Varied-shaped gold nanoparticles with nanogram killing efficiency as potential antimicrobial surface coatings for the medical devices

Medical device-associated infections are a serious medical threat, particularly for patients with impaired mobility and/or advanced age. Despite a variety of antimicrobial coatings for medical devices being explored to date, only a limited number have been introduced for clinical use. Research into new bactericidal agents with the ability to eradicate pathogens, limit biofilm formation, and exhibit satisfactory biocompatibility, is therefore necessary and urgent. In this study, a series of varied-morphology gold nanoparticles in shapes of rods, peanuts, stars and spherical-like, porous ones with potent antibacterial activity were synthesized and thoroughly tested against spectrum of Candida albicans, Pseudomonas aeruginosa, Staphylococcus aureus clinical strains, as well as spectrum of uropathogenic Escherichia coli isolates. The optimization of gold nanoparticles synthesis allowed to develop nanomaterials, which are proved to be significantly more potent against tested microbes compared with the gold nanoformulations reported to date. Notably, their antimicrobial spectrum includes strains with different drug resistance mechanisms. Facile and cost-efficient synthesis of gold nanoparticles, remarkable bactericidal efficiency at nanogram doses, and low toxicity, underline their potential for development as a new coatings, as indicated by the example of urological catheters. The presented research fills a gap in microbial studies of non-spherical gold nanoparticles for the development of antimicrobial coatings targeting multidrug-resistant pathogens responsible for device-associated nosocomial infections.


Results
Synthesis and physicochemical analysis of rod-and peanut-shaped gold nanoparticles. The morphology of the synthesized Au NPs was analyzed by scanning transmission electron microscopy (STEM) (Fig. 1). The nanoparticles synthesized over 0.5 h had an elongated rod-like shape with sharply-rounded ends (AuR NPs) with size of ~ 45 ± 8 nm along the longitudinal axis and ~ 10 ± 3 nm along the transverse axis, Fig. 1A4. In addition, it was noticed that when the transverse axis increased, the longitudinal axis decreased. Extending the synthesis time to 3 h, led to the formation of peanut-like gold nanoparticles with size of ~ 60 ± 5 nm along the longitudinal axis, ~ 25 ± 5 nm along the transverse axis, Fig. 1B4. However, in the STEM images of the AuP NPs it can be seen that the core of the nanoparticles had a rod shape and was surrounded by a gold shell, which caused the shape of the obtained nanoparticles to change their shape to peanut-like. This effect was caused by the CTAB and silver nitrate (AgNO 3 ) solutions, which were used up during the synthesis. The generated AgBr-ions block the growth of the AuNPs in the direction 26,27 . Therefore, the nanoparticles were expected to grow parallel to the 16 planes forming elongated structures as for the AuR NPs in Fig. 1A1,A2. However, it was noted that extending the synthesis time caused the ends of the edges to start growing in the direction, creating the AuP NPs (Fig. 1B1,B2), which was likely due to depletion of the CTAB and the AgNO 3 particles in the solution at the end of the synthesis 26 . In Fig. 1C1,C2 star-shape AuNPs, which cores have a cube-like shape and 144 nm size, as well as 243 nm at the farthest ends of the stars' arms the diagonal size. These shapes od Au NPs was caused by increasing the reducer concentration. It is responsible for the decrease of time, which was needed for the reduction of chloroauric acid (HAuCl 4 ). Furthermore, high reaction speed prevents any reaction between the CTAB and AgNO 3 solutions and consequently, nanoparticles can growth in non-dominant direction created star arms. Moreover, STEM images of AuSph (70C), Fig. 1D1,D2 showed spherical shape of NPs with sizes around 44 ± 5 nm, but these nanoparticles have pores. These pores were formed as a result of the high temperature synthesis. The condition under high temperature caused increase of the gold precursor reduction speed and unblocking of the nanoparticles growth in the direction by destabilization of CTAB molecules 26 . Moreover, STEM images of AuSph (CTAB) NPs, Fig. 1E1,E2 showed that obtained nanoparticles have spherical shape and average size around 10 nm. Very similar result was obtained for AuSph NPs synthesized using trisodium citrate, Fig. 1F1,F2, which also have spherical shape and around 8 nm size.
To determine the structure of the synthesized nanoparticles, selected area electron diffraction (SAED) patterns from the analyzed nanoparticle samples were acquired (Fig. 1A3-F3). The obtained results revealed that all synthesized Au NPs were crystalline. Moreover, the SAED patterns showed that the randomly oriented obtained gold nanoparticle crystallites could be indexed with the lattice parameter of gold 28  www.nature.com/scientificreports/ in SAED appear sharper and contain more and more spots with increasing nanoparticle size. Therefore, the SAED from AuS NPs (Fig. 1C3) have the sharpest rings, while SAED obtained for both spherical Au NPs have fuzzy rings, Fig. 1E3,F3. However, the SAED pattern taken from the AuP NPs was composed of blurred rings, in contrast to the pattern for AuR NPs, which showed visible spots. This may be because the AuP NPs were thicker (Fig. 1B1,B2), particularly at their ends, compared with AuR NPs (Fig. 1A1,A2). Moreover, UV-Vis spectra of Au NPs, showed, that the position of SPR peaks depends on the shape of nanoparticles, Fig. 1A5-F5. For AuR NPs, Fig. 1A5 and AuP NPs, Fig. 1B5, SPR positions were around 768 nm and 775 nm for rod-and peanut-shaped gold nanoparticles, respectively. For AuS NPs, the maximum absorbance was around 680 nm, Fig. 1D5, while porous nanoparticles have SPR peak with wide range between 500 and 900 nm, Fig. 1D5. Furthermore, SPR peaks position for both spherical Au NPs were around 525 nm, Fig. 1E5,F5. To confirm the stability of varied-shaped gold nanoparticles, zeta potential measurements were performed and results of these analyses are presented in Fig. 1A6-F6. Regardless of nanoparticles shape, all synthetized materials were positively charged in pH range 3.5-12.5. Accordingly, for AuR NPs the zeta potential values range from ~ 49 for pH = 3.5 to 18 mV for pH = 12.5, while for AuP NPs, these values are between 42 and 19 mV (Fig. 1A6,B6, respectively). For star-shaped gold nanoparticles zeta potential values range from 33 to 16 mV (Fig. 1C6). Zeta potential values for all spherical gold nanoparticles (Fig. 1D6-F6) are between 35 and 4 mV and depend on the pH solution. For all synthesized nanoparticles the potential value decreases as the pH value increases.
