Development of an Anti-Biofilm Screening Technique Leads to the Discovery of a Peptoid with Efficacy against Candida albicans

Abstract

Bacteria and fungi can secrete and reside within a complex polysaccharide matrix, forming a biofilm that protects these pathogens from the immune response and conventional antibiotics. Because many microbial pathogens grow within biofilms in clinical settings, there is a need for antimicrobial agents effective against biofilm-protected infections. We report the adaptation of a phenotypic high-throughput assay for discovering antimicrobial peptoids toward the screening of combinatorial libraries against established biofilms. This method, termed the Inverted Peptoid Library Agar Diffusion (iPLAD) assay, required optimization of growth media, reducing reagent, and fungal viability reporter. Once optimized, iPLAD was used to screen a combinatorial peptoid library against Candida albicans, a biofilm-forming fungal pathogen responsible for most hospital-acquired infections. This screening resulted in a lipopeptoid termed RMG9–11 with excellent activity against several species of Candida, including drug-resistant strains of C. albicans and the emerging and dangerous C. auris. Additionally, the cytotoxicity of RMG9–11 against several mammalian cell lines was minimal. This work provides a new method for the identification of compounds effective against biofilm-protected pathogens and demonstrates its utility by identifying a promising anti-Candida peptoid.

Graphical Abstract

Candida species are the leading cause of fungal hospital-acquired infections (HAIs) with over 7 million cases per year and a mortality rate of 33%.¹ C. albicans is commonly found on the body or in the mouth and generally does not result in health problems for the host unless there is a change in environment such as pH or nutrient availability.¹ Candidemia, which is a C. albicans infection of the blood stream, has a mortality rate of 40–60% in healthy individuals.² C. albicans can switch their morphology between budding yeast, pseudohyphae, or hyphae depending on the environment.³ In nutrient-rich environments, C. albicans are primarily budding yeasts, but as the environment becomes more unfavorable, they switch to a more robust hyphae morphology. In the hyphae morphology, C. albicans secrete polysaccharides and other biomolecules that form a 3D matrix termed a biofilm.⁴ Biofilms allow pathogens to grow on abiotic surfaces, such as counters, catheters, implanted prostheses,⁵ and ventilators.⁶ Biofilms comprise polysaccharides, proteins, extracellular DNA, and small molecules that are used for protection and quorum sensing (QS).⁷ The polysaccharide mixture makes it difficult for therapeutics to reach the pathogen, thus requiring alternative treatments in order to combat the infection. These biofilms can be composed of a single species, multiple species, or even pathogens from different kingdoms.⁷ Opportunistic bacteria can embed themselves into fungal biofilms to create a cross-kingdom biofilm. In these scenarios, the patient not only has to be treated for the initial infection but also requires multiple antibiotics to combat the more complex cross-kingdom infection. There have been documented cases of S. aureus being found in C. albicans biofilm where both species can release QS molecules to signal for either species to proliferate or slow down the production of compounds.⁸ Importantly, not all cross-kingdom biofilms are synergistic. A. baumannii will inhibit the growth of C. albicans biofilm during initial growth, but once the biofilm is established, C. albicans will inhibit the growth of A. baumannii.⁹ C. albicans can do this through the release of farnesol, a small molecule shown to inhibit A. baumannii growth. The development of an antimicrobial agent that not only targets planktonically growing microbes but also penetrates the pathogen’s biofilm would be very advantageous in treating HAIs.

Antimicrobial peptides (AMPs) have been shown to prevent biofilm formation and kill pathogens within an established biofilm.^10,11 The mechanism of action for these anti-biofilm peptides is generally membrane disruption, though interference in cell signaling systems or breakdown of the biofilm itself has also been observed.¹⁰ While being promising, AMPs unfortunately suffer from rapid proteolytic breakdown and short half-lives in vivo, limiting their ability to combat biofilm-embedded pathogens.^12,13 One alternative is antimicrobial peptoids, or N-substituted glycines. Peptoids maintain the amide backbone of peptides, but have the side chain shifted from the α-carbon to the amide nitrogen.¹⁴ This structural change improves the proteolytic stability of peptoids compared to peptides, with reported in vivo half-lives greater than 20 h.^15,16 Antimicrobial peptoids have been developed against bacteria, fungi, and parasites.^17,18 In an early study exploring the anti-biofilm properties of peptoids, Kapoor et al. demonstrated that a series of Peptoid 1 [(Nlys-Nspe-Nspe)₄] derivatives prevented P. aeruginosa biofilm formation and disrupted preexisting biofilm.¹⁹ Saporito et al. determined that a series of peptoids derived from the AMP GN-2 were less effective against planktonic cultures of E. coli but were more effective at eradicating established E. coli biofilms.²⁰ This illustrates that excellent antimicrobial activity of a compound against planktonic microbes may not translate into a desirable activity against microbes growing in a biofilm or vice versa. The first development of peptoids against fungal biofilms was reported by Luo et al.²¹ In this study, a series of 18 amphiphilic peptoids of varying lengths were evaluated against C. albicans, S. aureus, and E. coli biofilms, with longer peptoids tending to be the most effective. The most promising peptoids were also tested for activity against cross-kingdom polymicrobial biofilms, a clinically relevant scenario. These studies demonstrate the ability of antimicrobial peptoids to target microbes within biofilms. However, relying on peptoids with excellent activity against planktonic microbes to be effective against biofilms is speculative and a higher throughput technique is needed to identify peptoids with anti-biofilm activity.

We have previously reported the peptoid library agar diffusion (PLAD) assay to screen one-bead-on-compound libraries of peptoids quickly to identify compounds with antibacterial and antifungal properties.^22,23 This assay relies on a bead-immobilized chemical linker system that allows identical strands of a peptoid to be released in response to different chemical stimuli. The β-strand peptoid is released by a reducing reagent during screening to interact with the surrounding microbe, while the α-strand is released after screening by cyanogen bromide from beads displaying zones of inhibited growth. Under current conditions, the PLAD assay is setup to screen peptoid libraries against pathogens growing in colony form on solid media. The utility of the PLAD assay is that it can be adapted to a number of different microbial growth scenarios through modification of each aspect of the assay, including media, reducing reagent (to release β-strand peptoid), and hit reporter (to indicate inhibited microbial growth). Unfortunately, as developed, the PLAD assay cannot directly identify peptoids with anti-biofilm activity because the conditions of the assay do not cause microbes to form complex biofilms. Herein, we report the adaptation of the PLAD assay to screen combinatorial peptoid libraries against established C. albicans biofilms, termed the inverted PLAD (iPLAD) assay. This adaptation required optimization of assay setup, growth media, reducing reagent, and viability reporter. The iPLAD assay was subsequently used to screen combinatorial peptoid libraries against C. albicans biofilms to identify the lipopeptoid RMG9–11. RMG9–11 shows promise as a lead compound in the treatment of C. albicans infections with efficacy similar to some clinical anti-Candida compounds, pan-Candida activity, and good selectivity for C. albicans over mammalian cells.

