To analyse the biocidal efficacy of thermal sprayed copper surfaces.
To analyse the biocidal efficacy of thermal sprayed copper surfaces.
Copper alloy sheet metals containing >60% copper have been shown to exhibit potent biocidal activity. Surface biocidal activity was assessed by epifluorescence microscopy. After 2-h exposure at 20°C in phosphate-buffered saline (PBS), contact killing of Gram-negative Escherichia coli and Gram-positive Staphylococcus epidermidis by brass sheet metal and phosphor bronze was 3–4-times higher than that by stainless steel. SEM observations revealed that the surface membranes of both bacterial strains were slightly more irregular when exposed to brass sheet metal than stainless steel. However, when exposed to phosphor bronze coating, E. coli were 3–4 times larger with irregular membrane morphology. In addition, the majority of the cells were associated with spherical carbon-copper-phosphate crystalline nanostructures characteristic of nanoflowers. The membranes of many of the S. epidermidis exhibited blebbing, and a small subset was also associated with nanoflowers.
Our data indicate that increasing the surface roughness of copper alloys had a pronounced impact on the membrane integrity of Gram-positive and, to a lesser degree, Gram-negative bacteria. In the presence of PBS, carbon-copper-phosphate-containing nanoflowers were formed, likely nucleated by components derived from killed bacteria. The intimate association of the bacteria with the nanoflowers and phosphor bronze coating likely contributed to their nonreversible adhesion.
Thermal spraying of copper alloys provides a strategy for the rapid coating of three-dimensional organic and inorganic surfaces with biocidal copper alloys. Our study demonstrates that the macroscale surface roughness generated by the thermal spray process enhances the biocidal activity of copper alloys compared with the nanoscale surface roughness of copper sheet metals. Moreover, the coating surface topography provides conditions for the rapid formation of organic copper phosphate nanocrystals/nanoflowers.
Bacterial contamination of surfaces in hospitals, food-processing facilities and restaurants is the underlying cause of many, often life-threatening, microbial infections. It is estimated by the USA's Center for Disease Control and Food and the Drug Administration that approximately one-tenth of the population becomes ill as a result of infections by enteric pathogens such as Salmonella enterica and Campylobacter jejuni. Another foodborne enteropathogen, Listeria moncytogenes, is fatal in approximately 30 per cent of high-risk individuals such as women and newborn children, individuals with weakened immune systems and seniors. Extended periods of hospitalization increase the probability of nosocomial infection with spore-forming antibiotic-resistant strains of Clostridium difficile, a major cause of life-threatening pseudomembranous colitis. The problem is exacerbated by the formation of heat-resistant spores that are refractory to alcohol-based and other disinfectants. Consequently, there has been a great deal of interest in coating surfaces with agents that afford long-term protection against environmentally and institutionally derived pathogens.
While organisms require low concentrations of metal cofactors for various metabolic and reproductive processes, high concentrations of ions, such as copper, are biocidal (Grass et al. 2011). Hence, the coating of surfaces with copper-based alloys could provide a nontoxic, cost-effective and eco-friendly way of countering bacterial contaminations. The US Environmental Protection Agency (EPA) has acknowledged the antimicrobial efficacy of over 280 copper-based products against disease-causing bacteria with an average biocidal efficacy of approximately 99% within 2 h for alloys containing 60% or higher concentrations of copper (Rai et al. 2012). The incorporation of copper-containing alloys into hospital wards could significantly decrease bacterial contamination compared with stainless steel or polymer surfaces. How copper mediates its potent contact killing of bacteria is context and species-dependent. It is well established that copper ions, via Farber and Fenton-mediated reactions, generate highly reactive free radicals (Grass et al. 2011). Ultrastructural and molecular biology experiments have demonstrated that the plasma membranes of bacteria are compromised in the presence of copper, leading to the release of intracellular components (Espirito Santo et al. 2008; Grass et al. 2011). In many cases, genomic and extrachromosomal DNAs are also degraded (Espirito Santo et al. 2008; Grass et al. 2011). Whether these activities are mediated by free radical end products with copper ions serving as electron donors/acceptors remains to be determined. The biocidal activity of copper may also be due to the toxic effect of high metal ion concentrations on the biological activity of proteins required for cell survival.
