Silver Nanoparticles Templated by the M13 Phage Exhibit High Antibacterial Activity against Gram-Negative Pathogens and a Reduced Rate of Bacterial Resistance In Vitro

Abstract

Silver and silver nanoparticles (AgNPs) are well-known for their antibacterial properties. However, their low potency and the emergence of resistance are major barriers to the use of AgNPs in systemic therapy. Biological products can be used to reduce and cap AgNPs during synthesis from silver salts, but the structure and properties of biotemplated AgNPs are not understood well. We observed that AgNPs templated by the Escherichia coli phage M13 showed unusually high potency as well as activity against multiple Gram-negative pathogens, including E. coli, Pseudomonas aeruginosa, and Vibrio cholerae. The increased antimicrobial activity was attributable to the structural properties of the AgNPs rather than to the contributions from the phage. Furthermore, M13-templated AgNPs elicited bacterial resistance >10-fold more slowly than commercially purchased AgNPs and exhibited good cytocompatibility above the concentrations needed for bacterial inhibition. The improvements in antimicrobial properties obtained through biotemplating move AgNPs toward becoming a viable candidate for future systemic applications.

Introduction

Antimicrobial-resistant (AMR) bacteria caused >2.8 million infections in the United States in 2019, with a 20% increase in the number of cases caused by major AMR pathogens observed from 2019 to 2022. In light of the need for additional treatment options, silver nanoparticles (AgNPs) have been increasingly studied for their antibacterial properties. Multiple mechanisms for the antibacterial activity of AgNPs have been described − and may be present at the same time to varying degrees, depending on the system. AgNPs can interact directly with the cell membrane, leading to physical destabilization, cell leakage, and permeation of AgNPs into the cell. AgNPs also release ionic silver (Ag⁺), which interacts electrostatically with DNA and interferes with replication. At the same time, AgNPs and Ag⁺ interact with free radical scavengers such as glutathione, leading to the accumulation of reactive oxygen species (ROS), oxidative stress, and cell death.

While AgNPs are currently used for their antibacterial properties in a variety of products, including medical materials such as wound dressings and urinary catheters, , translation for therapy has been limited. Current clinical trials focus on topical therapy for nasopharynx, skin, and wound conditions. Recent reviews highlight the challenges of overly complex formulations, the lack of superiority to standard-of-care antibiotics in terms of efficacy, and potential side effects from long-term exposure and accumulation. , Potential toxicities are a significant concern; for example, some wound dressings contain AgNPs at a concentration of ∼50–100 mg mL^–1, while reduced cell viability and hemolysis are observed in vitro at substantially lower concentration ranges (<100 μg mL^–1). In addition, the development of bacterial resistance is a critical consideration for new antimicrobial agents. Although it has been suggested that bacterial resistance to AgNPs may be inhibited by multiple mechanisms of action, resistant strains are known clinically, and multiple studies have demonstrated the development of resistance while culturing Escherichia coli in subinhibitory concentrations of AgNPs. Resistance mechanisms include mutations in the copper efflux pump (cus) system to export silver ions and the production of flagellin to trigger aggregation and thus inactivation of AgNPs. − Therefore, AgNP syntheses yielding simple formulations, increased potency enabling smaller exposures, and delayed bacterial resistance are desirable.

AgNPs are commonly synthesized chemically from reduction of silver salts (e.g., AgNO₃) in the presence of a capping agent (e.g., citrate) or physically by top-down methods that break down silver materials (e.g., mechanical milling and laser ablation). Alternatively, biological syntheses, using biological products or extracts as the reducing and capping agents, may be advantageous for biocompatibility and environmental reasons or may yield AgNPs with different properties. , Biological synthesis of AgNPs has been performed using a variety of substrates, including plant leaves or plant extracts containing reducing molecules, fungal cells that reduce silver ions intracellularly, or microbes secreting reductase enzymes to yield extracellular AgNPs either attached to intact cells or reduced by the spent medium solution. , However, cells and cellular extracts are chemically complex and poorly defined, leading to uncertainty and batch-to-batch variability in the resulting bionanomaterial. When considering defined biological reagents, although isolated amino acid solutions are not capable of producing AgNPs, certain peptide sequences, especially containing glutamic acid residues, as well as double-stranded DNA, can mediate reduction and capping. − Silver reduction by peptides is also promoted by peptide display on a structured biological template, such as whole yeast cells or the phage M13. ,

M13 is a rod-shaped, nonlytic phage that infects F+ E. coli, having a diameter of ∼7 nm and a length of ∼1 μm. The capsid consists primarily of thousands of copies (∼2700) of major coat protein g8p, with minor coat proteins g3p and g6p, and g7p and g9p constituting the cylinder ends (Figure a). − The g8p protein consists of 50 amino acids, including two α-helices connected by a hinge region around residue 21 (Tyr21). The positively charged C-terminal domain (residues 40–50) interacts electrostatically with the phage genomic DNA inside the virion, while the negatively charged hydrophilic N-terminal domain (residues 1–20) is exposed to the extracellular milieu. The two domains are connected by an intermediate hydrophobic region (residues 21–39). The exposed N-terminal domain may present multiple sites for bioreduction at electron-rich residues. When incubated with AgNPs made by laser ablation, M13 can spontaneously form electrostatically assembled virion–AgNP networks for spectroscopic sensing applications. , M13-templated growth of AgNPs has been used to synthesize a chromium-sensing nanomaterial. Interestingly, these M13-templated AgNPs showed modest antibacterial activity (minimum concentration of 86.4 μg mL^–1 to inhibit E. coli growth). However, little is known about the antibacterial potential of the M13-templated AgNPs. In this study, we used M13 to synthesize AgNPs and characterized their structure and properties, demonstrating unusually high antibacterial potency, low cytotoxicity, and substantially slowed development of bacterial resistance.

1.

M13-templated silver nanoparticles. (a) M13 is a filamentous coliphage whose major coat protein, g8p, has a surface-exposed N-terminal domain. Solvent-accessible electron-rich residues include Lys8 and Tyr21. (b) M13-seeded silver reduction results in AgNP synthesis, as shown by the LSPR absorbance spectrum of AgM13_2h (blue) and AgM13_1d (red) compared to a AgNO₃ solution (yellow) and M13 alone (gray). AgNP spectra were normalized by the maximum value for comparison of the peak shape. The inset shows photographs of representative synthesis products. (c) TEM micrographs show AgNPs associated with phages for AgM13_1d (top row), which are larger than AgNPs from AgM13_2h (bottom row), consistent with the shift in the LSPR peak. Scale bar of 50 nm. (d) AgM13_2h synthesis was conducted with controls and various M13 phage mutants. Nanoparticle formation was quantified by calculating the area under the curve (AUC) of UV–vis absorbance spectra from 400 to 800 nm using Simpson’s rule implemented via the SciPy library in Python. AUC values represent the mean of three independent replicates, and error bars denote the standard error of the mean (SEM). Data for control samples are colored red and blue, those of wild-type (WT) M13 green, and those of M13 mutants purple, orange, and yellow.

Experimental Section

Amplification of M13 Phage

E. coli K12 ER2738 cells (New England Biolabs, USA) were cultured overnight in Luria broth (LB) supplemented with tetracycline (10 μg mL^–1) at 37 °C. A 200 μL aliquot of this overnight culture was inoculated into 20 mL of LB in a 250 mL Erlenmeyer flask. To propagate wild-type M13 phages, 1 μL of the M13KE phage solution (10¹³ pfu mL^–1) was added to the 20 mL culture, followed by incubation at 37 °C with shaking (250 rpm) for 4 h. The resulting culture containing both E. coli ER2738 and M13KE was transferred to 1 L of LB and incubated overnight at 37 °C while being shaken (250 rpm). Bacterial cells were pelleted by centrifugation (4500g for 10 min), and the supernatant was collected in a fresh tube. This centrifugation step was repeated thrice to ensure complete removal of bacterial debris. To precipitate the phages, 800 mL of the final supernatant was collected from top, transferred to a flask containing 200 mL of a 2.5 M NaCl/20% PEG-8000 solution, and mixed briefly. The solution was stored overnight at 4 °C, after which the phage precipitate was collected by centrifugation (8200g for 30 min at 4 °C) for subsequent use.

