

Methicillin‐resistant Staphylococcus aureus (MRSA) colonises skin, nasal passages and dermal wounds. Methods used to manage wounds infected and colonised with MRSA often include the use of topical antiseptics such as ionic silver and iodine. The objectives of this study were to determine the prevalence of silver‐resistance (sil) genes in MRSA and methicillin‐resistant coagulase‐negative staphylococci (MR‐CNS) isolated from wounds and nasal cavities of humans and animals, and also to determine the susceptibility of sil‐positive and sil‐negative MRSA isolates to a silver‐containing Hydrofiber^® (SCH) wound dressing, on planktonic silE‐positive and silE‐negative MRSA. Polymerase chain reaction was used to determine the presence of three silver‐resistance (sil) genes, silE, silP and silS in 33 MRSA and 8 methicillin‐resistant staphylococci (MR‐CNS). SilP and silS genes were absent in all isolates tested; however, two MRSA strains were found to contain the silE gene, together with one isolate of MR‐CNS. Phenotypic resistance of the silE‐positive strains and their susceptibility to the SCH dressing was evaluated using the zone of inhibition test on Mueller Hinton agar, and confocal laser microscopy using a live/dead fluorescent stain. Results confirmed that the SCH dressing was effective in killing all MRSA strains with and without the silE gene. First, this study showed that the prevalence of sil genes was low in the isolates investigated; and secondly, that the presence of a silver‐resistance gene (silE) in MRSA and MR‐CNS did not afford protection to the organism in the presence of a SCH wound dressing. The use of topical antiseptics in chronic wound care should be considered before the use of antibiotics that can result in their overuse and the risk of further resistance.

Introduction

Methicillin‐resistant Staphylococcus aureus (MRSA) is responsible for a wide range of infections ranging from localised skin conditions to life‐threatening conditions such as pneumonia (1). The prevalence of MRSA‐related infections is widespread, with levels as high as 54.5% being reported (2). Most strains of MRSA are resistant to antibiotics including β‐lactam antibiotics, chloramphenicol, clindamycin, tetracycline and fluoroquinolones (3). Presently, vancomycin is widely used to treat infections involving MRSA; however, because of reports of resistance to this antibiotic (4), alternative antimicrobial therapies are being investigated.

MRSA may also colonise infected chronic wounds where it presents an infection risk both in the wound and in the wider environment. Wound dressings containing antiseptic agents are being used to contain MRSA and generally control wound bioburden. One such antiseptic agent that is commonly used in wound management is ionic silver 5, 6. However, since the first report on the genetic and molecular basis of silver resistance in salmonella published by Gupta et al. (7), there has become an increasing awareness of the existence of silver‐resistance genes (silE, silS and silP) in bacteria 8, 9. To date, no MRSA strains have been found to possess silver‐resistance genes.

The aim of this study was to determine the prevalence of silver‐resistance genes in MRSA and also methicillin‐resistant coagulase‐negative staphylococci (MR‐CNS) isolated from wounds and nasal passages, and subsequently examine the efficacy of a silver‐containing wound dressing on subsequent sil‐positive and sil‐negative strains using a zone of inhibition (ZOI) test and rapid scanning confocal laser microscopy.

Methods

Organisms

Thirty‐three MRSA and eight MR‐CNS were used in this study. The identification of isolates and source site from which they were isolated are detailed in Table 1. All isolates were cultured and identified using both standard methods and polymerase chain reaction (PCR). PCR was used to confirm MRSA by identifying the presence of the femA and mecA genes, and MR‐CNS by amplification of the mecA gene as stated previously (7).

Table 1.

Source of all MRS isolates assessed for the presence of silS, silP and silE silver‐resistance genes

Isolate ID	Isolate species	Animal	Site
132	MRSA	Dog	Wound infection
133	MRSA	Dog	Wound infection
138	MRSA	Wallaby	Eye infection
139	MRSA	Vulture	Wound infection
137	MRSA	Human	Nasal carriage
136	MRSA	Horse	Wound infection
140	MRSA	Dog	Nasal carriage
141	MRSA	NA	Environmental
142	MRSA	Dog	Wound infection
143	MRSA	NA	Environmental
144	MRSA	Human	Nasal carriage
146	MRSA	NA	Environmental
147	MRSA	NA	Environmental
148	MRSA	Dog	Skin carriage
149	MRSA	Human	Nasal carriage
150	MRSA	Human	Nasal carriage
151	MRSA	Dog	Nasal carriage
152	MR‐CNS	Horse	Nasal carriage
153	MRSA	Dog	Wound infection
154	MRSA	Dog	Wound infection
155	MRSA	Dog	Wound infection
156	MRSA	Dog	Wound infection
157	Staphylococcus sciuri	Dog	Wound infection
160	MRSA	Dog	Joint infection
161	MRSA	Dog	Wound infection
162	MR‐Staphylococcus epidermidis	Horse	Wound infection
163	MR‐CNS	Horse	Joint infection
164	MR‐CNS	Chimp	Nasal carriage
165	MR‐CNS	Chimp	Nasal carriage
166	MRSA	Dog	Joint infection
167	MR‐CNS	Dog	Wound infection
168	MR‐CNS	Dog	Joint infection
169	MRSA	Dog	Wound infection
170	MRSA	Dog	Wound infection
171	MRSA	Dog	Skin infection
172	MRSA	Dog	Skin infection
134	MRSA	Dog	Abscess
135	MRSA	Dog	Wound infection
216	MRSA	Dog	Unknown
217	MRSA	Dog	Unknown
219	MRSA	Dog	Skin infection

