Introduction

Silver has been used for its antimicrobial properties since ancient times: Alexander the Great refused to drink water which had not been stored in silver vessels, and Paracelcus claimed the beneficial properties of silver towards health (Silver et al. 2006). When penicillin was discovered and the era of the antibiotics began, the use of silver for its antimicrobial properties decreased (Klasen 2000a). Since biocide-resistant strains emerged, the interest to use silver as an antimicrobial agent is rising again (Klasen 2000b). However, antimicrobial agents based on ionic silver (e.g., silver nitrate) have one major drawback: they are easily inactivated by complexation and precipitation and thus have a limited usefulness (Atiyeh et al. 2007). Zerovalent silver nanoparticles are considered as a valuable alternative for ionic silver. Due to their large specific surface-to-volume ratio, nanoparticles have different properties than bulk material (Rao and Cheetham 2001; Tilaki et al. 2006). It has been shown that silver nanoparticles are antimicrobial towards a broad spectrum of Gram-negative and Gram-positive bacteria (Kim et al. 2007; Panacek et al. 2006). Furthermore, silver nanoparticles show antifungal (Kim et al. 2008) and antiviral activity (Elechiguerra et al. 2005; Lu et al. 2008). Besides their antimicrobial properties, silver nanoparticles can be used for their catalytic, conductive, and optical features (Medina et al. 2005; Fuller et al. 2002; Kamat 2002).

Many methods exist to synthesize nanoparticles. Most chemical methods use a reducing agent (e.g., sodium borohydride) to reduce Ag+ to Ag0 and a stabilizer (e.g., polyvinylpyrrolidone) to control particle growth and prevent aggregation (Ledwith et al. 2007). However, these chemical fabrication methods often have problems with particle stability and are difficult to scale up (Tavakoli et al. 2007). In addition, there is a demand for more environment-friendly production methods. An alternative production method is based on biological production.

Some bacteria are known to sorb heavy metals from contaminated wastewaters. Some microorganisms not only have a strong biosorptive capacity but also can reduce and precipitate them in their metallic form. Fungi, such as Verticillium, produce silver nanoparticles intracellularly (Sastry et al. 2003), while Fusarium oxysporum and Aspergillus niger have shown to produce silver nanoparticles extracellularly (Ahmad et al. 2003; Duran et al. 2005; Sadowski et al. 2008; Sastry et al. 2003). It is presumed that the reduction of silver ions is catalyzed by enzymes, as described for Klebsiella pneumoniae and Pseudomonas stutzeri AG259 (Klaus et al. 1999; Shahverdi et al. 2007). Another frequently reported mechanism for the bacterial reduction of silver is the non-enzymatic interaction of Ag+ with cell wall functional groups. This mechanism was observed for Lactobacillus A09, Corynebacterium sp. SH09, and Aeromonas SH10 (Fu et al. 2006; Lin et al. 2005; Zhang et al. 2005, 2007). Non-enzymatic reduction makes it possible to apply special conditions such as high or low pH and temperature, potentially accelerating the silver reduction. Furthermore, high concentrations of Ag+ that are toxic for bacteria can be used (Zhang et al. 2007).

In this paper, we describe the synthesis of silver nanoparticles using bacteria as reducing and capping agent. The capacity to precipitate silver was evaluated for Gram-positive bacteria Lactobacillus spp., Enterococcus faecium, Lactococcus garvieae, Pediococcus pentosaceus, and Staphylococcus aureus and the Gram-negative bacteria Pseudomonas aeruginosa and Escherichia coli. A high pH was thereby providing reduction rates similar to chemical processes and much faster than other biological methods.

Materials and methods

Bacterial strains and growth conditions

Lactobacillus farciminis LMG 9189, Lactobacillus parabuchneri LMG 11772, Lactobacillus plantarum LMG 24830, Lactobacillus fermentum LMG 8900, Lactobacillus brevis LMG 11437, Lactobacillus rhamnosus LMG 18243, L. plantarum LMG 24832, E. faecium LMG 24831, L. garvieae LMG 8162, and P. pentosaceus LMG 9445 (LMG culture collection, Ghent University, Gent, Belgium) were grown for 48 h in Man–Rogosa–Sharpe (MRS) broth (Oxoid, UK) at 30°C. P. aeruginosa LMG 9009 and S. aureus LMG 8064 were grown in tryptic soy broth (Difco, USA), and E. coli LMG 10266 was grown in Luria Bertani broth (Difco, USA) for 24 h at 37°C. All strains were grown with agitation (120 rpm).