Synthesis of Silver Nanoparticles with the M13 Phage

Two protocols were used to synthesize silver nanoparticles with M13. In one protocol, the M13 precipitate described above was purified by a second precipitation, a AgNO₃ solution was added, and the mixture was stirred for 1 day, resulting in AgM13_1d. In the second protocol, a AgNO₃ solution was added to the M13 precipitate described above, the mixture stirred for 2 h, and then the second precipitation conducted, yielding AgM13_2h. For comparison, a third nanomaterial was prepared with AgNPs purchased commercially, incubated with M13, and precipitated, resulting in Ag_comm-M13. These three procedures were conducted as follows.

AgM13_1d

The M13 phage precipitate was dissolved in 50 mL of water and centrifuged again at 10 000 rpm to pellet cell debris. The top 40 mL of supernatant was transferred to a fresh tube. The M13 phages were precipitated a second time by mixing with 10 mL of the 2.5 M NaCl/20% PEG-8000 mixture. The solution was incubated at 4 °C for 1 h and centrifuged at 14 000 rpm for 10 min. The supernatant was discarded, and the pellet was resuspended in 10 mL of water. After measuring the phage concentration, the solution was stored at 4 °C. The phage concentration was determined spectrophotometrically by measuring the difference in absorbance at 269 and 320 nm and applying the following formula:

$$\mathrm{phage\; concentration}(\mathrm{virions}{\mathrm{mL}}^{-1})=\frac{A_{269}-A_{320}}{\mathrm{no.\; of\; bases\; per\; virion}}\times 6.0\times {10}^{16}$$

For the synthesis of silver nanoparticles (AgNPs), 0.05 mL of M13 phage (10¹³ pfu/mL) was added to 5 mL of Milli-Q water, and the pH was adjusted to 8.0–8.5 using dilute NaOH (1 mM). Subsequently, 0.05 mL of a 100 mM AgNO₃ solution was added dropwise under constant stirring, resulting in a final silver ion concentration of 1 mM. The appearance of an initially turbid solution implied the start of the nucleation process. The solution was stirred in the dark at room temperature for 1 day (24 ± 2 h). The change in color from the initial color of gray to light brown and finally to dark reddish-brown indicated the formation of AgNPs. The solution was centrifuged at 6000 rpm for 15 min and washed with Milli-Q water to remove excess phage and unreacted silver ions. The washing step was repeated thrice, and finally, the precipitate (named AgM13_1d) was resuspended in Milli-Q water and stored at 4 °C until further use.

AgM13_2h

Alternatively, the M13 phage precipitate was dissolved in 5 mL of water and centrifuged again at 10 000 rpm to pellet cell debris, and the supernatant was transferred to a fresh tube. The phage concentration was determined by spectrophotometry. Then, 0.05 mL of M13 phage (10¹³ pfu/mL) was added to 5 mL of Milli-Q water (pH 7, adjusted with 1 mM NaOH when needed). Subsequently, 0.05 mL of a 100 mM AgNO₃ solution was added dropwise under constant stirring. The solution was stirred at room temperature in the dark for 2 h; then, 1.25 mL of a 2.5 M NaCl/20% PEG-8000 mixture was added to the solution and briefly mixed. The solution was stored at 4 °C overnight. The gray precipitate was then centrifuged at 14 000 rpm for 10 min and washed with water to remove excess phage, unreacted Ag ions, and PEG. The washing step was repeated thrice, and finally, the precipitate (named AgM13_2h) was resuspended in Milli-Q water and stored at 4 °C until further use.

AgNP_comm-M13

As a control, citrate-capped silver nanoparticles with a particle size of 20 nm were purchased from Millipore and used without further modification. The commercial AgNPs were adsorbed with M13 phage (10¹³ pfu/mL) using overnight stirring. The precipitate was collected after centrifugation at 6000 rpm, washed with water, dried, and stored at 4 °C.

Acid Digestion of Silver–Phage Conjugates

Phage conjugate samples were transferred to a clean Teflon vessel, and digestion was carried out with concentrated HNO₃ (65–70%, trace metal grade, Fisher Scientific) with a supplement of H₂O₂ (30%, certified by ACS, Fisher Scientific) at 190 °C for 20 min in a microwave digestion system (Titan MPS, PerkinElmer). Once the sample was cooled to room temperature, it was diluted to a final volume of 10 mL by adding filtered deionized water.

Estimation of the Silver Concentration Using ICP-MS

To estimate the Ag concentration in each sample, inductively coupled plasma mass spectrometry (ICP-MS, NexION 2000, PerkinElmer) was performed. Samples were acid-digested prior to analysis. A calibration curve was established using a standard solution of a silver salt. The dwell time was set to 50 ms with 30 sweeps. Measurements were conducted in triplicate, with background correction was applied by subtracting background noise using blanks run before and after each sample.

Transmission Electron Microscopy

A carbon-coated EM grid (Ted Pella Inc., 200 mesh) was secured with reverse tweezers, ensuring the carbon-coated side facing upward. A 5 μL aliquot of the sample was applied using a 10 μL pipet and allowed to stand for 5 min. Excess solution was removed by blotting with a filter paper. The grid was then washed with 5 μL of ultrapure water, blotted, and stained with 2% aqueous uranyl acetate (UA) for 30 s. The staining solution was blotted with filter paper, followed by two washes with 10 μL of ultrapure water for 30 s each. The grid was then air-dried for 1 h and consequently imaged using an FEI Titan 80-300 TEM instrument (at an accelerating voltage of 300 kV) equipped with a Gatan Ultrascan1000 4MP CCD instrument.

Phage Mutants of g8p

Three g8p mutants were constructed: K8A, Y21A, and TEYI19–22AAAA (Table S1). The K8A and TEYI19–22AAAA phage mutants were constructed using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies, USA), following the manufacturer’s instructions. Mutagenic oligonucleotides were designed using the Agilent QuikChange Primer design program. The Y21A mutant was constructed using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs, USA), according to the manufacturer’s instructions, and transformed into Mix & Go Competent E. coli Cells (Zymo Research, USA) using kanamycin selection. Mutagenic oligonucleotides for this mutation were designed in Benchling (https://www.benchling.com/). All oligonucleotides for both site-directed mutagenesis kits were obtained from Integrated DNA Technologies (IDT, USA).

To confirm successful mutagenesis, transformed colonies were plated on LB agar supplemented with kanamycin and isopropyl β-d-1-thiogalactopyranoside (IPTG) and incubated overnight at 37 °C. Individual colonies were inoculated into LB medium containing 50 μg mL^–1 kanamycin and 200 μM IPTG and cultured overnight at 37 °C with shaking to induce phage production. Phage genomes were extracted with the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions and submitted for whole-plasmid sequencing (Plasmidsaurus, USA). Resulting sequences were analyzed by using Benchling.