Developed nanomaterials exert potent bactericidal and fungicidal activity against large spectrum of microbial pathogens. Varied-shaped gold nanoparticles were characterized by their antimicrobial activity against representative isolates of C. albicans fungi, Gram-negative bacteria from E. coli and Pseudomonas aeruginosa species, as well as Gram-positive Staphylococcus aureus bacteria. As determined using a colony counting assay (Fig. 2) and MIC/MBC/MBIC measurements (Table 1), non-spherical nanoparticles, as well as those porous ones, exert potent fungicidal and bactericidal activity against wild spectrum of microbial pathogens at nanogram doses. Although some variations in susceptibility are recorded among the species and  www.nature.com/scientificreports/ single isolates, a majority of microbial population is noted to be eradicated at concentration of 0.5-1 ng mL −1 (Fig. 2). At the same time, no significant effect was observed for both AuSph NPs (despite the employment of doses up to 25 µg mL −1 ) and CTAB-functionalized gold nanospheres, which point out the great potential of shaping morphology of nanomaterials in order to improve their killing activities. When comparing average survival of tested strains, no considerable differences between non-spherical gold nanoparticles were detected, since a median of survival is ranging mostly from 10 to 25% for all tested isolates. Only for C. albicans and E. coli isolates, AuR NPs were noted to be slightly weaker in microbial eradication, but this effect did not differ statistically ( Fig. 2B,D). When comparing the susceptibility of pathogens, E. coli and S. aureus bacteria were the most prone to non-spherical AuNPs-mediated treatment (Fig. 2B,D). High antimicrobial efficiency was also confirmed in the presence of growth medium following 24 h incubation with the tested compounds ( Table 1). The broth dilution method revealed that a relatively low dose of nanomaterials, i.e. ranging from 0.78 to 0.625 ng mL −1 should be considered as fungicidal and bactericidal, as demonstrated for C. albicans, E. coli and S. aureus isolates. Importantly, the minimum biofilm inhibitory concentrations were not significantly higher than the bactericidal ones; for the majority of strains the MBIC value was not greater than 0.625 ng mL −1 . Notably, P. aeruginosa isolates were considerably less susceptible to AuNPsmediated treatment when compared to other strains, since MIC/MBC/MBIC values recorded for 5 representative isolates were ranging from 20 to > 40 ng mL −1 . It is also consistent with the results of colony counting assay (Fig. 2), presenting lower killing efficiency of non-spherical gold nanoparticles against this species of bacteria. Regardless of this, activity of gold nanorods, nanopeanuts, nanostars and porous gold nanoseeds were recorded to be far more greater than those for spherical-shaped nanoparticles. AuSph NPs and AuSph (CTAB) NPs were ineffective at doses lower than 20 µg mL −1 (Table 1).
Antimicrobial mechanism of developed nanoparticles includes the induction of oxidative stress followed by destruction of pathogen's membranes and lysis of the cell. The integrity of microbial layers following treatment with the synthesized nanoparticles was investigated using two fluorimetric probes-highly hydrophobic NPN (N-phenyl-1-napthylamine) and cell-impermeable PI (propidium iodide). Increase of NPN fluorescence intensity is observed when probe enters the periplasmic space between outer and inner membrane of Gram-negative bacteria, as well as space between fungal cell wall and cell membrane membrane 29,30 . The ability of non-spherical gold nanoparticles to permeate the inner membranes of fungi and bacteria was investigated by analyzing the fluorescence signal recorded when PI binds to DNA of treated microbial cells. Developed nanoparticles cause significant permeability increase in the cell membranes membranes of all tested fungal and bacteria strains, which is evidenced by dose-dependent rise of NPN fluorescence intensity (increase by two-to seven-fold; Fig. 3A-C) and nearly threefold enhancement of PI-derived fluorescence signal when compared with the untreated control microbes (Fig. 3E-H). In contrast to that, spherical-shaped gold nanoparticles, both unfunctionalized (AuSph NPs) and CTAB-functionalized [AuSph (CTAB) NPs] did not affect Table 1. Minimal inhibitory concentrations (MIC; ng mL −1 or µg mL −1 ), minimal bactericidal/fungicidal concentrations (MBC/MFC; ng mL −1 or µg mL −1 ) and minimal biofilm inhibitory concentrations (MBIC; ng mL −1 or µg mL −1 ) of varied-shaped gold nanoparticles against C. albicans, E. coli, P. aeruginosa and S. aureus isolates. www.nature.com/scientificreports/  www.nature.com/scientificreports/ the membrane architecture of tested microorganisms, which is correlated with their poor killing efficiency. The above observation was confirmed by fluorescence microscopy using SYTO9/propidium iodide (PI)-dual staining, where microbes with intact membranes were stained green and those with impaired ones were indicated by red color. As presented in Fig. 3D, the SYTO9-positive populations of S. aureus decreased with increasing concentration of AuR NPs, AuP NPs, AuS NPs and AuSph (70C) NPs, while the population of dead bacteria rose. As expected, viable population of bacteria treated with AuSph NPs at microgram doses did not differ, when compared to control. Similar tendency was observed also for representative strains of E. coli, P. aeruginosa and C. albicans while analyzing all developed nanoparticles ( Supplementary Fig. 3). These results were also confirmed by AFM analysis, which revealed considerable alterations in the morphology of E. coli and P. aeruginosa bacteria treated with non-spherical nanoparticles. As demonstrated in Fig. 3I, upon exposure to gold nanorods, wrinkles on the surface of the treated bacteria were observed, which was in contrast to the control cells that were characterized by a smooth cell surface. In addition, the decrease in height of the tested samples was recorded, which indicates the decomposition of the cellular structure of microbe. These results confirm our hypothesis that the gold nanoparticles-mediated killing process is determined by membrane-permeabilizing properties. Notably, disruption of both the outer and inner membranes was detected in all tested microbial strains, regardless of the species and the mechanism of drug resistance detected for these isolates, which suggest that developed nanomaterials might be effective against broad spectrum of drug-resistant pathogens.