Reagents

As ionic silver potentially forms oxide precipitates at high pH, silver was added as a diamine complex for all experiments. To prepare the diamine silver complex, ammonium hydroxide (Acros Organics, Belgium) was added to AgNO3 (Sigma-Aldrich, USA) until the original brown precipitate disappeared again. Subsequently, the concentration was diluted with MilliQ water to 26 mg/mL AgNO3.

Ag(I) bioreduction assay

Biomass was harvested by centrifugation in a Sorvall RC5C centrifuge (USA) at 5,000×g and washed two times with MilliQ water. Subsequently, biomass was resuspended in MilliQ, and the OD was adjusted to obtain 5 g/L cell dry weight (CDW). Diamine silver complex was added to the alkalified biomass (0.025 N NaOH) in a concentration of 1:5 Ag/CDW, and the suspension was shaken at a rotation rate of 100 rpm at 28°C. After 24 h incubation, the supernatant was separated from biomass by centrifugation at 10,000×g for 15 min. Biomass was washed three times with MilliQ to remove all ionic silver that was not irreversibly sorbed or not reduced on the biomass.

Influence of pH on the Ag(I) bioreduction

The reduction of silver was performed at pH 2, 4, 6, 7, 8, 9, and 11.5 with L. fermentum LMG 8900. The washed biomass and diamine silver complex were adapted separately to the desired pH. One, 3, 5, and 10 min and 2 h after adding the silver complex to the biomass, UV–VIS analysis was performed with a Uvikon 932 spectrometer (Kontron Instruments, Switzerland). Wavelength scans were taken in a scan range of 200–650 nm. The data range was 1 nm, and scan speed was 200 nm/min.

Ag analysis by atomic absorption spectrometry

To analyze the concentration of silver precipitated on the cells and in the supernatant, biomass and supernatant were boiled for 2 h with 65% HNO3 (Normapur, VWR, Belgium) and 30% H2O2 (Merck, Germany) until the sample was clear. Subsequently, the silver concentration was measured using a Shimadzu AA-6300 atomic absorption spectroscope (Japan).

X-ray diffraction analysis

X-ray diffraction (XRD) of the silver on the biomass, dried at 30°C, was performed with a BRUKER D8 Discover. X-rays were generated by a copper X-ray tube with power 1.6 kW (40 kV, 40 mA). The wavelength CuKα corresponded to 1.54 Å. Measurements were made between 25° and 90° 2-theta with a step time of 153.6 s and a step size of 0.02° 2-theta.

Transmission electron microscopy

To perform transmission electron microscopy (TEM), ultrathin sections of approximately 50 nm were prepared with the Ultracut microtome (Reichert-Jung A.G., Vienna, Austria) according to a previously published procedure (Mast et al. 2005), omitting OsO4 fixation, uranyl acetate incubation, and Reynold’s lead citrate staining to avoid background staining. Ultrathin sections made were placed on polioform- and carbon-coated copper grids and were examined using a Technai Spirit TEM (from FEI, Eindhoven, the Netherlands) operated at 120 kV. Digital micrographs were made using a bottom-mounted 4 × 4 k Eagle camera (Fei) at various magnifications. Micrographs were analyzed using AnalySIS Solution of the iTEM software (Olympus Soft Imaging Solutions). Particles were detected in a pre-defined region of interest excluding border particles and particles of less than 5 pixels. Particles were analyzed in ten sections of bacteria using different micrographs. The mean diameter was determined for each particle. Also, the area of each bacterium-containing particle was measured. To determine the particle size distribution, particles were counted in size ranges of 1 nm. The counts per range were plotted in a graph, and the cumulative frequency distribution was calculated.

