Antibiofilm and Antivirulence Properties of Indoles Against Serratia marcescens

Indole and its derivatives have been shown to interfere with the quorum sensing (QS) systems of a wide range of bacterial pathogens. While indole has been previously shown to inhibit QS in Serratia marcescens, the effects of various indole derivatives on QS, biofilm formation, and virulence of S. marcescens remain unexplored. Hence, in the present study, we investigated the effects of 51 indole derivatives on S. marcescens biofilm formation, QS, and virulence factor production. The results obtained revealed that several indole derivatives (3-indoleacetonitrile, 5-fluoroindole, 6-fluoroindole, 7-fluoroindole, 7-methylindole, 7-nitroindole, 5-iodoindole, 5-fluoro-2-methylindole, 2-methylindole-3-carboxaldehyde, and 5-methylindole) dose-dependently interfered with quorum sensing (QS) and suppressed prodigiosin production, biofilm formation, swimming motility, and swarming motility. Further assays showed 6-fluoroindole and 7-methylindole suppressed fimbria-mediated yeast agglutination, extracellular polymeric substance production, and secretions of virulence factors (e.g., proteases and lipases). QS assays on Chromobacterium violaceum CV026 confirmed that indole derivatives interfered with QS. The current results demonstrate the antibiofilm and antivirulence properties of indole derivatives and their potentials in applications targeting S. marcescens virulence.


INTRODUCTION
Serratia marcescens is a Gram-negative, rod-shaped bacterium that belongs to the Enterobacteriaceae family and well-known for its ability to produce prodigiosin a red pigment . The bacterium is often isolated from clinical samples and is ubiquitous in nature. During the past four decades, S. marcescens has been increasingly recognized as an important opportunistic pathogen that has developed resistance to many antibiotics and is frequently associated with nosocomial infections in pediatric, adult, aged, and immunocompromised patients. S. marcescens causes surgical site, eye, respiratory tract, bloodstream, urinary tract, urinary catheter-associated, and gastrointestinal tract infections and endocarditis, meningitis, and other diseases (Cristina et al., 2019). It has been estimated that 6.5% of Gram-negative bacterial pathogenic infections are caused by Serratia spp. in the United States and Europe (Sader et al., 2014). Serratia spp. is also known to be the 7th and 10th most common cause of pneumonia and bloodstream infections, respectively, in United States and Europe (Acar and Goldstein, 1997;Jones, 2010). S. marcescens isolates of clinical origin have been shown to produce extended-spectrumβ-lactamase (Yang et al., 2012), and to acquire multiple drug resistance via horizontal gene transfer from other members of the Enterobacteriaceae family (Ivanova et al., 2008). The bacterium is known to utilize a broad array of nutrients and to grow in the presence of antiseptics, detergents, and disinfectants (Cooney et al., 2014).
Quorum sensing (QS) in S. marcescens plays important roles in antibiotic resistance, biofilm formation, synchronizing the productions of proteases, lipases, prodigiosin, and butanediol, and swimming and swarming motilities (Van Houdt et al., 2007). S. marcescens strains produce a wide range of N-acylhomoserine lactones (AHLs) (e.g., C4-AHL, C6-AHL, 3-oxo-C6-AHL, C7-AHL, and C8-AHL), which it uses as QS signal molecules (Eberl et al., 1996;Horng et al., 2002;Coulthurst et al., 2006). Furthermore, biofilms formed by S. marcescens clinical isolates are resistant to commonly used antibiotics (Ray et al., 2017), and biofilm formation by clinically important bacterial pathogens is primarily responsible for device-associated chronic infections. QS defective mutants have been reported to be less virulent and incapable of forming robust biofilms, and thus, QS inhibition is a strategy used to control the virulence of S. marcescens (LaSarre and Federle, 2013). Several bioactive compounds, including alpha-bisabolol and vanillic acid, have been successfully used to inhibit virulence factor production and biofilm formation by S. marcescens (Sethupathy et al., 2016b(Sethupathy et al., , 2017. In biofilms, bacterial cells are surrounded by a self-secreted polymeric matrix comprised of macromolecules such as proteins, lipids, carbohydrates, and extracellular DNA, which protect the bacterium from environmental stress factors, disinfectants, host immune system, and antibiotics (Stewart and Costerton, 2001;Høiby et al., 2010). Bacterial cells in biofilms are metabolically less active and grow more slowly and this characteristic facilitates their acquisition of antibiotic resistance (Stewart, 2002). Hence, it appears suitable combinations of biofilm/QS inhibitors and conventional antibiotics might usefully enhance the antibiotic susceptibilities of bacterial cells in biofilms (Brackman et al., 2011).
It has been reported that more than 85 species of Grampositive and Gram-negative bacteria can synthesize indole . Indole is an important bacterial signal molecule that regulates several important biological processes such as genetic stability, metabolism, biofilm formation, pathogenesis, antibiotic resistance, and oxidative stress responses and also acts as an interspecies and interkingdom signal to regulate diverse functions (Lee et al., 2015b). Several studies have described the antibiofilm and antivirulence activities of indole derivatives (e.g., 7-hydroxyindole, 3-indoleacetonitrile, 7-fluoroindole, 7benzyloxyindole, and methylindoles) against clinically important pathogens, such as enterohemorrhagic Escherichia coli (Lee et al., 2007), Pseudomonas aeruginosa (Lee et al., 2009(Lee et al., , 2012, Staphylococcus aureus (Lee et al., 2013), and Candida albicans Manoharan et al., 2018). In addition, indole and 3-indolylacetonitrile have been shown to inhibit the maturation of Paenibacillus alvei endospores (Kim Y.-G.et al., 2011), and halogenated indoles have been reported to have nematicidal and insecticidal potentials (Rajasekharan et al., 2019(Rajasekharan et al., , 2020. Although the QS inhibitory activity of indole in S. marcescens has been clearly reported in previous work (Hidalgo-Romano et al., 2014), the effects of indole derivatives on the biofilm and virulence of indole-negative S. marcescens have yet to be evaluated, and thus in the present study, we investigated the antibiofilm and antivirulence potentials of several indole derivatives against S. marcescens.