Phage mutants were propagated by inoculating their respective phage-producing E. coli ER2738 cells in LB medium with 50 μg mL^–1 kanamycin and 200 μM IPTG at 37 °C overnight with shaking. Cultures were centrifuged at 7200g for 15 min to pellet bacterial cells. The supernatant containing phage particles was sequentially filtered through a 0.22 μm syringe filter, concentrated using a 50 kDa Amicon Ultra Centrifugal Filter (Millipore Sigma, USA), and passed through a second 0.22 μm filter to ensure sterility. Purified phage preparations were resuspended into Milli-Q water and adjusted to pH 8, and silver nitrate was added as described for AgM13_2h.

Propagation of Bacteria

E. coli K12 ER2738 (F+), E. coli DH5α (F−) (both obtained from New England Biolabs, USA), and E. coli O157:H7 (ATCC700927 obtained from ATCC) were cultured in LB medium. A single colony of each E. coli stain was selected and grown in 5 mL of LB (for ER2738, the medium was supplemented with 10 μg mL^–1 tetracycline) at 37 °C while being shaken for 24 h.

Vibrio cholerae strain 0395 (a gift from M. Mahan, University of California, Santa Barbara, CA) was grown from inoculation by a single colony into 5 mL of LB without antibiotics at 37 °C with shaking overnight.

Pseudomonas aeruginosa (Schroeter) Migula (ATCC 25102) was propagated in ATCC Medium 3 nutrient broth with a single colony inoculated into 5 mL of broth without antibiotics and incubated at 37 °C with shaking for 24 h.

Bacterial Growth Assay

Overnight bacterial cultures were diluted in their respective growth media to an OD₆₀₀ of 1.25 × 10^–3 (approximately 10⁶ cfu mL^–1 assuming an OD₆₀₀ of 1.0 corresponds to ∼8 × 10⁸ cfu mL^–1). Growth was monitored in 96-well clear-bottom sterilized microplates by using a multimode microplate reader (Infinite 200Pro, Tecan, Switzerland). Then, 200 μL of the diluted bacterial cultures was transferred to each well. As a blank control, medium without any bacteria was used. The plates were agitated by linear shaking for 1 h at 37 °C. Then, nanoparticle samples (20 μL) at varying concentrations were added and mixed thoroughly. The samples were placed in the microplate reader and agitated by linear shaking for a total of 30 h at 37 °C, and OD₆₀₀ was measured at an interval of 30 min. Each condition was tested with three replicates. The mean OD₆₀₀ and the corresponding standard error of the mean (SEM) for each sample were calculated for each time point and plotted as a function of time.

To assess the effect of nanoparticle samples later in growth (i.e., exponential phase), the procedure described above was modified as follows. Nanoparticle samples were added after bacterial incubation for 6 h. The cultures were agitated by linear shaking for a total of 50 h at 37 °C, and OD₆₀₀ values were measured as described above.

Growth Rate and AUC Measurements

Growth rates were calculated by linear fitting of log­(OD₆₀₀) over time during the exponential phase. The area under the curve (AUC) was calculated using trapezoid integration over a 30 h period, with data points spaced at 30 min intervals.

Minimum Inhibitory Concentration Measurement

The minimum inhibitory concentration (MIC) was determined by the broth microdilution method in 96-well microtiter plate format, following an established protocol. The MIC was defined as the lowest silver concentration that completely inhibited visible bacterial growth in LB after 24 h, where visible growth was judged by the presence of turbidity (cloudiness) throughout the broth or a pellet of cells at the bottom of the well, as specified by the Clinical and Laboratory Standards Institute (CLSI) M07 standard.

Minimum Bactericidal Concentration Measurement

Overnight bacterial cultures of E. coli ER2738 or E. coli DH5α were inoculated in LB medium at 37 °C with shaking. The overnight cultures were diluted in LB medium to an optical density at 600 nm (OD₆₀₀) of 0.2. Bacteria at an OD₆₀₀ of 0.2 were then diluted 1:20 in LB medium. In triplicate, each bacterial strain was treated with AgNP (AgNP_1d, AgNP_2h, or AgNP_comm) at varying concentrations (4, 2, 1, 0.4, 0.2, 0.1, and 0.04 μg mL^–1) in a 96-well plate, along with untreated bacteria as a growth control. The plate was incubated statically at 37 °C for 24 h. After 24 h, the OD₆₀₀ was read in a plate reader (Infinite 200Pro, Tecan). For each well with an OD₆₀₀ of ≤0.1 as well as the lowest set of wells that grew to an OD₆₀₀ above 0.1, 50 μL of the culture from the 96-well plate was spread on an LB plate, which was then incubated for another 24 h at 37 °C. After 24 h, the plates were assessed for bacterial colonies. The lowest concentration of AgNP resulting in a 99.99% reduction in colony-forming units per milliliter was the MBC.

Measurement of the Development of Bacterial Resistance to AgNPs

Bacteria (E. coli K12 ER2738) were repeatedly exposed to subinhibitory concentrations of the AgNPs over 15 successive culture rounds in microplates, performed in triplicate. A dispersion of a AgNP solution (80 μg mL^–1 Ag) was serially diluted in LB to 40, 20, 10, 5, 2.5, 1.25, 0.63, 0.31, 0.16, 0.08, and 0.04 μg mL^–1. Each well was inoculated with bacteria at 10⁶ cfu mL^–1. Cultures were incubated at 37 °C for 24 h, and the minimum inhibitory concentration (MIC) was determined as the lowest silver concentration that prevented visible bacterial growth after 24 h. Immediately after each 24 h incubation period, 10 μL of medium containing surviving bacteria was collected from the three wells (when available) with the highest subinhibitory (i.e., below MIC) silver concentrations, subcultured in 15 mL of LB (supplemented with 10 μg mL^–1 tetracycline) at 37 °C for 24 h, and used to prepare a fresh inoculum (10⁶ cfu mL^–1) for the next cultivation round. After 15 cultivation rounds, the final MICs of AgNPs against the conditioned bacterial strains were determined using the same microdilution method, starting with a dispersion of AgNPs (160 μg mL^–1 Ag).

Bacterial Biofilm Dispersal Assay

Bacterial cells (E. coli K12 ER2738, E. coli DH5α, and E. coli O157:H7) were grown overnight at 37 °C with shaking at 250 rpm in their respective medium; the cell concentration was determined by OD₆₀₀, and the culture was diluted to a final estimated concentration of 5 × 10⁷ cfu mL^–1, as described above. For the crystal violet (CV) assay, 200 μL of the bacterial culture was added to each well of a 96-well glass-bottom microplate and incubated at 30 °C for 48 h without shaking. The liquid was removed from each well and replaced with fresh medium with or without nanoparticle samples. The microplate was further incubated at 30 °C for 24 h without shaking. The liquid was removed, and each well was stained with 125 μL of 1% (w/v) CV solution and incubated for 15 min at room temperature. The CV solution was removed completely, and wells were air-dried, followed by the addition of 125 μL of absolute ethanol to solubilize the remaining CV. The absorbance was measured at 520 nm by using a multimode microplate reader (Infinite 200Pro, Tecan). All conditions were assessed in triplicate.

Live/Dead Microscopy of Bacterial Biofilms

Overnight cultures of E. coli K12 ER2738 and E. coli O157:H7 were diluted to a final concentration of 5 × 10⁷ cfu mL^–1 as described above. An aliquot of 800 μL of diluted bacterial culture was added to each well of an eight-well chambered glass-bottom slide (ibidi, USA) and incubated at 30 °C for 48 h without shaking. The liquid was subsequently removed from each well and replaced with fresh medium, with or without nanoparticle samples, followed by incubation at 30 °C for 24 h without shaking. For staining, the medium was removed, and the wells were washed three times with PBS (pH 7.0) to remove planktonic cells. Then, 500 μL of a live/dead staining solution containing 3.3 μM SYTO 9 (Molecular Probes, Life Technologies, USA) and 10 μg mL^–1 propidium iodide (PI) (Sigma, St. Louis, MO) in PBS (pH 7.0) was added to each well. After incubation in the dark for 20 min, the staining solution was carefully removed, and the wells were rinsed twice with ultrapure water. Samples were embedded by using 1% low-melting point agarose for subsequent microscopic analysis.