MRSA, methicillin‐resistant Staphylococcus aureus; MR‐CNS, methicillin‐resistant coagulase‐negative staphylococci; NA, not applicable.

Antibiotic resistance was determined by disc diffusion assay according to the British Society of Antimicrobial Chemotherapy guidelines (10).

The MRSA and MR‐CNS isolates were sourced from a variety of animals, including dogs, horses and humans (working in veterinary medicine) along with environmental isolates from a veterinary hospital. The isolates were obtained from Leahurst Veterinary College during routine screening of animals and humans as outlined in the study undertaken by Baptiste et al. (9). The controls used in this study included Escherichia coli strain J53 containing plasmid pMG101 (positive control), E. coli strain J53 without plasmid pMG101 (negative control) and plasmid pKM1 (positive control – contains silE, silS, part silR and an ampicillin‐resistance marker) 7, 11 and a silE‐positive Enterobacter cloacae (positive control) that was kindly donated by Alan Lansdown, Imperial College Faculty of Medicine, London, UK (12). All the recombinant positive controls were maintained on ampicillin‐impregnated Luria Bertani (Sigma Aldrich, St Louis, MO, USA) media.

Dressings

The antimicrobial dressings evaluated in this study were a SCH (ConvaTec, Flintshire, UK) dressing and a non silver‐containing Hydrofiber^® (NSCH; ConvaTec) dressing.

Polymerase chain reaction

To prepare cell lysates, a single bacterial colony of each MRSA and MR‐CNS isolate was inoculated into 10 μl of Martindale water and heated at 95°C for 10 minutes, vortexed and then pulse‐centrifuged (6000 g) very briefly. After centrifugation, the supernatant fluid was used as the PCR template. All oligonucleotide primers used were obtained from MWG‐BIOTECH (Ebersberg, Germany). The primer composition for the silE, silS and silP genes can be found elsewhere (13). Before use, 15 μl of each primer (24–34 pmol/μl) was diluted to 100 μl using nuclease‐free water. The PCR reaction mixture was composed of 41 μl ReddyMix™ PCR Master Mix (Abgene, Epsom, UK), 3 μl of forward primer solution, 3 μl reverse primer solution and 3 μl of template preparation. The solution was then pulse centrifuged at room temperature to mix the contents. A drop of mineral oil was added onto the top of the mixture to prevent evaporation. About 3 μl of the template was pipetted through the oil layer into the mixture. The PCR reaction mixture preparation was repeated for silP, silE and silS for each isolate. PCR was run at an initial 95°C for 2 minutes, then 40 cycles of 95°C for 1 minute, 55°C for 1 minute and 72°C for 3 minutes, followed finally by 72°C for 5 minutes. The resulting product was then stored at 4°C. Products were subjected to electrophoresis in a 2% Tris–acetate–ethylenediaminetetraacetic acid (TAE) agarose gel, with a 100–1000 bp nucleic acid ladder to provide size standards, at 70 V for 60 minutes and stained with ethidium bromide (0.5 μg/ml). PCR products were then visualised with a UV transilluminator and recorded with a Gene Genius Bio‐imaging System version 6.03 (Syngene, Cambridge, UK). All experiments were carried out in duplicate.

Sequencing

All excised PCR products were purified using the Qiagen Q1A quick PCR purification kit (Qiagen, West Sussex, UK). Sequencing was carried out by the Advanced Biotechnology Centre, Imperial College London, UK using ABI 3100 16 capillary genetic analysers. The NCBI BlastN program was used to analyse all sequenced results.