Although the biocidal efficacy of copper alloys has been well known, healthcare institutions and industries continue to manufacture furniture and equipment using polymers and stainless steel due to the prohibitive cost of using copper alloy sheet metal. At the same time, in many industries, thermal spray processes are used for coating applications to protect substrates from wear, heat or corrosion. The thermal spray process utilizes energy of an electric arc or combustion to melt and propel material towards a substrate. Upon impact, molten particles spread and solidify, forming a coating (Herman and Sulit 1993). A critical feature of the thermal spraying process is the relatively low heat load to the substrate, creating an opportunity to apply copper alloy coatings on heat-sensitive surfaces such as wood, engineered medium-density fibreboard (MDF) or polymer substrates. The technology provides a cost-effective and rapid method for effectively decreasing bacterial contamination on surfaces. In addition to their aesthetic appearance, copper-based alloys have enhanced mechanical and anticorrosion properties, increasing the longevity of the coated materials/substrates.
This study was focused on the evaluation of the biocidal activity of a copper alloy coating deposited onto an organic substrate (MDF). The efficacy of the coating was compared with certified US EPA brass (manufactured by PMX Industries, Cedar Rapids, IA, USA) and 304L stainless steel sheet metals using two species of bacteria, Gram-negative Escherischia coli (E. coli) and Gram-positive Staphylococcus epidermidis (S. epidermidis).
A twin-wire arc spraying technique was used for coating deposition. A ValuArc instrument (Sulzer Metco, Westbury, NY, US) was used, which was equipped with a high-velocity gun. Spray process parameters were chosen as recommended by the manufacturer, wired feed rate of 7 m min−1, an arc current of 280 amps and a voltage of 29 volts. Thermally sprayed coatings have a rough surface topography with a high surface area due to coating formation by solidified multiple splats. Peaks and valleys of the coatings are random, and the surface features are most often characterized by the parameter Ra, which is the arithmetic average of the absolute values of peaks and valleys along the sampling length. Phosphor bronze (P. Bronze) was selected as the coating material due its high copper content (91·7%) to ensure antimicrobial properties. This alloy also has better corrosion and mechanical properties than pure copper. The coating was deposited onto MDF, a common material for manufacturing furniture for diverse use. The coating surface was sanded to reduce Ra up to 3 times. The maximum profile valley depth (Rv) also was reduced from Rv = 47 μm for as deposited coating to Rv = 22 μm after sanding. Brass sheet metal with a regular striated pattern from machining has a lower surface roughness than thermal sprayed alloys. The molecular composition of the copper alloys was performed by energy dispersive spectroscopy (EDS, Quantax 70 from Bruker Nano GmbH, Berlin, Germany). Surface topography measurements were performed with a diamond stylus profilometer (Surfometer 400, Precision Devices, Milan, MI, USA). All 3D surface images were obtained by merging four ESM images taken at different angles using 3D Image Viewer (Denshi Kougaky Kenkyusyo Co., Hiscope System Company, Closter, NJ, USA).
Inoculations were prepared by suspending a bacterial colony in 10 ml of sterile LB broth that was kept on a rotary shaker for 24 h at 37°C. Bacteria were then regrown for 3 h in fresh sterile LB broth until log phase. The bacteria were added on to the substrates in order to allow for culture for 2 h. After 2 h, the samples were washed with 10 ml sterile phosphate-buffered saline (PBS) and plated on agar plates at 37°C overnight. The colonies were used to quantify bacterial cells that survived on the coatings.