RESULTS AND DISCUSSION

The first task in adapting the PLAD assay to screen against C. albicans biofilms was to optimize the conditions for forming a C. albicans biofilm. Initial efforts studied the effect of different growth media on fungal morphology. Unlike the PLAD assay, which primarily screens against C. albicans in a budding yeast morphology growing on top of agar-immobilized peptoid library, the goal of the iPLAD was to interrogate peptoids against C. albicans in a hyphal and biofilm morphology. Therefore, a biofilm would need to be formed first on an abiotic surface before adding in agar-immobilized peptoid library, which is why the assay is referred to as “inverted” compared to the original PLAD design. To test various growth media, a saline solution of C. albicans (OD₅₃₀ = 0.15–0.25) was used to inoculate yeast peptoid dextrose (YPD), Spider media, or RPMI-MOPS broth, which was then transferred to a tissue culture (TC)-treated Petri dish and incubated overnight before capturing images of fungal growth on a stereoscope (Figure 1). Important to note is the use of TC-treated Petri dishes because biofilms grown on nontreated Petri dishes were easily washed away. YPD is a commonly used medium for the growth of C. albicans, but its use here was discontinued due to overgrowth of C. albicans and no formation of hyphae (an indicator of biofilm development) during iPLAD optimization.²⁴ Spider is a growth medium that is designed to specifically cause C. albicans to switch morphology from budding yeast to hyphae.²⁵ This medium successfully initiated hyphae morphology and modest biofilm development (Figures 1B and S2); however, there was still a robust amount of budding yeast colony formation similar to the use of YPD media. RPMI-MOPS was investigated because this medium is most commonly used in C. albicans biofilm microtiter plate assays.²⁶ Overnight growth of C. albicans in RPMI-MOPS in a TC-treated Petri dish generated scattered colonies with vigorous hyphae development. Even though both Spider media and RPMI-MOPS formed hyphae and biofilm, RPMI-MOPS was considered ideal for the development of the iPLAD due to a higher prevalence of hyphae and biofilm compared to Spider media. Additionally, the growth generated by Spider media was too vigorous to properly assess growth inhibition by peptoid libraries.

Figure 1.

C. albicans growth in TC-treated Petri dishes in (A) YPD, (B) Spider media, or (C) RPMI-MOPS broth after 24-h incubation. Increased hyphae and biofilm formation are observed with RPMI-MOPS versus the other media tested.

To further investigate the optimal time to form a robust biofilm, a solution of C. albicans in RPMI-MOPS was incubated in a TC-treated Petri dish for 24, 48, or 72 h. Biofilms were washed gently three times with phosphate buffered saline (PBS) to remove any planktonic fungi. Interestingly, at 48 h, biofilm displayed decreased adherence and robustness and tore easily, becoming even more fragile at 72 h. Therefore, 24 h was determined as the optimal incubation time for robust, established biofilm development (Figure S3), which is consistent with previous reports of C. albicans biofilm development.^27,28 With proper conditions for forming robust biofilm determined for the C. albicans iPLAD assay, analysis of a proper reducing reagent was assessed.

During PLAD screening, a reducing reagent is used to release the β-strand peptoid from the bead. We have demonstrated previously that different microbes have differing tolerances to reducing reagents;^22,29 thus, a suitable reducing reagent for C. albicans was determined. To test the tolerance of C. albicans biofilm to various reducing reagents, robust biofilms were first formed in RPMI-MOPS in microtiter plates for 24 h. After washing to remove planktonic cells, RPMI-MOPS containing varying concentrations of the reducing reagents dithiothreitol (DTT), beta-mercaptoethanol (βME), or tris-(2-carboxyethyl)phosphine (TCEP) was added and incubated overnight. Biofilm density was determined by imaging on a stereoscope and analyzing the images for luminosity (Figures S4 and S5). Fungal viability within the biofilm was determined using PrestoBlue. Deleterious effects on biofilm density and fungal viability were seen with high concentrations of DTT (Figure S5). However, concentrations of βME and TCEP as high as 14 mM had no effect on biofilm integrity or C. albicans viability. Given the success of TCEP during PLAD screenings,^22,23 this reducing reagent was utilized for subsequent iPLAD screenings.

One of the most important parts of a high-throughput assay is the “reporter”, which indicates that some desired biological activity has occurred within the assay. For the PLAD and iPLAD assays, this biological activity is the killing of microbes. For the PLAD assay, inhibited microbial growth is determined by a visible zone of inhibited growth around beads releasing peptoids with antimicrobial activity.²² However, this is not possible with the iPLAD because the killing of microbes within the biofilm does not always equate to visible eradication of the biofilm. Therefore, the use of both fluorescent and colorimetric hit reporters to evaluate cell viability within established biofilms was explored. Throughout this section, the iPLAD assay was set up using the optimized growth media (RPMI-MOPS) and reducing reagent (TCEP; 14 mM) conditions. Amphotericin B (AmpB) control resin was synthesized, where AmpB was attached to the β-strand portion of the PLAD linker (Figure S1), allowing this potent antifungal to be released during the iPLAD assay, thereby generating zones of inhibited growth within the biofilm that could be used to interrogate the reporter. Given their sensitivity, fluorescent reagents were first tested by adding them to the soft agar with resin before pouring onto an established C. albicans biofilm in a TC-treated Petri dish. Both PrestoBlue, a viability reporter used to measure cellular metabolic activity,³⁰ and propidium iodide, a dye that fluoresces once passing through the compromised membrane of dead cells, were explored. Unfortunately, during fluorescent microscopy analysis, the background fluorescence of the resin beads was far more intense than the fluorescence produced by either metabolized PrestoBlue or intracellular propidium iodide (Figure S6). It was ultimately concluded that fluorescent washout from the resin rendered the use of fluorescent reagents in the iPLAD untenable and colorimetric reagents were next explored.