Statistical analyses

SPSS for Windows version 16.0 was used for statistical analyses. Results were considered differently at an α two-tailed level of 0.05. Tests for normality of the data and equality of the variances were performed using the Kolmogorov–Smirnov and Levene’s test, respectively. Comparison of normal distributed data was performed with one-way analysis of variance. The nonparametric Kruskal–Wallis and Mann–Whitney U test were used to compare non-normal distributed data.

Results

Production of silver nanoparticles by different bacterial species

The ability of nine different Lactobacillus species to accumulate and subsequently reduce Ag+ was compared to P. pentosaceus, E. faecium, L. garvieae, S. aureus, E. coli, and P. aeruginosa. The test was performed under equal conditions for all species. For the different Lactobacillus species, the amount of silver associated with the biomass varied between 110 and 166 mg Ag/g cell dry weight (Table 1). Since 1,000 mg/L Ag+ was added as a diamine complex and the Ag/CDW ratio was 1:5, 200 mg/g was the theoretical maximum amount of silver that could be reduced. Consequently 83%, 75%, and 73% of the added silver was associated with the biomass of L. fermentum, L. farciminis, and Lactobacillus fructivorans, respectively. For the other lactic acid bacteria, P. pentosaceus, E. faecium, and L. garvieae, there was 141 mg/g CDW (70 %), 116 mg/g CDW (58%), and 146 mg/g CDW (73%) of biomass-associated silver, respectively. S. aureus demonstrated a recovery of 115 mg/g CDW (57%), while P. aeruginosa and E. coli had the lowest recovery of all with 107 mg/ g (53%) and 52 mg/g CDW (26%), respectively.

Table 1 The amount of silver associated with the biomass starting with an initial silver concentration of 200 mg/g CDW
Full size table

X-ray diffraction analysis was performed to evaluate the oxidation state of the silver associated with the biomass. Figure 1 shows the diffractogram for L. fermentum and S. aureus. In case of L. fermentum, typical peaks were present at 2θ = 38.24°, 44.46°, 64.21°, and 77.75° which is a typical diffraction pattern for metallic silver. The most intense peak corresponds to metallic silver (111) diffraction (2θ = 38.24°), while the less intensive peaks correspond to (200) diffraction (2θ = 44.46°), (220) diffraction (2θ = 64.21°), and (311) diffraction (2θ = 77.75°). All Lactobacillus spp., P. pentosaceus, E. faecium, and L. garvieae showed this typical diffraction pattern (data not shown). The diffractograms of L. fructivorans and P. pentosaceus indicated an additional peak that corresponded to the strongest Ag2O reflection (2θ = 33.33°). The diffractogram of S. aureus, P. aeruginosa, and E. coli did not show any typical silver peak nor the Ag2O reflection. Hence, only the lactic acid bacteria were confirmed to have the ability to produce Ag0.

Fig. 1
figure 1

XRD analysis of a L. fermentum with typical metallic silver diffraction peaks, b L. fructivorans with Ag2O and Ag(111) peak and c S. aureus without typical silver peaks (2-theta values typical for silver are mentioned between brackets)

Full size image

Particle size determined by transmission electron microscopy

TEM was performed on five of the best recovering and fast growing Lactobacillus spp.: L. fermentum, L. plantarum LMG 24830, L. plantarum LMG 24832, L. farciminis, and L. rhamnosus. TEM pictures showed that both particle localization and distribution inside the cell were dependent on Lactobacillus species (Fig. 2). L. fermentum produced two fractions of the silver nanoparticles. The most prevalent fraction consisted of small particles (5–15 nm) situated in the outer layers of the cell and on the cell wall. In contrast, the less prevalent fraction of silver nanoparticles, ranging from 15 to 40 nm, was intracellular and evenly distributed in the cytoplasm. Notably, the intracellular large nanoparticles included a few outliers between 60 and 80 nm. The nanoparticles produced by L. plantarum LMG 24830, L. plantarum LMG 24832, L. farciminis, and L. rhamnosus were situated more frequently inside the cytoplasm. Furthermore, the particles produced by these bacteria were somewhat larger compared to the particles produced by L. fermentum. In addition, outliers were more common than observed for L. fermentum, especially in case of the two L. plantarum strains. Finally, the nanoparticles appeared more clustered in the cytoplasm of L. fermentum.