To investigate whether the killing of fungi and bacteria by the nanoparticles developed in this study is dependent on oxidative damage-associated pathways, ROS generation upon AuNPs treatment was measured using 2,7-dichlorofluorescin-diacetate (DCFH-DA) as a fluorescent indicator of intracellular radicals. As shown in Fig. 4A, when E. coli were cultivated without test agents, only a residual fluorescent signal was detected, while a strong signal derived from DFCH-DA fluorescent probe was measured after treatment with increasing concentrations of non-spherical nanoparticles and those porous ones. No considerable effect was observed when bacteria were treated with AuSph NPs and AuSph (70C) NPs. Since analogous effect was observed for other pathogens as well ( Supplementary Fig. 4), we hypothesized that gold nanoparticle-induced ROS generation is responsible for their antimicrobial effect. Quantitative analysis have shown that treatment with non-spherical gold nanoparticles resulted in rapid augmentation of ROS levels compared with the untreated control, which indicates that the tested agents primarily kill microbes via ROS production ( Fig. 4B-E). Moreover, this free radical burst was strongly correlated with increases in membrane permeability and the antibacterial activity of the developed nanomaterials (Figs. 2, 3). This indicates that disrupting the membrane integrity through the generation of intracellular ROS is the most like mechanism for the antimicrobial activity of non-spherical gold nanoparticles and porous, spherical ones.
Moreover, to confirm the involvement of ROS in the bactericidal mechanism of non-spherical gold nanoparticles, the efficiency of AuR NPs against E. coli was evaluated in the presence of N-acetyl-cysteine (NAC). The data visualized in Fig. 4 show that the killing efficacy of the gold nanoparticles was significantly attenuated by the presence of antioxidant. Taken together, the collected data indicate that the superior antimicrobial activity of the developed nanoparticles is determined by ROS formation followed by membrane damage causing pathogens' death.
Developed nanoparticles are highly effective against uropathogenic E. coli and strongly limit the formation of biofilm on the surface of latex urinary catheters. Considering the potent antimicrobial activity of synthetized non-spherical gold nanoparticles presented in Figs. 2, 3, 4 and given our previous study, demonstrating anti-adhesive properties of gold nanorods 31 , we hypothesized that developed nanomaterials might be successfully used as biomedical devices coatings. Since tested nanoparticles were highly effective against representative E. coli isolates used in the previous stage of research, we decided to explore the potential of rod-and peanut-shaped gold nanoparticles as coatings for urinary catheters, which are generally made from materials that allow colonization of uropathogenic bacteria. For this purpose, we expanded our study on 10 clinical strains of E. coli isolated from the patients with urinary tract infections with internal numbers from 164/2 to 837. Strains used in this part of the study were isolated from subjects aged from 6 months to 84 years, 3 males and 7 females, with significant bacteriuria ranging from 10 4 to > 10 7 CFU/mL (Supplementary Table 1).
The antibiotics sensitivity of the tested strains was variable (Supplementary Table 2), showing a lower drug resistance ratio for children than adults, for which drug resistance was noted for at least 4 conventional antibiotics. More importantly, 6 and 7 of the 10 E. coli isolates were resistant to two antibiotics commonly used to treat UTIs-ciprofloxacin and trimethoprim/sulfamethoxazole, respectively. Furthermore, two of these isolates (E. coli 369 and E. coli 419) were also classified as ESBL (extended-spectrum β-lactamase) positive. Potent therapeutic efficiency was noted for gentamicin (9 strains sensitive, 1 strain intermediate), amikacin (10 strains sensitive), and nitrofurantoin (9 strains sensitive, 1 strain resistant). Based on these findings, the E. coli strains were divided into ESBL-positive, ciprofloxacin-resistant, and ciprofloxacin-sensitive groups for the testing and analysis in this research.
The bactericidal activity of rod-and peanut-shaped gold nanoparticles against isolates of E. coli was determined using a colony counting assay. As presented in Supplementary Fig. 1, both AuP NPs and AuR NPs exhibited potent bactericidal effects at low nanoparticle concentrations i.e. 0.125-0.5 ng mL −1 , and gold nanoparticles at a dose of 1 ng mL −1 were recognized as bactericidal for nearly all of the tested strains (Supplementary Fig. 1A,F). For the lowest tested dose, 0.125 ng mL −1 , the average survival rates over all of the tested stains were 42.68 ± 23.20% and 36.08 ± 20.82% for AuP NPs and AuR NPs, respectively, indicating their killing ability even at low doses ( Supplementary Fig. 1G). Notably, no considerable differences in bactericidal efficiency were noted between ESBL-positive, ciprofloxacin-resistant, and ciprofloxacin-sensitive strains. Statistical analysis of the differences in the killing abilities of AuP NPs and AuR NPs performed by comparing 30  www.nature.com/scientificreports/   1H). Both rod-and peanut-shaped gold nanoparticles exerted also high membrane activity as evidenced by β-galactosidase release from the treated bacteria cells. When the bacterial membrane is intact, β-galactosidase is localized intracellularly, however when the inner bacterial membrane is compromised β-galactosidase leaks out, allowing enzymatic hydrolysis of 2-nitrophenyl β-d-galactopyranoside (ONPG) to o-nitrophenol (ONP), which can be spectrophotometrically recorded at 420 nm. As demonstrated in Supplementary Fig. 2, nearly fourfold enhancement of β-galactosidase activity when compared with the untreated control bacteria was recorded. These data confirm the previous results demonstrating potent killing activity of non-spherical gold nanoparticles.