The first colorimetric reporter investigated was MTT, a standard reagent in detecting mammalian cell viability. MTT is a yellow pigment that is soluble in water; however, viable cells metabolize MTT to formazan, a purple, water-insoluble dye that stains cells purple. Our initial test concentration of MTT mimicked the standard concentration used in viability assays (0.5 mg/mL), followed by dilutions of 0.2, 0.1, and 0.05 mg/mL MTT. C. albicans biofilm was incubated with AmpB control resin in soft agar-containing TCEP and varying concentrations of MTT (Figure S7). At the highest concentration (0.5 mg/mL), there were clear, lightly colored zones around the resin. The resin had turned pink due to general dye absorption, and the rest of the biofilm had a purple-pink coloration, indicating healthy cells (Figure S7A). At 0.2 mg/mL MTT, a small, lighter-colored zone was observed around the resin followed by a dark pink halo (Figure S7B). At the two lowest concentrations, the rest of the biofilm did not exhibit a pink color and zones around the resin appeared dark due to biofilm eradication by the AmpB resin (Figure S7C–E). It was thought that the 0.2 mg/mL concentration could be adequate in the detection of hits during screening and proceeded to testing MTT as an iPLAD reporter in the screening of peptoid libraries.

Screening of AmpB control resin under these conditions again resulted in visible zones of biofilm eradication (Figure S8A). Screening of peptoid library RGL9, whose design is discussed below, produced an unusually high percentage of “hits,” that is beads with pink halos, when screened against C. albicans using the iPLAD set up with an MTT reporter (Figure S8B). As a negative control, resin with only the PLAD linker containing no peptoid was used in the iPLAD screening to determine whether the linker itself was interacting with MTT, giving false hits (Figure S8C). After incubation, the zones of lighter coloration and faint pink halos were observed around PLAD linker control resin, indicating that the pink halos were likely an artifact of the reporter and not an indication of biofilm inhibition. Therefore, MTT was abandoned as an iPLAD reporter. It was hypothesized that the reducing reagent used during iPLAD screening could interfere with reporters such as MTT and PrestoBlue that rely on redox chemistry to determine viability.

Phloxine B is a colorimetric reporter of C. albicans cell viability; however, unlike MTT, it is not a metabolic reporter and does not require reduction. Phloxine B passively diffuses into cells, and viable cells actively efflux the dye, preventing dye accumulation in live cells.³¹ Dead cells lack the means to secrete the dye and accumulate enough dye to be stained a red color that can be viewed by microscopy. We hypothesized that this detection method may be ideal for the iPLAD, in that it does not require cellular reduction or rely on fluorescence. The iPLAD assay was set up using AmpB control resin, TCEP, and phloxine B (10 μg/mL) under the hypothesis that AmpB released from the bead should result in accumulation of phloxine B, thereby staining the cells red. After incubation, AmpB resin produced eradicated biofilm around resin with no indication of red coloring (Figure S9). There was concern that because AmpB was potent enough to actually eradicate biofilm, then perhaps this control was not suitable for optimizing a reporter that relied on the presence of dead cells intact and embedded in a biofilm around beads. Therefore, a combinatorial peptoid library, RGL9, was used to explore the use of phloxine B as a reporter for the iPLAD assay. Unfortunately, phloxine B added to the soft agar prior to incubation gave no indication of cell viability, that is, no change in biofilm color. As an alternative to adding the reporter with the soft agar, a solution of phloxine B (10 μg/mL) was added on top of the solidified agar medium after overnight incubation with agar-immobilized peptoid library on top of biofilm. It was hypothesized that the phloxine B could diffuse through the agar and interact with cells in the biofilm, possibly giving a more accurate report of viability. After allowing phloxine B to diffuse for 1 h, several hits, having red zones around the beads where phloxine B was taken up by dead cells, were identified (Figures 2 and S10). The overall hit rate was reasonable and comparable to other PLAD screenings,^22,23 and the area around beads with no anti-biofilm activity remained white. It is important to maintain a relatively low concentration of phloxine B, as this compound can be toxic to C. albicans at higher concentrations.³²

Figure 2.

Representative image from the iPLAD screening of RGL9 against established C. albicans biofilm using phloxine B as a reporter. The “hit” (bead containing peptoid with anti-biofilm activity) can be seen as having a region of reddish cells around it.

The iPLAD assay can be understood in four stages (Figure 3). In the first stage, an overnight culture of C. albicans in RPMI-MOPS produces a mature biofilm on the surface of a TC-treated Petri dish. In the second stage, the medium is removed and the biofilm is washed with PBS before a mixture of PLAD-linked peptoid library beads and reducing reagent in soft agar media liquified at 42 °C is added on top of the biofilm and allowed to solidify. A minimal amount of agar is used to provide good contact between library beads and biofilm. Overnight incubation allows reducing reagent to release β-strand peptoid from library beads to interact with the biofilm. In the third stage, phloxine B is added on top of the soft agar and allowed to diffuse into the biofilm, revealing peptoids that have killed fungal cells within the biofilm by a pinkish ring around “hits.” It is possible that peptoids could inhibit the efflux of phloxine B without actually killing the fungi; however, given the generally accepted membrane disruption mechanism of action of antifungal peptoids, this is unlikely and such false positives would be ruled out in subsequent testing. Finally, in the fourth stage, hits are isolated and the structure of the α-strand is deconvoluted by tandem mass spectrometry (MS/MS).

Figure 3.

Diagram of individual stages for the Inverted Peptoid Library Agar Diffusion (iPLAD) assay. Stage 1: A mature C. albicans biofilm is formed on a TC-treated Petri dish. Stage 2: PLAD-linked peptoid library beads are embedded in soft agar containing a reducing agent on top of the established biofilm. Overnight incubation allows for the release of the β-strand peptoid from the bead to interact with fungi. Stage 3: The cell viability indicator phloxine B is added and diffuses through the agar. If the peptoid released from a bead kills the fungi, then those fungi are unable to pump out phloxine B, staining them a reddish color. Stage 4: Hit beads are removed, and the α-strand peptoid is cleaved for structure determination by tandem MS.