The microscopic visualizations were performed using a fluorescence microscope (BZ-X800, Keyence, USA) and a laser-scanning confocal microscope (CLSM) (TCS SP8 Digital Light Sheet Microscope, Leica Microsystems, Wetzlar, Germany). SYTO-9, which stains bacteria with intact membranes (i.e., live bacteria), was excited at a wavelength of 488 nm using an LED (BZ-X800) or a laser line (SP8). Emission was collected within the ranges of 500–550 nm (BZ-X800) or 500–530 nm (SP8). PI, which labels bacteria with compromised membrane integrity (i.e., dead bacteria), was excited with a 594 nm LED (BZ-X800) or 552 nm laser line (SP8) with emission collected within the ranges of 615–680 nm (BZ-X800) or 575–625 nm (SP8). Images were acquired sequentially using 20×, 40×, and 63× dry objective lenses.

Biofilm Quantification from Microscopy

All image (80 μm × 80 μm field of view) analyses were conducted using BiofilmQ version 1.0.1 in the MATLAB (version R2025a, Mathworks, USA) environment. Bacterial cells were distinguished from the background based on the PI/SYTO 9 fluorescence staining. Prior to segmentation, a mean filter with a kernel size of 19 pixels (x–y) and 7 pixels (z) was applied to reduce photon shot noise. The floating cell suppression algorithm in BiofilmQ was employed to minimize segmentation artifacts caused by planktonic cells. Cellular signals were enhanced relative to the background using a top-hat filter with a disk-shaped kernel larger than the expected bacterial size (5.21 μm diameter per plane or 65 voxels) applied to each x–y plane within the z stack. Segmentation was performed using an Otsu threshold with a sensitivity setting of 0.012. Following segmentation, the total biovolume was determined.

The image stacks (in separate fluorescence channels) were divided into cubic volumes with edge lengths of ≈1.2 μm (15 voxels), such that each volume contained approximately one cell on average. Binary images were used to measure local cell–cell distances (center-to-center distance), local cell density, mean fluorescence intensity, the ratio of the two fluorescent channels, and the distance of each cube from the biofilm–liquid interface (∼1 μm resolution) within the biofilm. All properties were quantified locally within a spherical region of 50 pixels (4 μm) around each cell for all biofilm cells. Data were exported as VTK files, and complex surface renderings, including the spatial distribution of object parameters, were visualized using ParaView (version 5.13.3, Kitware Inc., USA).

Biofilm Matrix Microscopy

For visualizing the density and architecture of the extrapolymeric substance (EPS) in biofilms, we used FilmTracer SYPRO ruby biofilm matrix stain (Thermo Fisher Scientific, USA) following the manufacturer’s protocol. In brief, the medium was removed, and the wells were washed three times with ultrapure water to remove planktonic cells; 500 μL of the biofilm matrix stain was added to each well. After incubation in the dark for 30 min, the staining solution was carefully removed, and the wells were rinsed twice with ultrapure water. Samples were embedded using 1% low-melting point agarose and imaged using CLSM (TCS SP8 Digital Light Sheet Microscope, Leica Microsystems). A 458 nm argon gas laser was used for excitation, and emission was collected at 615–680 nm.

Cytotoxicity Assay

The MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay was used to check the biocompatibility of AgM13_1d and AgM13_2h. HEK293T cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 10% fetal bovine serum (FBS) (Gibco) and 1.0% penicillin/streptomycin (Gibco) and cultured at 37 °C, 5.0% CO₂, and 95% humidity. The cells (1.0 × 10⁴) were seeded in each well of a 96-well plate and cultured in 10% FBS-supplemented DMEM overnight. The medium was changed, and the cells were then incubated with various concentrations of different nanoparticle samples in fresh medium for 24 h. Then, the media were changed with fresh media with MTT and incubated for 4 h, and the assay was performed using an MTT assay kit (CyQUANT MTT Cell Viability Assay, Invitrogen) following the manufacturer’s instructions.

Hemolysis Assay

AgNPs were diluted to concentrations of 0.2, 1, 1.5, or 2 μg mL^–1 in 1× PBS. Five milliliters of 10% sheep red blood cells (MP Biomedicals) was washed five times by spinning down at 500g for 5 min, removing the supernatant, and resuspending to the original volume with 1× PBS buffer. Washed blood cells were diluted in 1× PBS to obtain a 4% red blood cell solution. To a sterile 1.5 mL Eppendorf tube was added 300 μL of the 4% red blood cell solution with 300 μL of each sample (AgNP_1d, AgNP_2h, and AgNP_comm), as well as a 1× PBS sample for the negative control and a 2% Triton X-100 sample for the positive control. Mixtures were incubated for 1 h at 37 °C. The samples were then spun down at 500g for 5 min. In triplicate, 50 μL of supernatant and 150 μL of 1× PBS were added to a 96-well plate. The absorbance at 405 nm was measured using a plate reader, and the hemolytic activity was determined using the following equation:

$$\%\mathrm{hemolysis}=({\mathit{A}}_{\mathrm{sample}}-{\mathit{A}}_{\mathrm{PBS}})/({\mathit{A}}_{\mathrm{Triton\; X‐100}}-{\mathit{A}}_{\mathrm{PBS}})$$

where A _sample, A _PBS, and A _{Triton X‑100} are absorbance values at 405 nm of the sample, negative control, and positive control, respectively. Measurements were performed in triplicate. Calculations were made by using Microsoft Excel.

Stability of AgNPs in PBS

AgNPs were diluted to a final concentration of 0.2 μg mL^–1 in 1× PBS and then incubated with rotation at 37 °C. Aliquots were taken at specific time points (0, 1, 2, 4, 12, 24, and 30 h) for measurement of the hydrodynamic diameter, ζ potential, and UV–vis absorbance spectrum. The hydrodynamic diameter was measured by dynamic light scattering (DLS), and ζ potential was measured by electrophoretic light scattering using a Malvern Zetasizer Nano ZSP (Malvern Instruments, U.K.). Photons were collected at a scattering angle of 173°, and data were collected by Zetasizer Software. The cell contained 1 mL of a sample solution and was equilibrated for 10 s before each measurement. Measurements represent the average of three readings. To measure absorbance spectra, AgNPs were diluted from their stock solutions to a final concentration of 20 μg mL^–1 in 1× PBS and incubated at 37 °C. Aliquots of AgNPs were taken for UV–vis absorbance measurement at various time points (0, 1, 2, 4, 12, 24, and 30 h). Triplicate measurements were averaged.

Statistics

Quantitative data are expressed as the mean ± standard deviation (μ ± σ) or mean ± standard error of the mean (μ ± SEM), as stated. One-way analysis of variance (ANOVA) was followed by Tukey’s test and post hoc false discovery rate (FDR) correction (two-stage step-up method) for comparisons among multiple groups. Sample size refers to biological replicates. GraphPad Prism version 9.0 (GraphPad Software Inc., USA) was used for statistical analysis. Origin Pro 9.1 (OriginLab Corp., USA) was used for calculation of growth curve parameters. For all comparisons, a p value of <0.05 was considered statistically significant.