Using a simplified method to preliminarily quantify the viability of bacteria on the surface of the urinary catheters, we demonstrated the ability of nanorods and nanopeanuts to limit bacterial attachment to a material that is typically quite easily colonized by uropathogenic bacteria. A summary of the test is presented in Fig. 5A. The tested nanomaterials were highly effective for inhibiting bacteria attachment, proliferation, and survival on the surface of catheters. Compared with untreated latex catheters, biofilm formation by E. coli 369 was limited to 9.87 ± 5.57% at a AuR NP concentration of 0.5 ng mL −1 ; and at a dose of 25 ng mL −1 no bacterial growth was detected. In addition, the E. coli 419 strain was completely inhibited and eradicated even at the lowest dose of material. The lowest dose of AuR NPs was also able to disrupt biofilms pre-formed on the surface of urinary catheter. At a dose of 0.5 ng mL −1 the biofilms formed by E. coli 369 and E. coli 419 were disrupted to a value of 5.33 ± 4.71% and 3.76 ± 5.32%, which further indicated the high potential of non-spherical gold nanoparticles for use in coatings for medical devices, including urinary catheters (Fig. 5B). www.nature.com/scientificreports/ Rod-and peanut-shaped gold nanoparticles exert high biocompatibility at bactericidal doses. Above assumptions on potential utility of AuR NPs and AuP NPs as materials coatings were further supported by the satisfactory biocompatibility of the nanoagents in the bactericidal concentration range. Owing to the high membrane-permeabilizing properties of the non-spherical gold nanoparticles, a biocompatibility test based on hemoglobin release from damaged erythrocytes (hemolysis assay) was performed. As presented in Fig. 5C, damage of erythrocytes membranes was recorded when a dose of AuP NPs or AuR NPs of at least of 25 ng mL −1 was applied, which provides indication that in bactericidal doses (i.e. ~ 1 ng mL −1 ), the synthesized nanoagents should be recognized as safe for human cells.
Both AuP NPs and AuR NPs maintain their killing activity in the presence of urine and variable pH. For a material to be designed as a coating for biomedical devices, it is necessary to ensure that it will be effective in the physiological settings in which the device will be expected to perform. Since we propose that the rod-and peanut-shaped gold nanoparticles will be suitable as urinary catheter coatings, we made an attempt to verify that nanoparticles do not lose their activity in the presence of urine and variable pH. Notably, composition of urine and variable urinary pH affect the growth of uropathogens, and considerably alter the antimicrobial activity of some conventional antibiotics 32,33 . To test whether changes in urinary pH affect the killing abilities of the developed nanoparticles, minimum bactericidal doses were measured in growth medium adjusted to pH 5, 7, and 9 (Table 2 and Supplementary Table 3). As shown, at pH 5 the gold nanoparticles maintained their killing activity in 40% of the bacterial strains. The MBC values increased for 5/10 strains, nevertheless, they did not exceed 0.625 ng mL −1 . In alkaline environments a considerable decrease in the MBC values from 0.32 to 0.08 ng mL −1 and thus an improvement of the therapeutic efficiency was noted, suggesting the synthesized nanoparticles have potential not only against E. coli, but also urease-producing bacteria such as P. aeruginosa and P. mirabilis 34 . This behavior was very similar to that of ciprofloxacin, but more favorable than those of gentamicin, tobramycin, and doxycycline. Additionally, in order to introduce the developed nanoparticles into clinical settings, an appropriate antimicrobial efficacy in the presence of human urine must be achieved. The complex composition of human urine, which includes a variety of organic and inorganic ions, not only supports bacteria growth, but can also affect the solubility and stability of antimicrobials. To determine the impact of human urine on the killing properties of AuP NPs and AuR NPs, a fluorimetric-based viability assay was performed and the proliferation of bacterial strains in the presence of LB or LB supplemented with 50% urine was compared. As shown in Fig. 6A-F, some increases in proliferation were noted for bacteria cultured in the presence of human urine; the metabolic activity and survival of E. coli bacteria treated with the lowest dose of AuP NPs and AuR NPs, i.e. 0.5 ng mL −1 , were approximately 1.75-fold higher than those of bacteria cultivated in growth medium only. Significant differences were also noted for samples treated with 5 ng mL −1 of nanoparticles (increases from 13.65 ± 12.56% and 14.68 ± 13.06% to 42.10 ± 15.84% and 46.38 ± 14.86% for AuP NPs and AuR NPs, respectively). Under the same experimental conditions, nanoparticles at a dose of 25 ng mL −1 exerted bactericidal effects regardless of the addition of human urine (Fig. 6G-H). The killing activity of peanut-shaped gold nanoparticles appeared to be less affected by the addition of urine than rod-shaped particles, however no statistical differences between AuP NPs and AuR NPs were noted (Fig. 6I).