Once optimized, the iPLAD assay was used to screen combinatorial peptoid library RGL9 against C. albicans biofilms. Combinatorial libraries are advantageous as a source of lead molecules because a large collection of unique and diverse compounds can be synthesized and screened rapidly.³³ The design for peptoid library RGL9 has a general sequence of Ntri−N_R−N_R−N_C−N_R−N_R−N_C-linker where Ntri is a tridecylamine lipid tail, N_R are randomized aliphatic and aromatic monomers, and N_C are randomized cationic monomers (Figure 4A,B). Most AMPs are amphiphilic with an overall positive charge of two or more.³⁴ This is the basis of RGL9 construction, which includes two set positions for cationic monomers. The other randomized positions included aliphatic and aromatic monomers commonly observed in antimicrobial peptoids, such as the phenylethyl group, along with other lesser explored monomers of interest. Monomers that have not been used as commonly in antimicrobial peptoids, like naphthyl or nonaromatic cyclic monomers, were also added to explore greater chemical diversity in the library, a key advantage of screening combinatorial libraries. Addition of a tridecylamine tail at the N-terminus of combinatorial libraries to yield lipopeptoids is a strategy used to reduce library length and improve hit sequencing by MS/MS.^29,35 Through the addition of all these monomers, RGL9 had a theoretical diversity of roughly 1.6 × 10⁵ unique compounds. This theoretical diversity is limited by the amount of resin used during synthesis, which was 0.5 g, giving an actual diversity for RGL9 of approximately 3.2 × 10⁴ unique peptoids.

Figure 4.

(A) Structure of RGL9 on the solid-phase PLAD linker system. Positions C contain cationic monomers and positions R contain aromatic or aliphatic monomers. (B) Cationic, C, and aromatic/aliphatic, R, side chain groups incorporated into RGL9. (C) Structure of RMG9–11 identified from an iPLAD screening against C. albicans.

To screen RGL9 using the iPLAD assay, C. albicans mature biofilms were generated by inoculating RPMI-MOPS with a fungal cell solution and incubating in TC-treated Petri dishes overnight. Biofilms were gently washed with PBS to remove any planktonic or dead C. albicans cells. Liquified RPMI-MOPS soft agar with 14 mM TCEP and 2–5 mg aliquots of RGL9 resin equilibrated in PBS was added to washed biofilms, allowed to solidify, and incubated overnight. Unlike AmpB control resin, peptoids from RGL9 did not eradicate established biofilm. To determine cell viability and identify hits, phloxine B (10 μg/mL) was added on top of the soft agar and incubated for 1 h. Any beads with clearly observable red halos formed by dead cells unable to efflux phloxine B were considered “hits” and were manually removed, placed into individual tubes, and cleaned by boiling in 1% sodium dodecylsulfate to remove media and cellular debris. The α-strand peptoid was then cleaved from the bead using cyanogen bromide and analyzed by MS/MS to determine the unknown peptoid sequence.

In total, 17 hits were identified from screening ~3400 beads (0.5% hit rate). Due to sequencing challenges related to compound cleavage efficiency and instrument sensitivity, only one complete hit sequence was determined, leading to the discovery of lipopeptoid RMG9–11. Despite optimization efforts, deconvolution of the peptoid structure from a single bead by MS remains one of the most challenging aspects of combinatorial library screening. The structure of RMG9–11 (Figure 4C) contained many of the common elements found in AMPs. As has been observed in previous library screenings against C. albicans, the cationic guanidinium group was not prevalent in identified hits that were successfully sequenced, including RMG9–11 (data not shown). RMG9–11 contains cationic groups common among antimicrobial peptoids, such as Nlys and Nap.¹⁷ This peptoid also contains two residues found in a seminal, well-studied antimicrobial peptoid, termed Peptoid 1, which comprises a Nlys-Nspe-Nspe motif repeated four times.³⁶ Both Nlys and Nspe (labeled here as Npea) are seen in RMG9–11. While Peptoid 1 used an enantiomerically pure S phenylethylamine (Nspe) monomer, the phenylethylamine (Npea) used in the synthesis of RGL9 and ultimately identified in RMG9–11 was racemic. Derivatives of RMG9–11 containing enantiomerically pure R and S phenylethylamine were subsequently synthesized and tested but had no effect on potency, indicating that the stereochemistry of this monomer is not important for activity in this case. Another aromatic moiety found in RMG9–11, the furfuryl group, was previously identified in a peptoid effective against C. neoformans using the PLAD assay.²³ RMG9–11 is relatively short compared to Peptoid 1 or other antifungal peptoids,^21,36 primarily due to the pharmacological effect of the tridecyl lipid tail.

The minimum inhibitory concentration (MIC) of resynthesized RMG9–11 was initially determined against C. albicans, C. neoformans, and the ESKAPE bacteria following CSLI guidelines (Figure 5A).^37,38 The MIC for RMG9–11 against C. albicans was 6.25 μg/mL, which is comparable to other antifungal peptoids³⁵ and several clinical treatments for this fungal pathogen, including fluconazole and flucytosine.³⁹ As observed with previously discovered antifungal peptoids, RMG9–11 was more effective at killing C. neoformans than C. albicans,^16,40,41 although only at a twofold improvement from 6.25 to 3.13 μg/mL. The biofilm MIC (BMIC) of RMG9–11 increased 16-fold compared to the MIC (BMIC = 100 μg/mL; Figure 5B). BMIC is often elevated compared to MIC for a number of reasons, including a 100- to 1000-fold increase in cell number for the established biofilm. Other theories suggest that the negatively charged polysaccharide matrix of the biofilm can bind up cationic AMPs and peptoids, preventing their engagement of the pathogen target.¹⁰ The BMIC required to generate a hit during iPLAD screening is unknown, and determining the microenvironment concentration of peptoid released from a bead is challenging, factoring in release efficiency, diffusion, and peptoid affinity to the hydrophobic environment of the bead. A very rough calculation based on the radius of the zone of inhibition (0.3 mm) and the amount of compound theoretically released from one bead (6.4 nmol) would yield a millimolar microenvironmental concentration, which is concentrated enough to readily identify compounds with even modest anti-biofilm activity. The MIC and BMIC of clinical antifungals AmpB and fluconazole were also determined as a comparator. While AmpB is one of the most potent antifungals on the market, its efficacy against established C. albicans biofilms decreased 256-fold (Figure 5B). Similarly, while fluconazole was more effective than RMG9–11 against planktonic fungi (fluconazole MIC = 1 μg/mL), the BMIC was above the highest concentration tested, 512 μg/mL. Analysis of fungal inhibition within a biofilm at varying fold MIC of RMG9–11, AmpB, and fluconazole as measured by PrestoBlue indicated that RMG9–11 completely kills the fungi within the biofilm at 16× and 32× the MIC (Figure 5C). At these elevated fold MIC, AmpB and fluconazole show roughly 70 and 35% inhibition, respectively. Therefore, while the decreased efficacy of RMG9–11 against C. albicans biofilm seems disappointing, this peptoid reaches complete biofilm inhibition at fold-MIC values better than AmpB or fluconazole.