Results and Discussion

Synthesis of AgNP Nanomaterials Using Phage M13

M13 was used for in situ bioreduction of silver ions to silver nanoparticles (AgNPs) and self-assembly of a nanofibrous network. M13 was produced and secreted by E. coli, purified from the culture medium by precipitation with PEG/NaCl, and mixed with AgNO₃ in an aqueous solution under continuous stirring and ambient temperature. While M13 and AgNO₃ do not absorb detectably in the visible range, localized surface plasmon resonance (LSPR) from AgNPs produces a visible color. Overnight stirring yielded a reddish-brown suspension, a color change that indicated the formation of AgNPs, and the nanomaterial was collected by centrifugation and washed with water, yielding an AgNP nanomaterial called “AgM13_1d”. Alternatively, aiming to produce smaller particles, M13 was mixed with a AgNO₃ solution and reacted for a brief time (2 h), resulting in a blue-purple suspension. This product was precipitated with PEG/NaCl and washed, yielding an AgNP nanomaterial, called “AgM13_2h”. A solution lacking M13 and containing only AgNO₃ and PEG/NaCl showed no visible precipitation under the same conditions, indicating that M13 was necessary to produce the AgNPs.

The interaction of light with noble metal nanoparticles can cause a synchronized oscillation of conduction electrons known as localized surface plasmon resonance (LSPR). The absorption spectrum is highly dependent on nanoparticle geometry, particularly size. , The LSPR spectra of the AgNPs were determined by the UV–vis absorbance. Both AgM13_1d and AgM13_2h showed wide absorbance bands, indicating heterogeneous size distributions (Figure b). AgM13_2h exhibited a blue-shifted spectrum (λ_max ∼ 330 nm) compared to AgM13_1d (λ_max ∼ 410 nm), suggesting that a shorter stirring time resulted in smaller particles, as expected. The silver concentrations of the nanomaterials were determined by ICP-MS of acid-digested samples. ICP-MS analysis showed silver concentrations of 19.6 ± 0.2 and 19.9 ± 0.4 μg mL^–1 (μ ± σ) in AgM13_1d and AgM13_2h, respectively. The concentrations of silver nanoparticles in the studies described are given in terms of Ag concentration based on these experimentally determined silver concentrations such that comparisons can be made between different nanoparticles at the same Ag concentration. The synthesis yields of AgM13_1d and AgM13_2h were found to be 18.2% and 18.4%, respectively.

Transmission electron microscopy (TEM) of AgM13_1d and AgM13_2h showed electron-dense nanoparticles attached to filamentous phages in an interlinked structure (Figure c). As with the LSPR spectra, TEM indicated that AgNPs in AgM13_1d were larger than in AgM13_2h (diameters of 31.6 ± 8.4 and 16.3 ± 6.3 nm (μ ± σ), respectively) (Figure S1). AgM13_2h also showed association of phages with each other to form bundle-like structures, consistent with the PEG precipitation step used to collect the nanomaterial product.

High-Resolution TEM Characterization of AgM13_1d and AgM13_2h Nanomaterials

High-resolution TEM (HR-TEM) was used to analyze the internal structures of the AgNPs in AgM13_1d and AgM13_2h. The presence of distinct lattice fringes indicated that the AgNPs in both nanomaterials were crystalline, with faceted surfaces (Figure ). HR-TEM of AgM13_1d (Figure a) showed lattice spacings of 0.234 and 0.290 nm, attributable to the {111} and {110} planes of silver, respectively, indicating twinned particles with {110} and {111} facets exposed. The hexagonal symmetry of diffraction spots in the selected area electron diffraction (SAED) pattern (Figure b) indicated spherical face-centered-cubic (fcc) crystals bounded by the lowest-energy {111} facets. HR-TEM of AgM13_2h (Figure c) showed a lattice spacing of 0.167 nm attributable to {211} planes of silver. The corresponding SAED pattern (Figure d) confirmed preferential exposure of the {211} plane corresponding to fcc crystals.

2.

High-resolution TEM (HR-TEM) and selected area electron diffraction (SAED) pattern of silver–M13 nanoparticles. (a) Bright field HR-TEM image of a single AgM13_1d nanoparticle showing lattice fringes with interplanar spacings of 2.3 and 2.9 Å, corresponding to the (111) and (110) crystallographic planes, respectively. (b) SAED of the AgM13_1d nanoparticle showing a single-crystal diffraction pattern with 6-fold symmetry, characteristic of spherical face-centered cubic (fcc) nanocrystals predominantly exposing low-energy {111} facets. (c) Bright field HR-TEM image of a single AgM13_2h nanoparticle displaying lattice fringes with an interplanar distance of 1.67 Å, attributed to the (211) plane. (d) SAED pattern of the AgM13_2h nanoparticle that reveals preferential exposure of the {211} facets within the fcc crystal structure.

Influence of the Major Coat Protein (g8p) on Silver Reduction

The hydrophilic N-terminal domain of g8p comprises the sequence (residues 1–20) AEGDD­PAKAA­FNSLQA­SATE, followed by residues Y21 and I22 in the hinge region. To determine whether reduction of Ag⁺ ions to metallic Ag⁰ by M13 phage involved specific surface-exposed amino acid residues of the major coat protein (g8p), three g8p mutants were designed to test the effect of electron-rich residues: single mutants K8A and Y21A and a quadruple mutation of residues 19–22 (TEYI) to alanines (AAAA). These phage mutants were used to synthesize AgNPs (AgM13_2h), and the yield of nanoparticle formation was measured by absorbance spectroscopy (Figure S2). Both K8A and the quadruple mutant showed significant reduction in AgNP production compared to wild-type M13 (37% reduction, p = 0.0058, and 25% reduction, p = 0.016, respectively; two-tailed t-test) (Figure d). In contrast, the Y21A mutation did not affect the synthesis yield (p = 0.70; two-tailed t-test). These results support the involvement of multiple specific residues of g8p, particularly Lys8 in the reduction of silver ions.

AgM13_1d and AgM13_2h Inhibition of E. coli, V. cholerae, and P. aeruginosa Growth in a Liquid Culture

Antimicrobial activities of AgM13_1d, AgM13_2h, and commercially available citrate-capped AgNPs (30 nm diameter) were assayed by growth inhibition for four Gram-negative bacterial species: E. coli ER2738, E. coli DH5α, P. aeruginosa (ATCC 25102), and V. cholerae O395. Ag concentrations of 0.04, 0.1, and 0.2 μg mL^–1 were tested for each nanomaterial. E. coli ER2738 is an F+ strain that supports M13 phage infection. In the absence of AgNPs, E. coli ER2738 showed rapid growth with a stationary phase attained after ∼8 h (Figure a). However, cultures of E. coli ER2738 containing 0.2 μg mL^–1 AgM13_1d did not show growth within 30 h (Figure a, red). Growth inhibition was dose-dependent, as curves at lower Ag concentrations (0.04 and 0.1 μg mL^–1) exhibited extended lag phases (approximately 9 and 20 h, respectively). While AgM13_1d extended the lag phase at these lower concentrations, the growth rate (exponential phase) was affected only at the higher concentration (0.1 μg mL^–1) (Figure S3a). AgM13_2h showed greater potency, with complete inhibition of growth for 30 h at the lowest concentration (0.04 μg mL^–1) tested (Figure a, blue). In contrast, commercially purchased AgNPs (AgNP_comm) had no significant effect on the growth of E. coli ER2738 at the equivalent Ag concentration range (Figure a, yellow), consistent with previous measurements showing a minimum inhibitory Ag concentration of 26.75 μg mL^–1 (i.e., >100-fold higher than the concentrations tested here) for citrate-capped AgNPs on E. coli.

3.