Fig. 2
figure 2

TEM images of aL. fermentum, bL. farciminis, cL. plantarum LMG 24830, dL . plantarum LMG 24832, eL. rhamnosus after reduction of silver. The intense black spots are nanoparticles

Full size image

Based on the pictures taken by TEM (Fig. 2), nanoparticle size and the number of nanoparticles per section were determined for each of the five Lactobacillus species. Ten sections were analyzed for every species. Based on the frequency and cumulative frequency (Electronic supplementary material), 95% confidence interval (CI), mean particle size, mode, the number of particles per cell surface, the number of particles per cell volume, the number of particles per entire cell, and the concentration of Ag0 per cell were calculated (Table 2). Nanoparticles produced by L. fermentum were substantially different from those of the other Lactobacilli. While L. fermentum produced significantly more nanoparticles per surface area (280 ± 96 particles/μm2), volume (5,600 ± 1,922 particles/μm3), and cell (3,320 particles/cell) compared to the other species, the mean particle size (11.2 nm) was significantly smaller. The silver nanoparticles produced by L. fermentum showed the most narrow distribution as well, as reflected by the 95% CI (10.3 nm–12.1 nm). The concentration of Ag0 per cell (2.44 ± 0.45 × 10−11 g/cell) was equivalent with most other Lactobacillus species and was only significantly lower than that of L. plantarum LMG 24832 (6.77 ± 3.13 × 10−11 g/cell). Between the other Lactobacillus spp., the mean particle size, the number of particles per cell surface area, cell volume, and entire cell were comparable, with only L. rhamnosus having a smaller mean particle size (15.7 nm) and L. plantarum LMG 24832 producing significantly more particles per cell (1,565 ± 562). The mode of the particle size of silver nanoparticles produced by L. fermentum was between 6.0 and 8.0 nm. Similarly, the mode of the particle size produced by L. farciminis was between 6.0 and 7.0 nm. In contrast, the mode of particle size of L. plantarum LMG 24832 and L. rhamnosus was higher than that of L. fermentum, ranging between 9.0 and 10.0 nm

Table 2 Particle size and distribution for the particles on the cells were evaluated using TEM for five Lactobacillus species: mean particles size (95% CI), mode, particles/cell surface (standard deviation), particles/cell volume (standard deviation), and Ag0 concentration/cell (standard deviation) are given
Full size table

Influence of the pH on the silver reduction

The silver reduction in function of the pH was performed with L. fermentum (Fig. 3). There was a clear influence of the pH on the amount of silver associated with the biomass. When the pH increased, more silver was recovered; however, between pH 6 and 8, there was a slight decrease in the amount of silver associated with the biomass. The most alkaline pH tested, pH 11.5, gave the greatest recovery of silver.

Fig. 3
figure 3

The amount of silver associated with the biomass in function of the pH. Ag mg/g CDW increases with the pH. All assays were done in triplicate

Full size image

Additionally, pH influenced the rate of the reduction reaction. The suspension turned brown when silver was reduced, and the suspension coloring accelerated when increasing pH. At pH 11.5, the suspension colored dark brown within a minute. These observations were supported by UV–VIS analysis (Fig. 4). In case of pH 11.5, a broad absorption band appeared after 1 min. This plasmon resonance band was centered around 425 nm, and the intensity increased with time which indicated the rapid formation of silver nanoparticles. At lower pH, no such plasmon resonance bands were observed (results not shown).

Fig. 4
figure 4

UV–VIS spectra recorded as a function of time of the reaction of L. fermentum biomass with silver in alkaline conditions. A plasmon absorbance band centered at 425 nm increased with time

Full size image

Discussion

This paper describes the development of a method to produce silver nanoparticles using the biomatrix of bacteria both as a reducing and capping agent since the cell wall controls the growth of the nanoparticles and prevents their agglomeration. Several Gram-positive and Gram-negative bacteria were tested to reduce silver in a non-enzymatic way after application of a high Ag+ concentration in alkaline conditions. Only Lactobacillus spp. and the other lactic acid bacteria, P. pentosaceus, E. faecium, and L. garvieae, were able to reduce silver. TEM demonstrated that the localization of the nanoparticles inside the cell, the mean particle size, and the size distribution were species dependent (Fig. 2). The silver nanoparticles produced by L. fermentum had the smallest mean particle size and differed significantly of the nanoparticles produced by the other Lactobacillus species. Increasing pH enhanced the silver recovery and the reduction rate, creating a process which is fast compared to most other biological production methods (Nair and Pradeep 2002; Kalimuthu et al. 2008; Vigneshwaran et al. 2007). Moreover, the association of the nanoparticles with the bacterial biomass facilitates the concentration of the nanoparticles by centrifugation.