Figure 5.

(A) Minimum inhibitor concentration (MIC) values of RMG9–11 against fungal (blue) and bacterial (orange) pathogens. C. albicans clinical isolates include a reference strain (M1), fluconazole-susceptible strains (M4 and M6), and fluconazole-resistant strains (M2, M3, M5, and M7). ATCC 64124 is multidrug-resistant C. albicans which is resistant to AmpB, fluconazole, caspofungin, and flucytosine. (B) Comparison of MIC and BMIC values for RMG9–11 and clinical antifungals AmpB and fluconazole. MIC, minimum inhibitor concentration; BMIC, biofilm minimum inhibitor concentration; fold change, BMIC/MIC. (C) Percent inhibition of C. albicans grown in established biofilm at varying fold MIC for RMG9–11, AmpB, and fluconazole. ns, not significant compared to vehicle control (paired t-test p-value > 0.05, indicating no significant difference). *p-value ≤ 0.05. All data were collected in technical and biological triplicate.

RMG9–11 was evaluated against fluconazole-resistant and fluconazole-susceptible clinical strains of C. albicans generously provided by Dr. Mary Farone (MTSU Department of Biology) with MICs between 6.25–12.5 μg/mL regardless of fluconazole susceptibility (Figure 5A). Encouragingly, RMG9–11 maintained potency against ATCC 64124 (6.25 μg/mL), a C. albicans strain with AmpB, caspofungin, fluconazole, and flucytosine drug resistance. This led to investigation of RMG9–11 potency against C. auris. C. auris was identified in 2007 as a multidrug-resistant species of Candida and was observed first in Japan followed by cases in Africa and Europe.^42–45 Only a modest decrease in efficacy was seen for RMG9–11 with C. auris (12.5 μg/mL), which was encouraging, given that C. auris is resistant to most clinical antifungal agents.⁴⁶ Other common Candida species (tropicalis, krusei, glabrata, and parapsilosis) were evaluated to determine whether RMG9–11 could be used as a pan-Candida treatment. RMG9–11 appears promising, with MIC values ranging from 3.13 to 6.25 μg/mL against all Candida species tested, maintaining or even improving efficacy compared to C. albicans. This led to the evaluation of efficacy against bacterial pathogens, specifically the ESKAPE bacteria, to determine whether RMG9–11 could be useful on a broader scale, potentially against cross-kingdom infections. The ESKAPE bacteria consist of the most common multidrug-resistant and nosocomial bacterial pathogens, including both Gram-positive and Gram-negative species.⁴⁷ Encouragingly, RMG9–11 had moderate activity against both Gram-negative and Gram-positive bacteria (3.13–12.5 μg/mL; Figure 5A), with the exception of Gram-negative K. pneumoniae (50 μg/mL). The broad-spectrum antimicrobial activity of RMG9–11 naturally led to concerns about mammalian cytotoxicity, especially because the membranes of fungi and mammals have similar phospholipid compositions.

While the antimicrobial profile of RMG9–11 looked promising, a further look into mammalian cytotoxicity needed to be evaluated. Mammalian cell lines that were tested (HepG2 liver cells, 3T3 fibroblasts, and HaCat skin cells) were incubated with RMG9–11 for 72 h and cell viability was determined via mitochondrial activity using an MTT reduction assay where viable cells reduce MTT to formazan. Toxicity is reported as the concentration of peptoid, resulting in a 50% reduction in viable cells, termed toxicity dose 50% or TD₅₀. The cytotoxicity of RMG9–11 was first determined against HepG2 hepatocellular carcinoma cells with a TD₅₀ of 114 μg/mL and no toxicity at the MIC against C. albicans, giving a modest selectivity ratio (SR) of 18 (Table 1). SR is defined as the TD₅₀ divided by MIC and provides a measure of the therapeutic window for a compound. Typically for most lead compounds that would continue through development and optimization, an SR of 10 is preferred while an SR of 100 is more desirable for a compound moving into preclinical animal model evaluation. RMG9–11 displayed increased toxicity against mouse fibroblasts (3T3; TD₅₀ = 39 μg/mL) and keratinocytes (HaCat; TD₅₀ = 53 μg/mL) compared to HepG2 cells, though there was no measurable toxicity at the MIC. Hemolytic analysis showed RMG9–11 to be moderately toxic to human red blood cells (hRBCs) with an HC₁₀ of 29 μg/mL and an SR of 5. One hypothesis regarding the moderate toxicity is that it may result from the phenylethyl residue. Other compounds incorporating this peptoid monomer had a similar toxicity, including Peptoid 1 with an HC₁₀ of 16 μg/mL.⁴⁸ Studies are currently underway to improve the toxicity profile of RMG9–11 through an iterative structure–activity relationship study.

Table 1.

Cytotoxicity Analysis of RMG9-11 against Several Different Mammalian Cell Lines

cell Line	TD50 (μg/mL)	HC10 (μg/mL)	SR
HepG2	114 ± 5		18
3 T3	39 ± 2		6
HaCate	53 ± 5		8
hRBC		29 ± 2	5

Peptoid concentrations resulted in a 50% reduction in viable cells (TD₅₀; toxicity dose 50%) for liver (HepG2), fibroblast (3T3), and keratinocytes (HaCat). Peptoid concentrations resulted in 10% hemolysis (HC₁₀) of single donor hRBC. SRs were calculated as cytotoxicity divided by MIC against planktonic C. albicans. All data were collected in technical and biological triplicate.