Escherichia coli or S. Epidermidis were incubated for 2 h at room temperature; substrates were stained with LIVE/DEAD BacLight viability kit (Invitrogen, Burlington, Ontario, Canada). SYTO9, a green fluorescent nucleic acid stain, and propidium iodide (PI), a red fluorescent nucleic acid stain were used for the determination of viable bacteria. When SYTO9 was used independently, it labeled all the bacteria due to cell permeability properties. Propidium iodide is not cell permeable and hence is only able to stain cells where the membrane has been disrupted indicating nonviable cells. The costain was prepared by mixing 30 μl of SYTO9 and 30 μl of PI and diluting this solution to 1/200 in distilled water. 6 μl of the dye was poured on each substrate where the bacteria were inoculated. The staining was kept in the dark for 15 min. After that, the substrates were rinsed with distilled water. The fluorescent bacteria were visualized using fluorescence with Zeiss SteREO Discovery, V20.
Bacterial counts were performed by counting individual fluorescent spots within three random fields of view per sample at 120× magnification. Scanning electron microscopy (SEM) analysis revealed that a fluorescence spot 9·5 μm2 was representative of one bacterium, making it feasible to count individual cells. Large, irregular shape fluorescence stains were not counted. Dividing PI red fluorescence by SYTO9 green fluorescence staining of individual bacteria quantified lethality.
After inoculation for 2 h on the copper surface, bacterial cells were fixed using 4% formaldehyde in PBS buffer overnight at 4°C with rotation. Samples were then washed with PBS three times and then postfixed using 1% osmium tetroxide for 1 h at room temperature. The osmium tetroxide was then washed off with 0·1 mol l−1 PBS buffer three times for 5 min. The samples were then dehydrated in 50, 70, 80, 90 and 100% ethanol for 5, 10, 10, 15 and 2 × 10 min, respectively. Chemical critical point drying was achieved using hexamethyldisilazane series (HMDS) at 3 : 1, 1 : 1 and 1 : 3 parts ethanol to HMDS. Each treatment was kept for 30 min, and two changes of 100 HMDS were used for 15 min. The last change of HMDS was left to volatilize overnight in sterile Petri dish.
For SEM observations (Hitachi S2500, Hitachi Ltd, Tokyo, Japan), samples were then sputter-coated with gold–palladium.
The statistical program Graphpad® (Software Inc., La Jolla, CA, USA) Prism was used to calculate significant difference among results. The Kruskal–Wallis test was used with a Dunn modification testing for multiple sample comparisons.
A standard viable, plate count method was initially used to quantify the biocidal efficacy of all surfaces. Approximately 5000 Gram-negative E. coli and Gram-positive S. epidermidis bacteria in PBS buffer were plated onto 2-cm2 surfaces. Quantitative evaluation of the biocidal efficacy revealed that >80% of the E. coli and S. epidermis were killed by the exposure to brass sheet metal (EDS; 87% copper, 13% zinc), compared with <20% with stainless steel (data not shown). However, no live cells were observed on LB agar plates for the phosphor bronze (EDS; 91·7% copper, 7·5% tin) coatings. As it was highly improbable that the phosphor bronze coating, with a similar copper content as the brass sheet metal, would result in a 100% cell death only after a 2-h exposure, quantitative evaluation of biocidal activity was performed by the direct observation of bacteria on the surfaces by epifluorescence microscopy using SYTO9 and PI stains. Data shown in Fig. 1 indicate that a lethality ratio of 0·19 for E. coli and S. epidermidis was observed after a 2-h exposure to control stainless steel. By comparison, E. coli lethality ratios of 0·66, 0·75 and 0·81 were observed for brass sheet metal and unsanded and sanded coating surfaces, respectively. Lethality ratios of 0·68, 0·85 and 0·74 for S. epidermidis were observed on brass sheet metal and on unsanded and sanded coatings, indicating comparable biocidal efficacies by the different copper alloy surfaces for Gram-negative and Gram-positive bacteria. Statistically significant differences in lethality were only observed between stainless steel and the copper-containing alloys. Representative epifluorescence images of E. coli bacteria on the unsanded and sanded coatings are shown in Fig. 2, highlighting the fraction of cells with compromised membranes (red panels b and e) vs total (green, panels a and d) observed at 120× magnification. The yellow fluorescence seen in the merged images (panels c and f) indicates the majority of bacteria were killed. Similar images were obtained for S. epidermidis costained with SYTO9 and PI after exposure to stainless steel and brass sheet metal (data not shown).