Susceptibility of MRSA to a silver‐containing wound dressing

Thirteen MR‐CNS, which included three silE‐positive and two reference MRSA strains (EMRSA‐10 and NCTC12232), were investigated in this study. Blood agar (BA; Sigma Aldrich) plates were prepared as per manufacturers instructions. An overnight culture of each isolate was prepared at a concentration of approximately 1 × 10⁵ colony forming units/ml and inoculated onto the surface of BA in triplicate and then allowed to dry for several minutes. Samples of dressings (SCH and NSCH) were cut aseptically to a size of 2 cm². With the use of sterile forceps, each dressing type was placed onto each inoculated BA and pressed down gently to ensure close contact. All dressings were hydrated with 0.5 ml saline to more closely mimic hydration in wound conditions. Plates were incubated at 35 ± 3°C for 24 hours. After incubation, all plates were observed and the zones of inhibition around each sample on both the plates were measured. As the SCH and NSCH dressings were observed to contract after the addition of saline, inhibitory zones were calculated by subtracting the dimensions of the dressing after hydration from these zones and obtaining a mean.

Confocal laser microscopy

Three silE‐positive MRSA and MR‐CNS strains together with additional sil‐negative controls strains were investigated using confocal laser microscopy (14). Isolates were suspended in Maximum Recovery Diluent (Oxoid, Basingstoke, UK) to a concentration corresponding to a McFarland Standard of 3.0. Molecular Probes Live/Dead BacLight™ dye (Invitrogen, Paisley, UK) was prepared following manufacturers instructions. This solution was stored in the dark at room temperature until required. About 500 μl of BacLight stain was gently mixed with 500 μl of each MRSA or MR‐CNS isolate and incubated in the dark at room temperature for at least 10 minutes. About 1‐cm² samples of SCH dressing, gauze and the NSCH dressing were cut aseptically and placed on coverslips suspended over a 4‐cm² hole cut in the base of a Petri dish. Using sterile forceps, the dressings were gently teased apart to expose fibres to aid visualisation. To each of the dressings (n = 2), 150 μl of BacLight stained bacteria was added. The Petri dish was then covered and kept in the dark ready for confocal imaging. Samples were imaged using a Leica TCS SP2 inverted rapid‐scanning confocal laser microscope set to image wavelengths between 500 and 530 nm (Syto9 stains viable cells) and 610–640 nm [propidium iodide (PI) – stains dead cells) on two separate PMT channels. Images of each dressing were obtained at 3, 24 and 48 hours after the dressings were inoculated.

Statistical analysis

The ZOI for each isolate was determined in triplicate. The mean and standard deviation of ZOI for each dressing and bacteria were determined via Microsoft Excel 2000 (version 5.1.2600). Significance in difference of ZOI between different dressings and different bacterial isolates was examined using a t‐test (Minitab^® Release 14).

Results

This study identified only one silver‐resistance gene (silE) in 2 of 33 (6%) of the MRSA isolates studied (Figure 1). These were cultured from a wound and nasal passage of two dogs. The same single gene (silE) was also isolated from one MR‐CNS isolate (Staphylococcus sciuri) from the nares of a horse (Figure 1). Interestingly, the S. sciuri isolate was positive by PCR for both the mecA and femA genes, genes usually associated with S. aureus. Strain identification was confirmed using APIStaph (BioMerieux, Basingstoke, UK). The corresponding size of the silE PCR product was approximately 400 bp and was in accordance to that published by Silver (15). The results of sequencing showed that strain 133/03 (accession number AG067954) has 100% homology with the silE gene from pMG101 (7), strain 151 (accession number AG067954) has 95% homology with the silE gene from pMG101 (7) and strain 152 (accession number AG067954) has 100% homology with the silE gene. The sequence results showed that all three staphylococcus isolates, positive for the presence of the silE gene by PCR, had ≥95% homology for the pMG101 silver‐resistance gene (7).

Figure 1.

Analysis of polymerase chain reaction products from silver‐resistance (sil)E gene‐specific primers, and methicillin‐resistant Staphylococcus aureus and methicillin‐resistant coagulase‐negative staphylococci isolates of 152 (lane 2), 151 (lane 3) and 133/03 (lane 4) from wounds via agarose gel electrophoresis. Lane 1, Ready‐Load™ 1 Kb Plus DNA Ladder (Invitrogen); lane 5, water control; lane 6, Escherichia coli J53, negative control; lane 7, EC702 (CN, sil‐positive Enterobacter cloacae); 8, pKM1, positive control. The bands of 400 bp in size corresponds to silE. Analysis was carried out by staining the gel with ethidium bromide and using Gene Genius Bioimaging System version 6.03 (Syngene, Synoptic Ltd, Cambridge, UK).

An additional investigation was performed to determine whether the growth of silE‐positive and silE‐negative MRSA and MR‐CNS isolates were inhibited by the SCH wound dressing. The SCH dressing was shown to inhibit the growth of all strains evaluated (including silE‐positive strains) when compared with the control NSCH dressings. A ZOI range of 0.95–1.48 cm was measured against all sil‐positive and sil‐negative MRSA and MR‐CNS evaluated. The ZOI for sil‐positive strains 133/03 was 0.95 ± 0.05 cm, 1.22 ± 0.06 cm for strain 151 and 0.98 ± 0.08 cm for strain 152. There was no significant difference between the ZOI of sil‐positive and sil‐negative strains. All results were shown to be statistically significant (P < 0.05) when compared with the NSCH controls that showed no inhibition zones.