Effect of AgM13_1d, AgM13_2h, and AgNP_comm nanoparticles on the growth of various bacterial strains: (a) E. coli ER2738, (c) E. coli DH5α, (e) P. aeruginosa, and (g) V. cholerae following treatment with AgNPs at Ag concentrations of 0.04, 0.1, and 0.2 μg mL^–1 (Ag concentration determined by ICP-MS). Data are presented as means ± SEM (n = 3) biologically independent replicates. Areas under the curve (AUC), representing bacterial growth potential, are shown for (b) E. coli ER2738, (d) E. coli DH5α, (f) P. aeruginosa, and (h) V. cholerae. AUC values are expressed as the mean ± SD. Statistical significance was assessed using one-way analysis of variance (ANOVA) for all groups of the same species, followed by Tukey’s test with false discovery rate (FDR) post hoc correction (n = 3); p values compared to the control are indicated in the corresponding graphs.

The area under the growth curve (AUC, also called “growth potential”) , was used to quantify overall inhibition of bacterial growth. All tested concentrations of both AgM13_1d and AgM13_2h significantly reduced the growth potential of E. coli ER2738, while AgNP_comm did not (Figure b). In addition, the final OD₆₀₀ was determined from each growth curve (Figure S3e), showing the significantly reduced final OD₆₀₀ for AgM13_1d at 0.1 μg mL^–1 and above, and for all tested concentrations of AgM13_2h. Minimum inhibitory concentrations (MICs) were determined by the broth microdilution method for E. coli ER2738. The MICs of AgM13_1d and AgM13_2h were 0.08 and 0.16 μg mL^–1, respectively, while AgNP_comm showed an MIC of 5 μg mL^–1. These MIC measurements indicate that AgM13_1d and AgM13_2h are approximately >30 times more potent than AgNP_comm. Similarly, minimum bactericidal concentrations (MBCs) showed that AgM13_1d and AgM13_2h were 32-fold more potent than AgNP_comm against E. coli ER2738 (Table ), in agreement with the MIC improvements.

1. Minimum Bactericidal Concentrations (MBCs) for AgNP Types against Two E. coli Strains.

	MBC (ER2738) (μg mL–1)	MBC (DH5α) (μg mL–1)
AgM131d	0.4	0.4
AgM132h	0.4	1
AgNPcomm	16	16

To check whether inhibition of E. coli ER2738 was due to M13 infection or M13 targeting, we measured growth inhibition for E. coli DH5α, an F– strain that is not infected by M13 due to the absence of the bacterial receptor. Similar inhibition of growth was observed for DH5α (Figure c,d and Figure S3b,f), although DH5α was less susceptible to the lowest tested concentration of AgM13_2h. MBC measurements (Table ) also showed a ≤2-fold difference in potency against DH5α versus ER2738. This comparison indicates that M13 infection plays a minor role in growth inhibition compared to the silver nanomaterial. Extending this logic, growth inhibition was measured for other nonhost organisms. P. aeruginosa showed complete growth inhibition for 30 h at 0.2 μg mL^–1 AgM13_1d or 0.1 μg mL^–1 AgM13_2h (Figure e,f and Figure S3c,g), similar to E. coli DH5α. V. cholerae was more susceptible compared to the other bacterial species, with complete inhibition of growth for 30 h at 0.1 μg mL^–1 AgM13_1d or 0.04 μg mL^–1 AgM13_2h (Figure g,h and Figure S3d,h). In summary, both AgM13_1d and AgM13_2h significantly reduced the rate of growth of multiple Gram-negative bacterial species while the same concentrations of AgNP_comm did not inhibit bacterial growth.

AgM13_1d and AgM13_2h Inhibition of E. coli Biofilm

E. coli ER2738, E. coli DH5a, and E. coli O157:H7 were cultured on glass-bottom microplates under static conditions at 30 °C for 48 h to allow the initial formation of biofilms. The culture medium was then replaced with fresh medium containing AgM13_1d or AgM13_2h (0.2 μg mL^–1 Ag) or AgNP_comm (0.2 μg mL^–1 Ag), followed by incubation for an additional 24 h without shaking. After treatment, biofilms were washed and stained with crystal violet to quantify the extracellular polymeric substances (EPS). The results show >50% reductions in EPS upon treatment of each of the three E. coli strains with AgM13_1d or AgM13_2h (Figure a–c). AgNP_comm showed less inhibition, with EPS reduced by 10–40% for the three E. coli strains.

4.

Influence of AgM13_1d, AgM13_2h, and AgNP_comm nanoparticles on E. coli biofilms. (a–c) Quantification of biofilm biomass by the crystal violet assay for three different E. coli strains. The absorbance at 520 nm reflects the extent of biofilm formation following exposure to different AgNP formulations. Data are expressed as the mean ± SD (n = 3). Statistical significance was assessed using one-way ANOVA for groups of the same species, followed by Tukey’s test with false discovery rate (FDR) post hoc correction (n = 3); p values for comparison to the control are indicated in the corresponding graphs. (d) Confocal laser scanning microscopy (CLSM) maximum projection images of E. coli ER2738 biofilms treated with various silver nanoparticles, stained with SYTO 9 and propidium iodide (PI). SYTO 9 (green fluorescence) indicates viable cells, while PI (red fluorescence) indicates compromised membranes, suggesting dead cells. (e) Three-dimensional reconstructions of corresponding biofilms showing the spatial distribution of live (green) and dead (red) cells following nanoparticle treatments as labeled.

The E. coli K12 ER2738 and E. coli O157:H7 biofilms were subjected to SYTO9/PI live/dead staining and confocal laser scanning microscopy (CLSM). SYTO9 permeates all bacterial cells, producing green fluorescence, while propidium iodide (PI) selectively stains cells with compromised membranes with red fluorescence. Maximum projection images (Figure d and Figure S4a) indicated bacterial cell death and reduced growth following AgM13_1d or AgM13_2h treatment, compared with no treatment or AgNP_comm treatment. Similarly, the bacterial biovolume was determined by computational image analysis and showed similar qualitative trends (Figure e and Figure S4b). Staining with a SYPRO Ruby biofilm matrix also showed a microscopically less dense matrix after treatment with AgNPs (Figure S4c), supporting the reduction observed from the crystal violet assay.

AgM13_1d Inhibition of Exponentially Growing E. coli

To determine whether AgM13_1d was able to inhibit and/or kill cells growing in the exponential phase, varying concentrations (0.1–0.6 μg mL^–1) of AgM13_1d were added at midexponential phase (OD₆₀₀ of 0.2–0.4) of an E. coli ER2738 culture. A concentration-dependent inhibitory effect was observed (Figure a). Addition of intermediate concentrations (0.3 and 0.4 μg mL^–1) of AgM13_1d caused a transient decrease in OD₆₀₀, which then recovered some growth after 24 h. Addition of high concentrations (≥0.5 μg mL^–1) of AgM13_1d resulted in reduced OD₆₀₀, suppression of growth for >48 h, and cell death, indicated by a lack of colony-forming units (CFUs) (Figure b).

5.

Bactericidal activity and bacterial resistance to AgM13_1d. (a) E. coli culture was grown to midexponential phase, and then AgM13_1d nanomaterial was added. Growth without AgM13_1d (black curve) or with different concentrations of AgM13_1d (red curves) was monitored by OD₆₀₀. The concentration of AgM13_1d is shown on the right side in micrograms per milliliter. (b) End point samples were plated for the control (no AgM13_1d), 0.2 μg mL^–1 AgM13_1d, or 0.6 μg mL^–1 AgM13_1d, to count colony-forming units (CFUs). No CFUs were observed at 0.6 μg mL^–1 AgM13_1d, consistent with the observed decrease in OD₆₀₀. (c) The E. coli culture was propagated over 15 rounds of serial passage in the presence of AgM13_1d (red), AgM13_2h (blue), or commercially purchased AgNPs (yellow). The minimum inhibitory concentration (MIC) was determined after each round, and the development of resistance was monitored by an increase in MIC. Commercial AgNPs show a high starting MIC and a 2-fold increase in MIC every two or three rounds. Black dots indicate the MIC above the detection range (>40 μg mL^–1) for AgNP_comm; the MIC after round 15 was determined to be 120 μg mL^–1. In contrast, AgM13_1d and AgM13_2h exhibit low starting MICs and a 2-fold increase in MIC after eight rounds.