In summary, the PLAD assay has been adapted to screen for peptoids with activity against C. albicans in established biofilms. Treatment of pathogens within a biofilm matrix is notoriously difficult because the biofilm protects microbes from most conventional antibiotics. The optimization of growth media, reducing reagent, and viability reporter was important to the success of this assay. With C. albicans, RPMI-MOPS was proven to be the media that sufficiently caused the pathogen to switch morphology to hyphae and created robust biofilms. TCEP as the reducing reagent for the release of the β-strand from the PLAD linker proved to have no effect on C. albicans viability and needed no further optimization for the iPLAD assay. The biggest challenge was in determining a reliable hit reporter. Fluorescent reporters, PrestoBlue, and propidium iodide were first explored; however, resin fluorescence rendered fluorescent reporters untenable. The colorimetric reporter MTT was also explored and showed early promise. However, this promise proved to be false positives, and MTT was abandoned due to interference between MTT and the reducing reagent TCEP. Lastly, the colorimetric reporter phloxine B was tested. While adding phloxine B within the soft agar was not successful, the addition of a solution of phloxine B on top of the agar after incubation clearly identified compounds that inhibit C. albicans growth while embedded in biofilms. Identifying a suitable hit reporter will likely remain the biggest challenge in the adaptation of this assay toward screening combinatorial peptoid libraries against other biofilms formed by fungal, bacterial, or cross-kingdom colonizations. However, if the iPLAD assay can be adapted to screen libraries against cross-kingdom biofilms, it will hopefully result in compounds that have both antifungal and antibacterial activity, resulting in a single treatment for both pathogens within a complex biofilm scenario. This would greatly relieve the stress on patients who need intensive treatment for cross-kingdom biofilm infections. The iPLAD assay was ultimately used to screen peptoid library RGL9 against C. albicans and identify RMG9–11, which displays pan-Candida activity, including modest activity against notoriously dangerous and resistant C. auris. While the antifungal activity makes RMG9–11 an attractive lead compound, the moderate mammalian cytotoxicity merits continued improvement of this peptoid. This can hopefully be achieved through derivatization and/or removal of toxic residues through an iterative structure–activity relationship study.

METHODS

Materials.

Reagents were purchased from Fisher Scientific (Waltham, MA), Alfa Aesar (Haverhill, MA), Amresco (Solon, OH), TCI America (Portland, OR), Anaspec (Fremont, CA), EMD Millipore (Billerica, MA), Peptides International (Louisville, KY), and Chem-Implex (Wood Dale, IL). All reagents used were of greater than 95% purity. hRBCs were acquired from Innovative Research (Novi, MI). Boc- and Mmt-protected diamines were purchased from Chem-Impex (Wood Dale, IL) or synthesized as previously described.⁴⁹ Microscopic images were captured using a Leica M165FC microscope. All mass spectra were acquired on a Waters Synapt HDMS QToF with Ion Mobility, and all NMR spectra were acquired on a JOEL ECA 500 NMR spectrometer. All fluorescence and absorbance readings were acquired on a Spectramax M5 plate reader. Purification of compound was achieved by Varian Prepstar SD-1 with Supelco Ascentis C18 column (5 μM; 25 cm × 21.2 mm; Sigma-Aldrich 581347-U) and a 0–100% gradient of water to acetonitrile containing 0.05% trifluoroacetic acid (TFA).

AmpB Control Resin Synthesis.

The PLAD linker was synthesized as described previously.²² PLAD linker resin was Boc-deprotected using 95% TFA, 2.5% water, and 2.5% triisopropylsilane (TIS) and agitated for 1 h. Boc deprotection was confirmed by a colorimetric Kaiser test.⁵⁰ Bromoacetic acid (BrAcOH; 2 M) and diisopropylcarbodiimide (DIC; 3.2 M) in anhydrous dimethylformamide (DMF) were agitated for 10 min and washed three times with DMF, and coupling was confirmed by Kaiser test. AmpB (AmpB; 2 M) in DMF was added to resin and agitated for 45 min and washed three times with DMF, and coupling was confirmed by Kaiser test. Synthesis was confirmed by cleaving the disulfide bond with TCEP hydrochloride (0.29 mg/mL) in water and verifying by MS.

Media Optimization.

Candida albicans were streaked onto YPD from frozen stocks and incubated overnight at 37 °C. Single colonies were added to saline (0.85%) solution until an absorbance of OD₅₃₀ = 0.15–0.25 was achieved. The media of interest (6 mL) was inoculated with cells (100 μL) and added to the cell-treated Petri plates (60 mm). Plates were incubated overnight, washed gently three times with PBS, and analyzed by microscopy to observe growth morphology (budding yeast, hyphae, or biofilm formation). YPD, RPMI with 3-(N-morpholino)propanesulfonic acid (0.1 M, pH 7.0; RPMI-MOPS), and Spider (1% Difco nutrient broth, 1% mannitol, 0.2% dibasic potassium phosphate, pH 7.2) media were assessed.

Reducing Reagent Optimization.

Several reducing reagents (DTT, βME, and TCEP) were evaluated for their toxicity against C. albicans at 0, 2, 6, 10, and 14 mM. C. albicans colonies were transferred from a streaked YPD plate to 0.85% saline to reach an OD₅₃₀ between 0.15 and 0.25. This inoculant was diluted 1:100 into RPMI-MOPS and then further diluted 1:20 into RPMI-MOPS. Two hundred microliters of the inoculant was seeded in each well of a 96-well black-walled plate. Plates were incubated overnight at 37 °C. The medium was gently removed, and biofilms were gently washed three times with PBS. RPMI-MOPS (190 μL) was added to each well. Solutions of reducing reagent (10× final concentration) were prepared in water, and 10 μL of the reducing reagent was added to each well in triplicate and then incubated at 37 °C for 24 h. Each well was imaged on a stereoscope with an angled baselight to provide a bright biofilm against a dark background. The images were analyzed in Adobe Photoshop for luminosity, which provides a measure of biofilm density and integrity. The medium was gently removed, and biofilms were gently washed with PBS. RPMI-MOPS (200 μL) was added to each well, followed by PrestoBlue (20 μL), and the plate was incubated at 37 °C for 1 h before measuring fluorescence on a SpectraMax M5 plate reader (Ex. 555 nm; Em. 585 nm).

Hit Reporter Optimization.