The non-enzymatic production of silver nanoparticles by Lactobacillus spp. allowed the application of high silver concentrations potentially toxic for the cell. We applied a very concentrated Ag+ solution (1 g/L), which is much higher compared to enzymatic processes. Enzyme-catalyzed silver reduction is restricted to lower silver concentrations ranging between 10 and 100 mg/L Ag+ (Kalimuthu et al. 2008; Kowshik et al. 2003; Shahverdi et al. 2007; Vigneshwaran et al. 2007). The amount of silver recovered depended on the species used. L. fructivorans recovered 73% and L. farciminis 75%, while L. fermentum even recovered 83%. These silver recoveries are comparable if not greater than the silver recovery of non-enzymatic processes using Corynebacterium SH09 (50%) and Aeromonas SH10 (67%) to produce silver nanoparticles (Zhang et al. 2005; Zhang et al. 2007).

The recovery of silver and the reduction rate were pH dependent. When pH increased, both recovery and the reduction rate increased (Figs. 3 and 4). After adding the AgNO3 solution to the biomass at pH 11.5, the suspension started to color brown within a minute. This is an indication that Ag+ was reduced to Ag0 and nanoparticles were formed, which was confirmed by UV–VIS analysis. The silver reduction was rapid when compared to other biological precipitation processes. For example, Bacillus licheniformis required 6 h before silver was reduced (Kalimuthu et al. 2008), whereas 12 and 72 h, respectively, were required before silver reduction by Aeromonas spp. SH10 (Nair and Pradeep 2002) and Aspergillus flavus (Vigneshwaran et al. 2007) was observed. Only Enterobacteria could synthesize silver nanoparticles rapidly: K. pneumoniae reduced silver in less than 5 min (Shahverdi et al. 2007). Since such rapid silver reduction rates are typically only achievable using chemical processes, our method combines the eco-friendliness of biological silver nanoparticle production with the speed of chemical processes.

Only the Lactobacillus species, P. pentosaceus, E. faecium, and L. garvieae were able to reduce silver as indicated by XRD (Table 1). S. aureus, however, and the Gram-negative bacteria P. aeruginosa and E. coli did not reduce silver using this method. The amount of silver recovered with the biomass was species dependent as well. Some lactic acid bacteria recovered more than 70%, and L. fermentum even recovered 83%. The Gram-negative bacteria on the other hand showed a lower recovery. P. aeruginosa recovered 53.5% and E. coli 26%. The species dependency of silver recovery is potentially due to differences in the specific composition of the biomass between species. Gram-positive bacteria possess a cell wall that consists of a thick layer of peptidoglycan, teichoic acids, lipoteichoic acids, proteins, and polysaccharides. Therefore, these bacteria have many anionic surface groups which may act as sites for the biosorption and subsequent reduction of silver cations (van Hullebusch et al. 2003). In contrast, Gram-positive S. aureus retained lower concentrations of silver on its biomass and, as demonstrated by XRD, did not reduce silver. Lactic acid bacteria are known to produce exopolysaccharides (EPSs) (Aslim et al. 2006; Boonaert and Rouxhet 2000; van Hullebusch et al. 2003). These EPSs may serve in the protection of the bacterial cells against toxic compounds such as metal ions (Aslim et al. 2006). Furthermore, EPSs, which provide additional sites for biosorption of silver ions, consist of repeating sugar units, mainly glucose, galactose, and rhamnose in different ratios (Jolly et al. 2002). Lin et al. (2005) suggested that reducing sugars of bacterial cells may function as the electron donor in the reduction of Ag+ to its metallic form Ag0. A suggested reduction mechanism for our production method is given in Fig. 5. Whenever pH increases, more competition occurs between protons and metal ions for negatively charged binding sites. Lin et al. (2005) discovered a slight drop in pH after biosorption of Ag+ on the biomass of Lactobacillus A09, which indicated the competition between protons and metal ions. High pH catalyzes the ring opening of monosaccharides such as glucose to their open-chain aldehyde (Fig. 5b). It is the aldehyde which delivers the reducing power. When metal ions are present, the aldehyde will be oxidized to the corresponding carboxylic acid, and at the same time, the metal ions will be reduced (Fig. 5c).