Surface topography plays a significant role in the adherence of microbes to their substrates. To determine differences between the bacterial adhesions to the sheet metals compared with the coating, surface topography was analysed. Ra measurement revealed that surface roughness of stainless steel, brass sheet metal, sanded phosphor bronze and unsanded coating is 0·18, 0·54, 4·3 and 12·85 μm, respectively. Consistent with the large variation in Ra values, SEM revealed a relatively smooth, striated surface for brass sheet metal (Fig. 3a) compared with the highly variable topographical appearance of unsanded (Fig. 3b) and sanded (Fig. 3c) coatings. Three-dimensional analysis of the SEM images highlighted the different degrees of surface roughness between brass sheet metal (Fig. 3d) and the unsanded coating (Fig. 3e). Sanding of the coating reduced roughness by removing the peaks, leaving valleys intact (Fig. 3f).
To further investigate why the bacteria were not released from the phosphor bronze coating, SEM was used to observe the morphology of the cells after a 2-h incubation. The majority of E. coli on the control stainless steel was typically rod-shaped with smooth surfaces (Fig. 4a). Similarly, the surfaces of the spherical S. epidermidis appeared smooth (Fig. 4d), indicating that control stainless steel had no significant impact on the morphology of Gram-negative and Gram-positive bacteria. In contrast, the surface morphology of E. coli and S. epidermidis was slightly more irregular when exposed to the brass sheet metal (Fig. 4b,e). While there was no significant difference in biocidal activity between brass sheet metal and the unsanded or sanded phosphor bronze coatings (Fig. 1), there was a dramatic increase in the surface roughness and a 3- to 4-fold increase in the size of E. coli (Fig. 4c) exposed to the coatings with a minor subset lysed. The majority of E. coli appeared to be in intimate contact or enclosed by porous spheres after 2 h with an average size of 3·5–5 μm (Fig 4c), similar in size and appearance to hybrid organic–inorganic structures referred to as ‘nanoflowers’ that were recently reported by Ge et al. (2012). EDS analysis of a representative nanoflower revealed a composition of 16·3% carbon, 10·6% copper, 5·0% phosphorus and 2·5% oxygen (Fig. 5a,b). Sphere-free regions of the phosphor bronze coating were composed of 95·6% copper and 4·8% phosphorus (data not shown), indicating that the carbon atoms associated with the spheres were likely derived from components of killed E. coli. Because the porous spheres appeared similar in structure and composition to the protein Cu3(PO4).H20 nanoflowers reported by Ge et al. the spheres in our study will henceforth be referred to as nanoflower.
A time course analysis revealed that petal-like structures are associated with the E. coli as early as 30 min of exposure (Fig. 6a), growing in size until reaching maximum size after 2 h of incubation (Fig. 6b,c). Rod-like extensions extending from the swollen E. coli cells likely represent sites of crystal nucleation (Fig. 4c) as they do not have the long, threadlike appearance of fimbriae/pili. In contrast to E. coli, no significant difference in size was noted in S. epidermidis exposed to the phosphor bronze coating, although cells with extensive membrane blebs were often noted (Fig. 4f). A minor subset of the cells with membrane blebs was associated with nanoflowers.
In order to determine whether nanoflower nucleation was mediated by organic components derived from killed bacteria, crystal formation was analysed in the absence of bacteria. In the presence of PBS and the absence of bacteria, nonspherical, fibrous-like microcrystals were seen (Fig. 6d). In the presence of PBS and LB, without bacteria, nonporous formations with a mulberry-like surface topography were observed (Fig. 6e). When saline was substituted for PBS, swollen bacteria were observed. However, neither nanoflowers, crystals nor evidence of biofilm formation was observed (Fig. 6f). Likewise, nanoflowers and crystals were not observed following the chelation of copper ions with EDTA in the presence of PBS and bacteria (Fig. 6g). Thus nanoflowers formed only in the presence of PBS and bacteria and were composed of protein-copper-phosphate crystals.