Confocal imaging was used to observe the dynamic effect of the silver‐containing dressing against MRSA and MR‐CNS bacteria stained with a live/dead stain allowed cell death over time (Figure 2A–H). Results showed that sil‐negative MRSA control strains were killed after a 3‐hour incubation period in the presence of ionic silver within the SCH dressing (Figure 2B). In the presence of the NSCH dressing, both the control sil‐negative strains (Figure 2A) and the silE positive strains (Figure 2C, E, G) were shown to survive for the full 48‐hour time course and therefore maintained the green Syto9 fluorescent stain. In the presence of ionic silver, all MRSA and MR‐CNS strains previously shown to contain the silE gene (MRSA 133/03, 151 and 152) were killed within 48 hours after inoculation into the dressing. Images are shown for strain 133/03 and reflect the results seen for all three positive isolates (Figure 2D–H). Cell death started to occur in some cells after 24 hours, as shown by the appearance of PI‐stained red bacteria (Figure 2F). Total bacterial cell death was apparent when the dressings were imaged after 48 hours and showed that all cells had stained red.

Figure 2.

Confocal laser microscope images showing the efficacy of ionic silver present in SCH on methicillin‐resistant Staphylococcus aureus (MRSA) viability. Bacteria labelled with a live:dead BacLight stain were shown as green when alive and red when dead. Incubation of a non silver‐resistant strain of MRSA with SCH showed cell death occurred within 3 hours of inoculation (B), compared with the NSCH control, where cells remained viable (A). MRSA 133/03, found to contain silE, showed no significant cell death after 3 hours with SCH (D). After 24 hours, low levels of cell death were apparent (F) and after a maximum of 48 hours, all bacteria were dead (H). Incubation with a NSCH (control) showed that cells survived within the dressing when silver was absent (C, E and G).

Discussion

Before this study, no silver‐resistant MRSA isolates from wounds or other sites in humans or animals have been reported. Recent reports have suggested that pet animals could act as a reservoir for MRSA infections and re‐infection in humans (16). Therefore, there is a possible risk of silver‐resistant MRSA being transferred from animal to humans or vice versa (11).

The presence of silE alone in MRSA isolates in this study was inconsistent with the previous finding that the three silver‐resistance genes (silE, silS and silP) are typically associated with each other (15). Investigation into the presence of silver and other heavy metal‐resistance genes, in the genome or on plasmids, may offer further insight as to why only this single gene is present. The silE gene encodes a periplasmic protein capable of binding metal ions. The gene product has also been shown to be homologous to PcoE gene found in E. coli and known to be responsible for copper resistance (15). The physiological and phenotypic effects of these missing genes are still unclear. However, based on the results of PCR and bacterial sensitivity to a silver wound dressing in our case, it is believed that simultaneous presence of silP, silE and silS might not be the factor necessary to confer silver resistance to the bacteria. This also correlates with the data collected by Gupta et al. (7).

ZOI testing and confocal microscopy showed excellent activity of the SCH wound dressing against silE‐positive MRSA. It is probable that the presence of all sil genes might be essential for bacteria to display phenotypic resistance to silver. Confocal imaging of cell death within the SCH dressing showed that, although silE‐positive MRSAs were eventually killed by ionic silver, there was a lag period when compared with the control MRSA during which the bacteria were able to survive. This may reflect a partial resistance of the bacteria to the effects of silver that is eventually overcome as the sustained provision of free silver ions from the dressing led to accumulation of ions within the bacterial cells. Further investigation of this delayed effect by varying the concentrations of silver to which the bacteria are exposed may confirm this. This study did not determine the exact locations of silE in the MRSA 133/03, 151 and 152 that warrants further investigation to determine if silver resistance is plasmid determined or chromosomal. This could then provide information on whether the sil genes are linked to antibiotic‐resistance genes and further establish the switch and mechanism necessary for sil genes to be turned on and if transfer of these plasmids is possible in vivo.

In summary, it was shown that the prevalence of silver‐resistance genes in MRSA and MR‐CNS from various sources was low and restricted to a single gene (silE). Furthermore, investigation of the susceptibility of silE‐positive isolates to a SCH showed that the dressing remained effective. This study was undertaken to study the prevalence of sil genes in MRSA. The study shows therefore that the silE gene, responsible for the generation of the silE protein, did not provide significant resistance to the isolates as indicated by the ZOI and confocal results.