Fig. 5
figure 5

Suggested bacterial reduction of Ag+ to Ag0 at the cell surface. a Bacterial cell with reducing sugars such as glucose and protonated anionic functional groups (–RH). b When pH increases, protons dissociate and create negatively charged adsorption sites for Ag+. The reducing sugars turn into their open-ring structure and are now able to reduce Ag+. c The aldehyde function of the reducing sugar is oxidized to its carboxylic acid, while Ag+ is reduced to Ag0

Full size image

There are two kinds of EPSs: homopolymeric EPSs, which are synthesized by excreted enzymes, and heteropolymeric EPSs, which are synthesized in the cytoplasm and are subsequently polymerized on the outer cell prior to secretion (Jolly et al. 2002). Heteropolymeric EPSs may explain why the localization of the nanoparticles varies between the Lactobacillus species. Some Lactobacillus species produce an S-layer as well (Boonaert and Rouxhet 2000; Delcour et al. 1999). S-layers form a highly porous protein meshwork; therefore, they potentially act as a template for the nucleation of nanocrystals. (Merroun et al. 2007) used the S-layers of Bacillus sphaericus JG-A12 to produce gold nanoparticles. The silver nanoparticles produced by L. fermentum were distributed mainly in the outer layers of the cell and have a more narrow size distribution than the nanoparticles produced by the other Lactobacillus species. However, more research is necessary to clarify which cellular structures are truly responsible for the biosorption and reduction of silver.

The mean diameter of the silver nanoparticles produced by our method varied with the Lactobacillus spp. used. The smallest nanoparticles were produced by L. fermentum and had a diameter of 11.2 nm. This is smaller than the silver nanoparticles produced by the bacteria K. pneumoniae (52.5 nm), Aeromonas sp. SH10 (20 nm), and B. licheniformis (50 nm) (Kalimuthu et al. 2008; Sharverdi et al. 2007; Zhang et al 2007). In contrast, Zhang et al. (2005) reported the bacterial production of silver nanoparticles with a diameter of 9.9 nm produced by Corynebacterium SH09. The fungi Phoma sp.3.2883 and Verticillium produced silver nanoparticles with a mean particle diameter of 71.06 and 25 nm, respectively (Chen et al. 2003; Mukherjee et al. 2001), which is larger than the nanoparticles produced by our L. fermentum. A. flavus, however, produced smaller silver nanoparticles (8.92 nm) (Vigneshwaran et al. 2007). Of course, it is not just the mean size of the nanoparticle that is important but the size distribution as well. Of all the Lactobacillus species, L. fermentum had the most narrow size distribution, with a 95% confidence interval between 10.3 and 12.1 nm.

Silver nanoparticles can be applied in many areas; however, the most important field of application is likely the use of nanosilver for disinfection. Also, our silver nanoparticles showed antimicrobial properties (data not shown). Silver is used commonly in wound dressings for the treatment of burns (Atiyeh et al. 2007). Because silver ions are easily inactivated upon reaction with complexing agents in biofluids, they are often replaced by silver nanoparticles. Silver nanoparticles can be used for their catalytic, conductive, and optical properties as well (Medina et al. 2005; Fuller et al. 2002; Kamat 2002). Until now, most research was performed using chemically produced nanoparticles. As the characteristics of nanoparticles such as size, stability, shape, and therefore the activity depend on the production method, it will be important to investigate the activity of biologically produced nanoparticles in the future. Therefore, the spectrum of the antimicrobial activity towards different microorganisms will be examined more thoroughly in future research.