C. albicans colonies from a streaked YPD plate were added to saline (0.85%) solution to reach an absorbance between OD₅₃₀ = 0.15–0.25. RPMI-MOPS broth (6 mL) was inoculated with 100 μL of cell solution, added to a cell-treated Petri dish, and incubated overnight at 37 °C to form established biofilms. The medium was removed, and the biofilm was washed gently three times with PBS. Soft RPMI-MOPS agar (0.75% w/v; 3 mL) was melted and cooled to 42 °C. TCEP (14 mM) reducing reagent and an aliquot of resin, either AmpB on PLAD control resin or peptoid library RGL9 (5 mg in 0.5 mL of PBS), were added to the liquified agar, which was then mixed quickly and added on top of the established biofilm. Hit reporters to be studied [PrestoBlue (9% v/v), propidium iodide (9% v/v), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 0.5, 0.2, 0.1, and 0.05 μg/mL), and phloxine B (10 μg/mL)] were added either during addition of soft agar or after overnight incubation (only phloxine B). The images of the reporter interacting with the biofilm were analyzed by microscopy after a 24 h incubation at 37 °C.

RGL9 Library Synthesis.

The PLAD linker was synthesized as previously described.²² RGL9 was synthesized using a split-and-pool method onto TentaGel containing the PLAD linker. Briefly, resin (500 mg) was bromoacylated using bromoacetic acid (BrAcOH; 2 M) with diisopropylcarbiimi-dine (DIC; 3.2 M) in DMF and agitated for 10 min after microwave assistance (10% power; 15 s; 2×). Resin was then evenly divided into either 4 cationic or 10 noncationic reaction vials and treated with amine solutions (2 M) in DMF and agitated for 45 min after microwave assistance. Resin was then pooled together, and the entire bromoacylation, split, amination, and pool process was repeated six times. Positions four and seven were reserved for cationic submonomers, and positions two, three, five, and six were reserved for noncationic submonomers. The amines used during split-and-pool synthesis were Boc-ethylenediamine, Boc-diaminopropane, Bocdiaminobutane, Mmt-diaminopropane, homotryptamine, furfurylamine, 4-(aminomethyl)phenol, benzylamine, cyclohexylamine, methoxyethylamine, naphthylamine, isopropylamine, (+/−) phenylethylamine, and propylamine. This gave RGL9 a theoretical diversity of 1.6 × 10⁷. After split-and-pool synthesis was complete, a final addition of tridecylamine was completed using peptoid chemistry. The terminal amine was then Boc-protected using di-tert-butyl dicarbonate (Boc₂O) in 5% N-methylmorpholine (NMM) in DMF at 10 molar equivalents compared to resin loading capacity. Amines intended for guanidinylation to form arginine mimics were Mmt-deprotected using TFA (1%) in dichloromethane and agitated with resin (5×, 10 min). Guanidylation was achieved by agitating the library overnight with pyrazole carboxamidine (10 molar equivalents) with 4-dimethylaminopyridine (1 molar equivalent) in 5% NMM-DMF. Global Boc deprotection was achieved by treating with TFA, water, and TIS at a 95:2.5:2.5 ratio, respectively, for 1 h followed by washing with dichloromethane to yield the final RGL9 peptoid library.

iPLAD Assay.

A YPD agar plate was streaked with frozen stocks of C. albicans and incubated overnight at 37 °C. Single colonies were added to the saline solution (0.85%) until a turbidity between OD₅₃₀ = 0.15–0.25 was achieved. The cell solution (100 μL) was added to RPMI-MOPS (pH 7; 6 mL) and plated onto a cell-treated Petri plate (60 mm) and incubated overnight at 37 °C. The medium was removed, and biofilms were washed gently with PBS (3×). RPMI-MOPS soft agar (0.75% w/v) was liquefied and maintained at 42 °C until use. RGL9 (2–5 mg in 0.5 mL PBS) and TCEP (100 mM; 580 μL) were added to soft agar (4 mL total volume) and plated on top of biofilm. The plates were incubated at 37 °C overnight. Phloxine B (10 μg/mL; 1 mL) was added on top of the soft agar and incubated for an additional 1 h. Beads with red halos were extracted with surgical tweezers, placed in individual microcentrifuge tubes, and boiled in sodium dodecyl sulfate (1%) for 10 min, followed by washing with PBS (3×). The remaining α-strand peptoid on these beads was cleaved from resin by incubating with cyanogen bromide (40 mg/mL) in acetonitrile:water (1:1) with hydrochloric acid (0.1 M) overnight at room temperature in the dark. The cyanogen bromide solution was removed in vacuo, resuspended in acetonitrile:water (1:1) with TFA (0.05%), analyzed via MS, and sequenced using MS/MS to obtain structures of the unknown peptoids.

Peptoid Synthesis.

Peptoid sequences determined by MS/MS were resynthesized for antimicrobial characterization using the standard submonomer approach.¹⁴ Peptoids were synthesized on polystyrene rink amide resin. Resin was swelled in DMF for 20 min. Resin was Fmoc-deprotected using 20% piperidine in DMF and agitated for 10 min twice. Removal was confirmed by the colorimetric Kaiser test. Resin was bromoacylated using BrAcOH (2 M) and DIC (3.2 M) in anhydrous DMF and agitated for 10 min following microwave assistance. After washing with DMF, amines were coupled by agitating resin with the amine of choice (2 M) in anhydrous DMF for 25 min after microwave assistance followed by washing with DMF. This process was repeated to form the desired peptoid. Peptoids were cleaved from resin, and Boc groups were simultaneously deprotected by agitating with TFA, water, TIS (95:2.5:2.5) for 1 h. TFA solution was removed by bubbling the solution with air followed by resuspension in acetonitrile:water (1:1) with TFA (0.05%) and verification by MS. Peptoids were then purified using RPHPLC with a Supleco C-18 column and a water-to-acetonitrile (0 to 100%) gradient containing 0.05% TFA. Solvent was removed in vacuo to yield peptoids as white powders.

The amines used for the synthesis of RMG9–11 were Boc-1,3-diaminopropane, benzylamine, furfurylamine, Boc-1,4-diaminobutane, (±)-2-phenylethylamine, 1-aminopropane, and 1-aminotridecane. Synthesis starting with 0.26 g of Rink Amide resin (0.75 mmol/g loading capacity) yielded 37 mg of peptoid (16.4% yield). The structure of this compound was confirmed by ESI-MS; expected m/z = 1044.42 da; observed m/z 1044.62 da (Figure S11).