AuP NPs and AuR NPs exert potent anti-biofilm activity. Biofilms are communities of surfaceattached microorganisms embedded in a polymeric matrix consisting of self-produced substances, which protect the bacteria against chemical and mechanical stress factors and increase their resistance to applied antibiotics 35 . www.nature.com/scientificreports/ The production of biofilms is recognized as a crucial determinant of long-term persistence of bacteria in the urinary tract and a factor determining recurrence of urinary tract infections 36 . In effect, targeting biofilms might offer an opportunity to control biofilm infections 35,37 . In this study, to assess the anti-biofilm properties of the nanomaterials, we used resazurin biofilm staining, which is based on the ability of metabolic enzymes in the cytoplasm of viable bacteria to convert rezasurin into highly fluorescent resofurin 38 . The rezasurin viability assay is considered more accurate than conventional biofilm-counting methods 39 . Two approaches were used to investigate the anti-biofilm properties of the developed nanoparticles: in the first, the preventive properties against  www.nature.com/scientificreports/ E. coli-formed biofilms were assessed (Figs. 7A-F, 8A-F), in the second, the ability of AuP NPs and AuR NPs to disrupt pre-formed biofilms were determined (Figs. 7G-L, 8G-L). The peanut-and rod-shaped nanoparticles were highly effective in inhibiting biofilm formation. At a dose of 5 ng mL −1 no viable biofilm-embedded cells were detected in any of the tested isolates. When averaged, AuP NPs and AuR NPs were far more effective than the conventional antibiotics ciprofloxacin and gentamicin, even when the antibiotics were used at 1000 times higher doses (i.e. 0.125-25 µg mL −1 ). When the survival points for all of the tested doses were compared, the biofilm formation by ESBL-positive, ciprofloxacin-resistant, and ciprofloxacin-sensitive strains was inhibited to 43.55%, 44.18%, and 32.07% (for peanut-shaped gold nanoparticles) and 42.08%, 46.71%, and 49.82% (for rod-shaped nanomaterials), respectively. In addition, ciprofloxacin and gentamicin were much less effective as 1000-fold doses were required to achieve poorer results (Figs. 7D-F, 8D-F). Notably, further anti-biofilm performance was observed when the synthesized nanoparticles were tested as biofilm disrupting agents. Compared with the untreated pre-formed E. coli biofilms, AuP NPs and AuR NPs were able to significantly disrupt the tested biofilms even at low nanomaterial doses. The disruption ability of the developed nanoparticles at a dose of 0.125 ng mL −1 ranged from 35.52 ± 3.83% to 96.11 ± 1.04% (71.07 ± 20.17% on average) and from 34.79 ± 0.68% to 111.15 ± 33.26% (68.52 ± 32.36% on average) survival for AuP NPs and AuR NPs, respectively. Doses of 5-10 ng mL −1 were found to completely disrupt pre-formed established biofilms of E. coli (Figs. 7G-I, 8G-I). Ciprofloxacin and gentamicin were less effective against all of the tested strains, even when used at greater than 1000-fold higher concentration (Figs. 7J-L, 8J-L). Notably, the developed nanoparticles were more effective at disrupting the established bacterial biofilms than preventing the formation of new ones, which suggests that AuP NPs and AuR NPs possess some ability to interact with/and penetrate the extracellular matrix in which the bacteria are embedded, and breaking down the biofilm matrix makes them more prone to nanoparticle-mediated treatment.

Discussion
The antimicrobial effects of metal nanoparticles are a result of their morphology and surface charge, which are the critical factors that affect the interaction of NPs with microorganisms [40][41][42] . AuNPs have therefore been synthesized in a broad spectrum of sizes and shapes, including spherical nanoparticles 22,43 , nanorods 22 , nanocubes 26 , nanostars, and nanoflowers 25 . In our study, through the addition of CTAB-a cationic agent commonly used in the synthesis of varied-shape nanomaterials 44 -and by varying the synthesis time, we developed a series of varied-shaped gold nanoparticles in shape of nanorods (AuR NPs), nanopeanuts (AuP NPs) and nanostars (AuS NPs), as well as spherical-like porous nanoparticles [AuSph (70C) NPs]. For the purpose of additional controls, unfunctionalized gold nanoseeds (AuSph NPs) and CTAB-functionalized gold nanoparticles [AuSph (CTAB) NPs] were prepared. Modifications of synthesis methods allowed us to obtain nanomaterials, which are characterized by potent fungicidal and bactericidal activity against fungal strains, Gram-negative and Gram-positive bacteria at doses unreported before as low 0.5-1 ng mL −1 (Fig. 2, Table 1). In the majority of previous studies, the killing activity of gold nanoparticles was achieved in concentrations ranging from 5 to 512 µg mL −122,25,43 , i.e., thousands-fold higher doses than those reported in this study. Two potential explanations for the superior bactericidal activity are: (i) "a side effect" of CTAB, used as a shape-directing agent during the synthesis of the nanoparticles or (ii) a unique physicochemical feature of the nanoparticles themselves. Some reports suggest that CTAB might determine the bactericidal efficiency of nanorods against Staphylococcus aureus and Cutibacterium acnes and induce a superoxide stress in E. coli 45,46 . In this context, the possibility that gold nanoparticles could serve as a drug delivery system for the CTAB could be considered, however, the experimental data obtained exclude this possibility. In our experimental protocol we washed the CTAB out of the AuNP reaction mixture after finalizing the synthesis, and the amount of CTAB left on the surface of the nanorods and nanopeanuts was determined by FT-Raman spectroscopy ( Supplementary Fig. 5). The calculated percentage of CTAB remaining in the nanoparticle solutions after rinsing was 6% and 10% for the AuP NP and AuR NP solutions, respectively ( Supplementary Fig. 3). Since the bactericidal doses of the tested nanoparticles were below 1 ng mL −1 , it can be concluded that the concentration of CTAB would be as low as 60-100 pg mL −1 , which, to our knowledge, is unlikely to exert such potent killing behavior.