AgM13_1d and AgM13_2h Show Slower Bacterial Resistance to Silver Nanoparticles

To assess the rate of bacterial resistance development, we conducted 15 successive cultivation rounds for E. coli ER2738 in medium containing subinhibitory concentrations of AgM13_1d, AgM13_2h, or AgNP_comm and determined the MICs of the silver nanomaterials after each round (Figure c and Table S2). Resistance to AgNP_comm developed rapidly, with a 2-fold increase in the MIC observed every two or three rounds, such that the MIC had increased to 40 μg mL^–1 by round 9. The MIC after round 15 was determined to be 120 μg mL^–1. In contrast, the MICs for AgM13_1d and AgM13_2h remained stable at 0.08 and 0.16 μg mL^–1, respectively, through round 8. A 2-fold increase in the MIC was observed in round 9 for both AgM13_1d and AgM13_2h, with MICs reaching 0.16 and 0.31 μg mL^–1, respectively. The MIC for AgM13_1d then remained constant through round 15, and the MIC for AgM13_2h increased 2-fold after round 14. These results demonstrate that subinhibitory exposure led to the rapid development of resistance against AgNP_comm, which showed an MIC increase of 24-fold over 15 rounds. In comparison, resistance developed slowly against AgM13_1d and AgM13_2h, whose MICs increased only 2- and 4-fold, respectively, over the same period.

Effect of AgNP_comm-M13 on E. coli Growth

The enhanced inhibitory effect of AgM13_1d and AgM13_2h compared to commercially purchased AgNPs could be due to either the filamentous, networked structure of the nanomaterial or different properties of the silver nanoparticles themselves when grown with M13. To determine whether a filamentous network structure was sufficient to enhance the inhibitory effect, we prepared a noncovalent nanoconjugate by mixing AgNP_comm with M13, producing “AgNP_comm-M13”. TEM analysis showed the formation of a filamentous network in AgNP_comm-M13 (Figure a). We evaluated whether AgNP_comm-M13 inhibited the growth of E. coli ER2738. While cultures with AgM13_1d (0.2 μg mL^–1) and AgM13_2h (0.2 μg mL^–1) showed no growth for 24 h, as expected, cultures with AgNP_comm-M13 (0.2 μg mL^–1) showed little effect on growth (Figure b). These results indicated that a networked structure containing AgNP_comm and M13 was not sufficient to cause the inhibitory effect observed for AgM13_1d and AgM13_2h.

6.

AgNP_comm-M13 and acid-digested AgM13_1d. (a) TEM micrographs showing the networked structure of nanomaterial (AgNP_comm-M13) made from commercially purchased AgNPs and M13. (b) AgNP_comm-M13 did not exhibit substantial antibacterial activity compared to AgM13_1d and AgM13_2h at the same concentration (0.2 μg mL^–1 Ag) on E. coli ER2738. (c) TEM micrographs of acid-digested AgM13_1d indicate a loss of phage structures and the presence of aggregates of silver nanoparticles and amorphous organic matter. (d) Acid-digested AgM13_1d showed antibacterial activity with a modest decrease in potency on E. coli ER2738, as digestion caused a 4–6 h decrease in lag time compared to undigested AgM13_1d at the same concentration. No nanomaterial was added to the control.

Effect of Acid-Digested AgM13_1d on E. coli Growth

If the inhibitory effects of AgM13_1d and AgM13_2h were due to properties of the silver nanoparticles when grown with M13, then removal of the biological matter and networked structure should preserve the antibacterial activity. AgM13_1d was subjected to digestion with concentrated nitric acid and H₂O₂ to degrade the M13 phages. TEM of the acid-digested material showed aggregates of AgNPs and unstructured organic matter without filamentous phage structures (Figure c). When tested on a liquid culture of E. coli ER2738, acid-digested AgM13_1d (0.03 and 0.09 μg mL^–1) largely retained inhibitory activity, with a <3-fold loss of potency (Figure d; digested AgM13_1d at 0.09 μg mL^–1 showed greater activity than undigested AgM13_1d at 0.03 μg mL^–1). These results support the attribution of the enhanced antibacterial properties of AgM13_1d to the silver nanoparticles synthesized by this method.

Cytotoxicity and Hemolysis Assay of AgM13_1d and AgM13_2h on Mammalian Cells

AgM13_1d and AgM13_2h were added to human embryonic kidney (HEK293) cells and assayed by the MTT assay for metabolic activity. There was no significant decrease in cell activity at 0.2–1.9 μg mL^–1 AgM13_1d or 0.2–2.8 μg mL^–1 AgM13_2h (Figure a), suggesting cytocompatibility at concentrations >10-fold above the MICs on E. coli ER2738. The hemolytic activity was measured by incubating 0.2–2 μg mL^–1 AgNPs with sheep red blood cells. Less than 2% hemolysis was observed over this concentration range (Figure b), indicating a lack of hemolytic activity.

7.

Cytotoxicity assessment of AgM13_1d and AgM13_2h nanoparticles. (A) Metabolic activity was measured in human embryonic kidney (HEK293) cells using the MTT assay. Cell activity is expressed as the mean ± standard deviation (SD) from three biologically independent replicates (n = 3). Statistical significance was determined by one-way analysis of variance (ANOVA), followed by Tukey’s test and false discovery rate (FDR) post hoc correction. Corresponding p values for pairwise comparisons are indicated in the graphs. No significant decreases were noted compared to those of the control. (B) Hemolytic activity was measured in sheep red blood cells (n = 3). Less than 2% hemolysis was observed.

Stability of AgNPs in a Saline Solution

The salt concentration is known to cause aggregation of AgNPs, including biogenic AgNPs. To assess the colloidal stability of the AgNPs (0.2 μg mL^–1) in PBS, the ζ potential was measured by electrophoretic light scattering. ζ potentials (ζ) for all AgNPs were negative, as expected, with the ζ of AgM13_1d in the range of −16 to −25 mV and the ζ for AgM13_2h in the range of −8 to −17 mV. AgNP_comm showed an upward trend in ζ, while there was no obvious trend for AgM13_1d and AgM13_2h over 30 h (Figure S5a). The hydrodynamic diameter (D _h) and polydispersity index (PDI) in PBS were estimated by dynamic light scattering. The PDI was high (0.4–1) for all nanoparticles, consistent with the appearance by TEM for AgM13_1d and AgM13_2h, (Figure c). The PDI and average D _h of AgM13_2h appeared consistent over time, but for AgNP_comm and AgM13_1d, both PDI and D _h showed an increasing trend over 30 h (Figure S5b,c). UV–vis absorbance spectra in PBS showed peak broadening over time for AgNP_comm and AgM13_1d (20 μg mL^–1) with AgM13_2h showing less pronounced broadening (Figure S5d–f). These results are consistent with aggregation over time in PBS, with the stability of AgM13_1d being similar to that of commercially purchased AgNPs, and the stability of AgM13_2h being superior.