Several studies have demonstrated that exposure of bacteria to copper alloys (>60% copper) for 2 h at 37°C results in the killing of approximately 90% of the bacteria (Grass et al. 2011). Consistent with the inverse relationship between biocidal activity and copper content, our data indicate that 80% of the Gram-negative E. coli and Gram-positive S. epidermidis were killed when exposed for 2 h at room temperature to brass sheet metal with 87% copper content. The biocidal efficacy was increased by 10–15% when cells were exposed to phosphor bronze coatings with slightly higher copper content (91·7% copper). Unexpectedly, in contrast to control stainless steel and brass sheet metals, neither viable E. coli nor S. epidermidis was released from sanded and unsanded coatings despite rigorous washing in the presence of glass beads, which could have been attributed to different surface roughness. Analysis by epifluorescence microscopy revealed that the biocidal activity of brass sheet metal and the phosphor bronze coating had comparable biocidal activities despite the differences in surface roughness. Hence, the differential cell adhesion between brass sheet metal and phosphor bronze coatings was likely due to a number of variables that included changes in surface topography.
Adhesion of bacteria to abiotic surfaces involves a stereotypic series of steps. The first step involves a gravity-mediated association with abiotic surfaces, a process that is accelerated by flagellar movement (Anselme et al. 2010). The second step, adhesion, is promoted by several factors, such as the membrane composition of the bacteria, the presence of fimbriae/pili, the formation biofilm by bacterial aggregates, as well as the surface topography of the substrate. The transition during this second step from ‘reversible’ to ‘nonreversible’ adhesion can be triggered by the formation of biofilm by bacteria that have made contact with a solid substrate (Anselme et al. 2010). Furthermore, analysis of biofilm production by aggregates of the genetically tractable E. coli over abiotic surfaces is partly promoted by flagellated strains (Pratt and Kolter 1998). However, E. coli DH5α and S. epidermidis, which have no flagella, also tightly adhered to phosphor bronze coating. Additionally, in contrast to the mainly amorphous appearance of extracellular polymeric biofilms observed under SEM that are formed by bacterial colonies (Flemming and Wingender 2010), petal-like structures were in intimate contact with the swollen E. coli and a subset of S. epidermidis. Increase in biofilm mass is dependent on bacterial proliferation and the continuous recruitment of free-floating bacteria. Hence, the presence of biocidal levels of copper is likely to be refractory to the growth of biofilms. Although it cannot be discounted that biofilm may have formed that was undetectable by SEM, the combined data indicate that biofilm-mediated adhesion is unlikely to have made a significant contribution to the irreversible adhesion of E. coli and S. epidermidis to the phosphor bronze coating.
Although poorly understood, there is a growing body of evidence that sessile bacteria sense and respond to the topography of their microenvironments, promoting or decreasing their surface adhesion depending on the size, morphology and physiochemical properties of the bacteria. However, with respect to nanostructure surfaces, contradictory results have been reported on the impact of surface roughness and the number of bound bacteria. As reviewed by Anselme et al. the contradictory results in bacterial adhesion are due to a combination of differences in the chemistry, wettability and nanotopography of surfaces. To circumvent issues associated with the impact of variances in substrate chemistry, the adhesion of different bacteria was investigated on glass slides with distinctive degrees of surface roughness, but with no measurable differences in surface chemistry (Mitik-Dineva et al. 2009). Their study demonstrated that E. coli attached readily to the smooth rather than rough glass surfaces. However, binding of the spherical S. aureus was not as affected by changes in surface roughness in the nanoscale range. We did not observe any significant difference in the number of E. coli and S. epidermidis bound to stainless steel with a Ra value of 180 nm. Approximately 50% more bacteria were associated with the brass sheet metal with a Ra value of 540 nm than with stainless steel. SEM images revealed that the surface of both bacterial species appeared rougher when exposed to brass sheet metal. The change in membrane morphology, combined with the rougher surface of brass sheet metal, may have resulted in a higher number of bacteria being retained on brass sheet metal compared with stainless steel.