This conclusion was supported by the results of additional experiments intended to estimate the bactericidal activity of CTAB against the tested fungal and bacterial strains. The MIC/MBC values of CTAB were much higher than those reported for the developed nanoparticles (> 32 µg mL −1 vs. 0.078-40 ng mL −1 for CTAB and AuNPs nanoparticles, respectively; data not shown). More importantly, we did not observe any considerable effect of CTAB-functionalized spherical gold nanoparticles [AuSph (CTAB) NPs], which confirms that the presence or CTAB on the surface of gold nanoparticles is not a main determinant of their killing ability. Likewise, the observation of Castillo-Martínez et al. that gold nanorods fabricated in the presence CTAB are bactericidal for E. coli at a concentration of ~ 7 µg mL −1 , does not indicate that CTAB markedly improved the bactericidal activity  www.nature.com/scientificreports/ of the final product in this study 47 . In contrast, the second hypothesis linking the superior bactericidal effect to morphological futures of the AuNPs is supported by a compelling number of reports. Non-spherical nanomaterials that feature sharp edges or protrusions with a high aspect ratio may cause high local stress on the bacterial cells causing membrane rupture 48,49 . Similarly, the increased surface area to volume ratio of non-spherical nanoparticles and their enhanced adsorption and binding of compounds, as well as increased ROS generation are considered crucial factors for their biological activity 22,26,41 . For instance, the abrasiveness of the surface of zinc oxide nanoparticles was demonstrated to determine their enhanced antibacterial effect 50 . Likewise, better antibacterial activity has been reported for graphene nanowalls compared with graphene nanosheets owing to the sharper edges of the nanosheets, which considerably compromised cellular membrane and increased membrane permeability and leakage of intracellular material 48 . Therefore, we suggest that a similar type of shape-dependent interaction resulting in an increased ability of nanoparticles to interact with the microorganism is the basis for the superior bactericidal activity of the nanoparticles developed in this research. Since we do not observe a significant differences between killing efficiency of non-spherical gold nanoparticles, we believe that inducting of local stress on bacteria membrane and increasing the surface of contact between bacteria and nanoparticle is far more important than developing a specific shape of nanoparticle by itself. However, it is possible that shape and size of developed nanoparticles will affect the behaviors other than killing bacteria, such as activity in highly viscous body fluids (e.g. pus) or distribution in biofilms. This issue requires however more experimental work. Previous studies have reported the mechanisms by which gold nanoparticles exert bactericidal effects against bacteria 23,51 . For instance, Ortiz-Benitez et al. demonstrated that gold nanoparticles promote lysis of Streptococcus pneumoniae as a result of recruitment of lipids, proteins, and carbohydrates, which leads to pore formation and subsequent cell damage 24 . Furthermore, Cui et al. demonstrated that AuNPs in E. coli induce membrane potential collapse, and inhibit ATPase activity, or in another reported pathway, inhibit a subunit of ribosome from binding tRNA 51 . In contrast, Lee et al. revealed that the mechanism of AuNP activity against E. coli involves DNA damage through apoptosis-like pathways 23 . In our study, we demonstrated that non-spherical gold nanoparticles exert bactericidal effects through formation of reactive oxygen species followed by permeabilization of the outer and inner bacterial membranes, which ultimately led to bacterial death (Figs. 3, 4, Supplementary Fig. 2). Interestingly, in previously published reports the mechanism of AuNP-mediated killing was independent of reactive oxygen species formation, which is in contrast to data acquired in this research 52 . Nevertheless, it should be noted that in both of the reports referenced spherical gold nanoparticles were used, in contrast to the non-spherical ones used in this research. Although the effect of residual CTAB on the surface of the developed nanoparticles cannot be excluded entirely 46 , it should be noted that a number of studies indicate ROS-promoting killing mechanisms for gold nanoparticles against microbial pathogens, including methicillin-resistant Staphylococcus aureus 53,54 and Corynebacterium pseudotuberculosis 55 . Notably, in research by Nasser et al., positively-charged gold nanorods were more efficient in ROS triggering than spherical nanogold. It was demonstrated that the increased surface area and higher surface area to charge ratio of the nanorods promoted further ROS production 56 . We believe that a similar phenomenon occurs under our experimental conditions and the ROS-promoting properties are determined by the non-spherical shape of the nanomaterials, which contributes to increased local stress on the bacterial wall followed by ROS production and bacteria killing.
A potential application of the nanoparticles developed in this study is as antimicrobial coatings for urinary catheters, which are generally made from materials that allow colonization by fungi and uropathogenic bacteria.
Although some examples of silver NP-containing antimicrobial catheters have been demonstrated to date, there remains a need to improve their effectiveness. Only bactericidal, biocompatible, and anti-adherence agents are therefore considered as coating materials for medical devices. Despite numerous agents currently being tested as antimicrobial coatings 57 , the data on gold-coated urinary catheters remains very limited and gold is not utilized in a single medical device coating. The main factors contributing to the persistence and recurrence of urinary tract infections include (i) drug resistance, (ii) biofilm formation, (iii) chronic inflammatory states, and (iv) attachment of bacteria to medical devices. Based on the presented data, we believe that employment of non-spherical gold nanoparticles might successfully address these issues. Firstly, they are characterized by potent antibacterial activity against broad spectrum of pathogens, including E. coli bacteria (Fig. 2, Supplementary Fig. 1), which is recognized as a primary cause of catheter-associated urinary tract infections in hospital and in outpatients 58 . Crucially, this effect was maintained in the presence of urine (Fig. 6). Secondly, the synthesized nanorods and nanopeanuts exhibited anti-biofilm properties, both inhibiting biofilm formation and killing bacteria in already developed biofilms (Figs. 7, 8), which is in agreement with previously published reports 59,60 . Despite these promising data there is however a need for more detailed analyses in which urinary catheters are coated with nanogold; or AuNPs are incorporated into a polymeric catheter material. The potential of the tested nanomaterial as a medical device surface coating indicated by the data is preliminary and certainly has some limitations. However, the shape-controlled activity of gold nanoparticles should be subjected to further analysis for the development of new coating strategies that could considerably contribute to the reduction of hospital-acquired infections.