In biological syntheses of AgNPs, biomolecules act as both a reducing agent and a capping agent. In peptides, promotion of AgNP synthesis is sequence-dependent. With rod-shaped viral templates, introduction of cysteine residues, silver-binding peptides, or tetraglutamic acid peptides (E4) can promote silver reduction as nanoparticles or nanowires. Basic residues (Lys and Arg) can also interact with silver surfaces. Although amine groups are generally considered weak reductants, primary amines within amino acids can serve as binding sites for silver ions and may reduce silver ions. It is noteworthy that silver binding alone does not imply AgNP synthesis and may actually inhibit aggregation of silver atoms to form the AgNP nucleus. Considering the surface-exposed domain of g8p in M13 (AEGDD­PAKAAF­NSLQAS­ATEYI), mutant K8A was found to reduce the level of AgNP synthesis by 37%, implicating this specific residue. Aromatic amino acids are also electron-rich, and indeed, tryptophan (Trp) residues in peptides have been demonstrated as reducing agents. Tyrosine can reduce silver ions at high pH, with oxidation of an ionized phenol to quinone leading to AgNP synthesis (Figure S6). However, mutation Y21A had little effect, indicating that tyrosine was not important for the reduction in this system. Other likely contributors are acidic residues, such as E20. Overall, silver reduction by M13 in our system appears to be mediated by multiple charged residues of g8p.

The M13-templated AgNPs were surprisingly potent as antibacterial agents, with an MIC as low as 0.08 μg mL^–1, compared to that of approximately size-matched commercially purchased AgNPs tested here (AgNP_comm, 20 nm diameter, having an MIC of 5 μg mL^–1) as well as previously reported M13-templated AgNPs (MIC of 86.4 μg mL^–1). For context, most currently approved antibiotics have MIC susceptibility break points for pathogens in the range of 0.03–16 μg mL^–1 (Clinical Laboratory Standards Institute). Although a direct comparison between conventional antibiotics and nanomaterials is difficult, lower MICs would be desirable for AgNPs given bioaccumulation concerns. AgM13_1d and AgM13_1d were also effective at reducing E. coli biofilms at relatively low concentrations (0.2 μg mL^–1) compared to that of AgNP_comm.

To determine whether the antibacterial effect was due to the silver nanoparticles themselves versus the phage network, we conducted complementary tests: disrupting the phage network of AgM13_1d by acid digestion and creating an M13 network together with AgNP_comm. Acid digestion had a modest effect on antibacterial activity (<3-fold decrease in potency), which may be due to disruption of the network and/or alteration of the AgNP surfaces by digested organic matter. Conversely, combining M13 with AgNP_comm also gave only a weak effect on antibacterial activity compared to AgNP_comm alone. This result is consistent with other studies showing that the simple combination of phage with AgNPs does not produce major increases in potency. In particular, the combination of a silver-binding mutant of lytic coliphage T7 with commercially purchased AgNPs resulted in a <2-fold improvement over the antibacterial activity of T7 alone, and the combination of a lytic Staphylococcus aureus phage with AgNPs showed a straightforward additive antibiofilm effect. Overall, our results indicated that, once the AgNPs had been synthesized, the phage played a minor role in the antibacterial effect, thus attributing the main antibacterial effect to the synthesized silver nanoparticles themselves. Consistent with this, AgM13_1d and AgM13_2h exhibited antimicrobial activity on multiple Gram-negative pathogens (E. coli, P. aeruginosa, and V. cholerae), regardless of expression of the bacterial receptor (F pilus), indicating that phage specificity does not affect the antimicrobial activity of the nanomaterials.

Multiple factors influence the antibacterial potency of AgNPs, including size and crystalline ultrastructure, which are in turn determined by the synthesis parameters, including temperature, pH, concentration, and incubation time. The literature suggests that bacterial responses to silver nanoparticles are generally dissimilar to bacterial responses to ionic silver. Indeed, the MICs of AgM13_1d and AgM13_2h (0.08–0.16 μg mL^–1 silver) were substantially lower than the MICs of AgNO₃ (approximately 2–10 μg mL^–1 silver), , indicating that bulk dissolved silver is unlikely to account for most of the antibacterial property of AgM13_1d and AgM13_2h. To understand the structural basis for high potency, high-resolution TEM on AgM13_1d showed preferential exposure of the {111} crystal plane, consistent with its low surface energy (∼0.76 J m^–2) compared to the other crystal planes in the Ag nanocrystal. , The {111} facets have been shown to be more antibacterial than the {100} facets that tend to be predominant in other silver nanoparticles. , Interestingly, the {211} plane was also observed at the surface of AgM13_2h, which indicates the presence of a more reactive surface, as well as multiple modes of crystal growth. Thus, reduction and growth on M13 resulted in AgNPs with unusual crystal structures and antimicrobial potency. Whether these structures translate into increased potency via activity of the nanoparticle surface itself or localized dissolution of silver is unknown.

Despite early optimism that silver might avoid bacterial resistance due to its multiple mechanisms of antibacterial action, empirically, bacterial resistance to silver nanoparticles has been observed during serial passaging under subinhibitory conditions. , The major mechanisms of resistance appear to be the production of flagellin, which promotes AgNP precipitation, and mutations in the heavy metal efflux system (particularly cusS in E. coli). Previous efforts to overcome resistance focus on resensitizing the bacteria, including inhibiting flagellin production, rather than AgNP optimization. Here, we observed that M13 templating can slow the rate of emergence of resistance by >10-fold. For example, while AgNP_comm treatment resulted in a 24-fold increase in MIC after 15 rounds of bacterial passaging, AgM13_1d treatment resulted in an only 2-fold increase in MIC after the same number of passages. Although we do not know the mechanism, reduced bacterial resistance would be consistent with the crystal structure of M13-templated AgNPs, which display more reactive facets. In addition to the large increase in potency and the apparent cytocompatibility observed, the substantial reduction in the rate of bacterial resistance observed in vitro suggests that M13 templating of AgNPs could have significant advantages over previously studied AgNPs. However, further study is required regarding the stability of AgM13_1d and AgM13_2h, particularly with respect to aggregation, in relevant physiological media.

Conclusions

Bacterial resistance to current treatments, particularly among Gram-negative pathogens, motivates a search for new antibacterial strategies, including nanomedicine. While silver is well-known for its antibacterial properties and silver nanoparticles are considered to be promising formulations, their potency, potential for bacterial resistance, and cytocompatibility are important issues that must be addressed to consider systemic treatment. Here we report that silver reduction and AgNP synthesis on phage M13, promoted by charged residues on the major coat protein g8p, result in AgNPs with >30-fold increased potency, >10-fold slower emergence of bacterial resistance, and cytocompatibility in vitro. These properties may be attributable to greater exposure of reactive planes of the silver crystals when templated on M13. Looking ahead, while phage manufacturing is still developing with respect to batch-to-batch consistency, purity, and GMP compatibility, a number of clinical trials attest to the technical ability to overcome these challenges. Continued improvements in biotemplation could lead to AgNPs that may enable future systemic applications pending in vivo evaluation.

Data for all results are presented in the manuscript and Supporting Information files.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.langmuir.5c03695.

• Tables of primer sequences and MICs during serial culturing and figures showing size distributions, UV–vis spectra, growth rates, confocal microscopy images, stability data, and a diagram of the mechanism (PDF)

D.B. and A.A. contributed equally to this work. Conceptualization: D.B. and I.A.C. Investigation and formal analysis: D.B., A.A., K.M., D.K., and Y.C. Validation: A.A. and K.M. Visualization: A.A., D.B., and K.M. Writing of the original draft: D.B., A.A., K.M., and I.A.C. Review and editing: A.A., D.B., K.M., D.K., Y.C., and I.A.C. Funding acquisition and supervision: I.A.C. Revision: A.A., K.M., and I.A.C.

The authors declare the following competing financial interest(s): I.A.C. is a co-founder of Paralos Bioscience.