Fungal MIC Assay.

The MIC of peptoids against fungal pathogens C. albicans and C. neoformans was determined following CLSI guidelines.³⁸ Colonies were transferred from a streaked YPD plate to 0.85% saline to reach an OD₅₃₀ between 0.18 and 0.25. This inoculant was diluted 1:100 into RPMI-MOPS and then further diluted 1:20 into RPMI-MOPS. Then, 198 μL of inoculant was seeded in each well of a 96-well black-walled plate. Twofold serial dilutions of 100× peptoid solutions were prepared in water, and 2 μL of peptoid was added to each well in triplicate. The plate was incubated at 37 °C for 72 h for C. neoformans and 24 h for C. albicans before evaluating the MIC by manual observation. MIC was defined as the lowest concentration of compound preventing fungal growth. This assay was repeated three times on separate days with each compound.

Fungal BMIC Assay.

The BMIC of peptoids against fungal pathogens C. albicans was determined following CLSI guidelines.³⁸ Colonies were transferred from a streaked YPD plate to 0.85% saline to reach an OD₅₃₀ between 0.15 and 0.25. This inoculant was diluted 1:100 into RPMI-MOPS and then further diluted 1:20 into RPMI-MOPS. Then, 200 μL of inoculant was seeded in each well of a 96-well black-walled plate. Plates were incubated overnight at 37 °C. The medium was gently removed and gently washed three times with PBS. RPMI-MOPS (198 μL) was added to each well. Twofold serial dilutions of 100× peptoid solutions were prepared in water, and 2 μL of peptoid was added to each well in triplicate and then incubated at 37 °C for 24 h. PrestoBlue (20 μL) was added to each well and incubated at 37 °C for 1 h before measuring fluorescence on a SpectraMax M5 plate reader (Ex. 555 nm; Em. 585 nm).

Bacterial MIC Assay.

The MIC of peptoids against the ESKAPE bacteria (Enterococcus faecium ATCC 6569; Staphylococcus aureus ATCC 29213; Klebsiella pneumoniae ATCC 13883; Acinetobacter baumannii ATCC 19606; Pseudomonas aeruginosa ATCC 25619; Enterococcus faecalis ATCC 29212; and Escherichia coli ATCC 25922) was determined following CLSI guidelines.³⁷ The MIC against Mycobacterium smegmatis was also determined following CLSI guidelines.⁵¹ Colonies picked from streaked tryptic soy agar plates were added to tryptic soy broth to achieve a turbidity of OD₆₀₀ = 0.08–0.15. The inoculant was diluted 1:200 into cation-adjusted Mueller Hinton Broth, and 90 μL was plated into each well of a 96-well black-walled plate. Twofold serial dilutions of 10× peptoid (10 μL) were added to each well in triplicate and incubated at 37 °C for 24 h for the ESKAPE bacteria and 72 h for M. smegmatis. Tetracycline (20 μg/mL) was used as a positive control, and deionized water was used as a vehicle control. Following incubation for the ESKAPE bacteria, Presto Blue (10 μL) was added to each well and incubated at 37 °C for 1 h before measuring fluorescence on a SpectraMax M5 plate reader (Ex. 555 nm; Em. 585 nm). Following incubation for M. smegmatis, wells were scored following CLSI guidelines to determine MIC. This assay was repeated three times on separate days.

Mammalian Cytotoxicity.

Hepatocellular carcinoma (HepG2), mouse fibroblast (3 T3), and human keratinocyte (HaCat) cells were cultured in Dulbecco modified Eagle media (DMEM) containing 10% fetal bovine serum and 1% penicillin, streptomycin, and glutamine at 37 °C and 5% CO₂. Cells were collected and adjusted to 1 × 10⁵ to 4 × 10⁵ cells/mL in phenol-red free DMEM and plated (100 μL) into 96-well plates. Twofold serial dilutions of 10× peptoid (11.1 μL) were added to each well in triplicate. Water (vehicle) was used as a negative control. Plates were incubated at 37 °C in 5% CO₂ for 72 h. Thiazolyl blue tetrazolium bromide (MTT) was added to each well (5 mg/mL; 20 μL) and incubated for 3 h. The medium was removed, dimethyl sulfoxide (100 μL) was added, and plates were incubated at 37 °C for 10 min. Plates were read on a SpectraMax M5 plate reader (Abs. 570 nm). The concentration of compound resulting in a 50% reduction in growth compared to control (toxicity dose 50%; TD₅₀) was determined using GraFit. This procedure was repeated three times on separate days.

Hemolytic Assay.

Hemolytic activity was determined using single donor hRBCs. hRBCs were washed with PBS and centrifuged (1000 rpm; 10 min) three times, resuspended in PBS, and aliquoted (100 μL) in 96-well plates. Twofold serial dilutions of 10× peptoid final concentrations in PBS were prepared and added to wells in triplicate. PBS was used as a vehicle control, and 1% Triton X-100 was used as a positive control. Plates were incubated for 1 h (37 °C; 5% CO₂) and then centrifuged (1000 rpm; 10 min), and the supernatant was diluted 1:20 into PBS in a new 96-well plate. Plates were read on a SpectraMax M5 plate reader (Abs. 405 nm). Percent hemolysis was determined by the following equation

$$\%\text{hemolysis=}\frac{{\text{OD}}_{\text{405}\text{nm}}\text{sample}-{\text{OD}}_{\text{405}\text{nm}}405\text{negative}\text{control}}{{\text{OD}}_{\text{405}\text{nm}}\text{positive}\text{control}-{\text{OD}}_{\text{405}\text{nm}}\text{negative}\text{control}}\times 100$$

GraFit was used to determine concentrations at 50% hemolytic activity (HC₅₀) and the Hill coefficient (H). Hemolytic activity at 10% (HC₁₀) was then determined by the following equation

$${\text{HC}}_{\text{10}}={\text{HC}}_{\text{50}}{\left[10\%/\left(100\%-10\%\right)\right]}^{1/H}$$

ABBREVIATIONS

iPLAD

Inverted Peptoid Library Agar Diffusion

AMP

antimicrobial peptide

YPD

yeast extract peptone dextrose

MIC

minimum inhibitory concentration

TD₅₀

toxicity dose 50%

HC₁₀

hemolytic concentration 10%

SR

selectivity ratio