In summary, the novelty of the presented research lies in a modulation of the synthesis of gold nanoparticles that resulted in the formation of a series of non-spherical gold nanoparticles with substantially reduced bactericidal concentrations, i.e. from µg mL −1 to ng mL −1 , which has not been reported for nanoparticles synthesized by previous methods. The newly synthesized gold nanoparticles therefore provide a unique opportunity to develop more efficient antimicrobial coatings for medical devices and support the treatment of ever-growing antibiotic-resistant urinary tract infections. Notably, the AuP NPs and AuR NPs were effective against all E. coli strains, regardless of the identified mechanism of drug resistance. According to previously published data 61 , it is possible that the interaction of nanoparticles with the surface of bacteria is purely physically driven. We suggest that there is an electrostatically-governed association of CTAB-stabilized nanoparticles with the lipopolysaccharides of E. coli outer membranes, and this binding initiates the observed killing activity of AuNPs. Assuming that the majority of conventional antibiotics interact with specific targets in the bacterial cell, and bacteria can www.nature.com/scientificreports/ develop resistance through modification, the non-specific mechanism of the gold nanoparticle action would be favorable for preventing drug resistance. For star AuS NPs the reaction was stopped after 30 min, however, not 70 μL, but 210 μL of 7.8 × 10 −2 м C 6 H 8 O 6 , was added, while AuSph (70C) NPs was synthesized in the 70 °C. Furthermore, spherical AuNPs was obtained by two different synthesis methods. As one of the spherical Au NPs was used gold seeds obtained in the first step of prepared fancy shape Au NPs (AuSph (CTAB) NPs), while second spherical Au NPs (AuSph NPs) was synthesized as following: HAuCl 4 with concentration c = 1 mм was used as source of gold ions and trisodium citrate with concentration c = 34 mм (Na 3 C 6 H 5 O 7 ) was applied as a reduction and stabilizing agent. Trisodium citrate (980 µL) was heated until boiling on a magnetic stirrer, then the HAuCl 4 (20 µL) was added. Heating was continued till the color of the solution changed from colorless to pink/red.

Methods
Physicochemical characterization of spherical, rod-, star-, porous and peanut-shaped gold nanoparticles. Scanning transmission electron microscopy (STEM) using a high-angle annular dark-field detector (HAADF) was employed to examine the morphology of the synthesized nanoparticles. Selected area electron diffraction (SAED) patterns were acquired in TEM mode. All measurements were made on an aberration-corrected FEI Titan electron microscope operating at 300 kV, equipped with a FEG cathode. The average size of the longitudinal and transverse axes was calculated from the TEM photos (100 nanoparticles per calculation). Furthermore, optical properties of AuR NPs and AuP NPs was obtained by Ultraviolet-visible spectroscopy (UV-Vis) using Lambda Bio20 instrument from Perkin Elmer. The resolution was chosen to be 1 nm and the scan speed was 240 nm min −1 . To collect data about the stability of developed nanoparticles, zeta potential values in a pH range from 3.5 to 12.5 were measured using Zetasizer Nano Series from Malvern Instruments. The microelectrophoretic method and Smoluchowski model were employed to determine the zeta potential distribution. Each value was obtained as an average of three subsequent runs of the instrument with at least 20 measurements.
Antibacterial testing. The bactericidal efficiency of the developed non-spherical nanoparticles was tested using a colony counting assay (killing assay) according to a previously published procedure 37,62 . Briefly, bacterial and fungal cells were grown to mid-log phase at 37 °C, brought to 10 5 CFU mL −1 with sterile phosphate-buffered saline (PBS) and incubated in the presence of developed nanoparticles at concentrations ranging from 0.125 to 1 ng mL −1 . After 1 h of incubation at 37 °C the plates were transferred to ice and the suspensions were diluted 10to 1000-fold in PBS. Then, 10 μL aliquots were spotted on Saboraud agar plates (Candida isolates), MacConkey agar plates (E. coli isolates) or LB agar plates (P. aeruginosa and S. aureus isolates) for overnight culture at 37 °C and the CFUs were determined. Microbial cell survival after exposure to the tested agent was expressed as log (CFU mL −1 ) or a percentage of the control. In another set of experiments, the killing efficiency of gold nanoparticles in the presence of highly nutritious growth medium was tested. The minimum inhibitory concentration (MIC), minimum bactericidal concentration (MBC), and minimum biofilm inhibitory concentration (MBIC) were determined. The final concentrations of the tested compounds ranged from 20 µg mL −1 to 39 ng mL −1 [for AuSph NPs and AuSph (CTAB) NPs] and from 20 to 0.039 ng mL −1 [for AuR NPs, AuP NPs, AuS NPs and AuSph (70C) NPs]. The MIC values were determined visually as the lowest concentration of tested agent that showed no microbe growth after 24-48 h of incubation. Fungicidal and bactericidal activity was determined by plating each sample (10 µL) on Saboraud agar plates (Candida isolates), MacConkey agar plates (E. coli isolates) or LB agar plates (P. aeruginosa and S. aureus isolates). The minimum biofilm inhibitory concentration was assessed using an MTT test based on the reduction of tetrazolium salts 63 . Membrane permeabilization assays. An N-phenyl-1-napthylamine (NPN) uptake assay was used to assess the outer layers permeability of fungi and Gram-negative bacteria treated with varied-shaped gold nano- www.nature.com/scientificreports/ particles. C. albicans, E. coli and P. aeruginosa cells were suspended in PBS (OD 600 ~ 0.5 for fungal and OD 600 ~ 0.1 for bacterial samples) and then nanoparticles were added at concentrations of 0.5, 5, and 25 µg mL −1 (for AuSph NPs) or 0.5, 5, and 25 ng mL −1 (for other nanoparticles). NPN was then added to give a final concentration of 0.5 mM, and the mixture was incubated