A striking difference in bacterial morphology was observed between the solid metals and the phosphor bronze coating. This was particularly evident for E. coli cells that were approximately 3- to 4-fold larger with compromised membranes when plated on the sanded and unsanded phosphor bronze coating. The increased swelling in the presence of a hypotonic PBS solution may reflect that the cell walls of the bacteria were compromised by the copper ions. Swelling was observed after only 30 min of exposure to the biocidal surface, indicating that aberrant membrane permeability occurred rapidly, leading to osmotic stress due to the influx of water. Whether the cell walls were damaged by the generation of hydroxyl free radicals by Haber–Weiss and Fenton reactions of reduced copper ions remains to be determined. It is also likely that the E. coli genome was also rapidly degraded by the resultant free radicals as demonstrated for E. coli by Espirito Santo et al. (Espirito Santo et al. 2008). As noted by Warnes et al. (Warnes et al. 2010), PI does not effectively bind to degraded DNA. It is, therefore, conceivable that a subset of the E. coli on brass sheet metal and the phosphor bronze coating may not have been stained with PI, leading to an underestimate of biocidal efficacy. Moreover, intact bacteria with degraded DNA would have been nonviable, which may have affected the viable cell count for E. coli incubated on brass sheet metal.
No significant difference in the size of Gram-positive S. epidermidis was observed by the exposure to all substrates used in this study. Warnes et al. did not observe a change in the size and membrane morphology of Gram-positive Enterococcus faecalis and Enterococcus faecium when exposed to copper alloys with a copper content ranging from 60 to 95%. Bacterial killing was attributed to an inhibition of cellular respiration and DNA degradation by ROS (Warnes et al. 2010). In contrast to our studies with S. epidermidis where viable cells were detectable after 2 h of exposure to brass sheet metal, no viable E. faecalis and E. faecium cells were observed after a 1-h exposure to the copper alloys. As the authors hypothesized, it is conceivable that for Gram-positive cells, the absence of an outer cell wall and periplasmic space facilitates the intracellular penetration of toxic ROS, leading to cell death with minimum impact on cell membrane. Our data indicate that a subset of the S. epidermidis had compromised cell membranes when exposed to phosphor bronze coating, probably reflecting species-specific differences in the response of Gram-positive cells to toxic levels of copper, or that macroscale differences between peaks and valleys enhance bacterial killing by increasing the concentration of copper within the valleys where the majority of cells were observed. It is interesting to note that a subset of the S. epidermidis with membrane blebs was also associated with nanoflowers in the presence of PBS, indicating the organic material released from the damaged cells promoted the nucleation of organic copper phosphate crystals.
The formation of porous spheres following the exposure of bacteria to the phosphor bronze coating was surprising. These spheres were remarkably similar in appearance and size to nanoflowers that were self-assembled in the presence polypeptides, CuSO4 and PBS pH 7·4 after 3 days of incubation (Ge et al. 2012). However, in our study, primary crystals were visible as early as 30 min, reaching, within 2 h, a size comparable to those formed with purified proteins after 3 days. Ge et al. (2012) proposed that amide residues of proteins act as nucleators to promote the formation of copper phosphate petal-like structures that coalesce into protein inorganic nanoflower crystals.
By a similar mechanism, it is possible that a complex mixture of organic compounds derived from bacteria and a high accumulation of copper ions within the valleys where cells were concentrated greatly augmented the rate of crystal nucleation. Consistent with a nucleation role by bacteria-derived components, membrane disruption was also evident after only 30 min of incubation. The combined data indicate that a change from nanoscale to macroscale topography has pronounced impact on the biocidal efficacy of copper alloy.
The project was supported by the National Science and Engineering Research Council of Canada through its Engage Program to M.R. Authors also greatly appreciate the support of the Centre for Advance Coating Technologies of University of Toronto and for providing the Centre facilities for substrate evaluation.