for 5 min. The fluorescence increases due to entering NPN into periplasm space was measured (excitation/emission wavelengths of 348/408 nm) using a Labsystem Varioscan Lux microplate reader (Thermo Fisher Scientific, Waltham, USA). Microscopy visualization of the membrane integrity was carried out using a SYTO9-PI dual staining. Permeabilization of the inner membrane upon treatment with gold nanoparticles was investigated by measuring PI-derived fluorescence signal. The final cell suspension was adjusted to OD 600 ~ 0.5 or ~ 0.1 prior to addition of gold nanoparticles at doses of 0.5, 5, and 25 µg mL −1 (for AuSph NPs) or 0.5, 5, and 25 ng mL −1 (for other nanoparticles). The increase of fluorescence signal (excitation/ emission wavelengths of 535/617 nm) was determined after 1 h incubation. Investigation of potential of gold nanoparticles as antimicrobial catheters coatings. The ability of rod-shaped gold nanoparticles to limit bacterial attachment and biofilm viability on latex urinary catheters was investigated using a colony counting assay based on a previously published protocol with some modifications 65  To assess the preventive effect of AuP NPs on biofilm formation, nanoparticles with concentrations of 0.5, 5, and 25 ng mL −1 were added at day 0. The urine was then replaced once a day without the further addition of nanoparticles. To determine the ability of the nanoparticles to disrupt pre-formed biofilms, nanoparticles were added to wells containing latex urinary catheters on day 2 and left to incubate for a further 24 h. At the final time points, the catheter pieces were transferred to individual test tubes containing 0.5 mL of sterile PBS. The tubes were sonicated for 5 min in a bath sonicator to detach all biofilm-embedded bacteria and thoroughly mixed. After sonication, viable bacteria in PBS from each tube were enumerated after serial dilutions and plated on MacConkey agar plates, as previously described. The values obtained in control cells (without test agents) was taken as 100%.
Biocompatibility evaluation. Disruption of red blood cells (RBCs) membrane upon treatment with tested nanoparticles was employed as an indicator of toxicity of AuP NPs and AuR NPs against host cells. The hemolytic propertie of the examined nanoparticles was tested using RBCs isolated from the blood of healthy volunteers and suspended in sterile PBS to obtain a hematocrit of ~ 5%. Briefly, peanut-and rod-shaped nanoparticles in doses ranging from 0.125 to 50 ng mL −1 were incubated in 96-well plates with the RBC solution at 37 °C for 1 h. After incubation, intact erythrocytes were concentrated by centrifugation (2500 rpm, 10 min) and the supernatant was transferred to a new clear bottom 96-well plate. The release of hemoglobin from damaged RBCs was measured by monitoring the absorbance at 540 nm. Samples exposed to the test agents were compared with wells containing Triton-X100-treated RBCs (denoted 100% hemolysis).
Investigation of AuP NP and AuR NP killing activity in the presence of urine. In order for the synthesized nanoparticles to be used for the eradication of uropathogens and as coatings for urinary catheters, it is crucial that the appropriate killing efficacy can be achieved in the presence of urine. To assess whether the developed nanomaterials were able to maintain killing effectiveness in the presence of urine, a resazurin-based proliferation assay was performed 66  www.nature.com/scientificreports/ employed, the final concentrations of AuP NPs and AuR NPs were adjusted to 0.5, 5, and 25 ng mL −1 . Changes in fluorescence recorded using excitation/emission wavelengths of 520/590 nm were measured using a Labsystem Varioscan Lux microplate reader (Thermo Fisher Scientific, Waltham, USA).

Influence of pH on the therapeutic efficiency of the developed nanoparticles.
To investigate whether differences in the urinary pH influenced nanoparticle efficiency, microdilution assays were performed as described above, but with pH adjustment to pH = 5, 7, or 9 using 0.1 M HCl and 0.1 M NaOH. The pHdependent changes in antibacterial activity were also compared with those for clinically relevant antibiotics: ciprofloxacin, gentamicin, tobramycin, and doxycycline. The final concentrations of the tested compounds ranged from 10 to 0.019 ng mL −1 for AuP NPs and AuR NPs, and from 256 to 0.5 µg mL −1 for conventional antibiotics.
Anti-biofilm properties. The bactericidal efficiency of the peanut-and rod-shaped nanoparticles against E. coli cells embedded in biofilms was determined using resazurin-based fluorimetric staining, which allowed the quantification of viable biofilm cells grown in microtiter plates 39 . To evaluate the ability of AuP NPs and AuR NPs to prevent the formation of biofilms, E. coli bacteria suspended in LB (OD 600 ~ 0.1) were added to black bottom 96-well plates and cultivated for 48 h with increasing concentrations of AuP NPs and AuR NPs ranging from 0.125 to 25 ng mL −1 . To determine the ability of the nanoparticles to disrupt biofilms, agents were added after 48 h of biofilm cultivation and left for incubation for another 1 h. At the end time point, resazurin sodium salt was added to give a final concentration of 0.2 mg mL −1 . The fluorescence generated after 1 h of incubation at 37 °C was measured at excitation/emission wavelengths of 520/590 nm. The values obtained for control cell cultures (without tested agents) were taken as 100%. The anti-biofilm properties of the developed nanoparticles were compared with the efficiency of ciprofloxacin and gentamicin at concentrations ranging from 0.125 to 25 µg mL −1 .
Statistical analysis. The data provided are results from 3 independent experiments ± SD. The significance of differences was determined using the two-tailed Student's t-test. Statistical analyses were performed using OriginPro 2020 (OriginLab Corporation, Northampton, USA). p < 0.05 was considered to be statistically significant.
Ethics approval and consent to participate. This study was approved by the Bioethics Committee at

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).