Indole Compounds
Indole and 50 indole derivatives (Supplementary Figure 1) were purchased from Sigma-Aldrich (St. Louis, MO, United States) and Combi-Blocks, Inc. (San Diego, CA, United States), dissolved in dimethyl sulfoxide (DMSO) to produce 1 M stock solutions, and stored at −20 • C. DMSO (0.1% v/v) was used as the negative control and at the concentrations present (<0.1%) did not affect bacterial growth or biofilm formation.

Bacterial Culture Conditions
S. marcescens ATCC 14756 was streaked on Luria-Bertani (LB) agar and incubated at 30 • C for 24 h. Plates were stored at 4 • C until required. For biofilm and other assays, a single colony of S. marcescens was inoculated in LB broth and cultured at 30 • C for 12 h at 160 rpm. Two percentage of the overnight culture (adjusted to 0.5 McFarland containing ∼1 × 10 8 CFU mL −1 ) was used as an inoculum for biofilm and other virulence assays.

Prodigiosin Assay
S. marcescens was grown in the absence or presence of indole or indole derivatives at 30 • C for 20 h and centrifuged at 12,000 rpm for 10 min. Acidified ethanol (1 mL of 4% 1 M HCl in ethanol) was added to the cell pellets obtained and vortexed vigorously to extract prodigiosin. After centrifugation, supernatant absorbances were measured at 534 nm as previously reported (Slater et al., 2003).

Biofilm Inhibition Assay
S. marcescens cells were inoculated in 2 mL LB and incubated at 30 • C for 20 h. To assess biofilm inhibitory activity, cells were re-inoculated into LB (dilution ratio 1:50) and cultured overnight in 96-well plates in the absence or presence of indole or indole derivatives under static conditions at 30 • C for 20 h. After incubation, planktonic cell densities were measured at 620 nm and culture supernatant were discarded. Residual planktonic cells were removed by washing plates three times with water. Biofilms that formed on the plates were stained with crystal violet (0.1%) for 20 min, excess dye was removed by washing three times, and bound crystal violet was solubilized in 95% ethanol. Absorbances were measured at 570 nm using a Multiskan EX microplate photometer (Thermo Fisher Scientific, Waltham, MA, United States) .

Growth Curve Analysis and Determination of Minimum Inhibitory Concentrations (MICs)
S. marcescens was grown in the absence or presence of indole or selected indole derivatives, as described above, and planktonic cell growth was monitored periodically at 600 nm for 24 h using an Optizen 2120UV spectrophotometer (Mecasys Co., Ltd., Daejeon, Korea). MIC was defined as the lowest concentration that inhibited planktonic cell growth by 80% and also confirmed by colony counting. MICs were determined for 3-indoleacetonitrile, 5-fluoro-2methylindole, 5-fluoroindole, 6-fluoroindole, 5-methylindole, 7-methylindole, and indole as previously described (Lee et al., 2013).

Microscopic Observation of Biofilms
S. marcescens cells were inoculated in 2 mL of LB and incubated at 30 • C for 20 h. Cells were then re-inoculated in LB at a dilution ratio of 1:50, cultured overnight in 96well plates with indole, 6-fluoroindole, or 7-methylindole at 1 mM, and then incubated at 30 • C for 24 h without shaking. For confocal laser scanning microscope (CLSM) analysis, cells were stained with 100 µL of pre-warmed PBS containing carboxyfluorescein diacetate succinimidyl ester for 20 min at 30 • C (final concentration, 5 µM) and then washed with PBS . Cells were visualized by a CLSM (Nikon Eclipse Ti, Tokyo, Japan) using a 20× objective and an Ar laser (excitation wavelength 488 nm, emission wavelength range 500-550 nm). In each experiment, at least 10 random positions in three independent cultures were chosen for microscopic analysis.

Swarming and Swimming Assay
Swimming agar (1% peptone, 0.5% NaCl, and 0.3% agar) and swarming agar plates (1% peptone, 0.5% NaCl, and 0.6% agar) were prepared with or without indole (0.5 or 1 mM) or selected indole derivatives. Plates were spotted with 5 µL overnight culture of S. marcescens, incubated in an upright position at 30 • C for 16 h, and then swimming and swarming motility inhibitions were assessed as previously described (Pearson, 2019).

Protease Assay
Extracellular casein degrading protease activities of supernatants from S. marcescens grown in the absence or presence of indole or indole derivatives were measured using 2% w/v of azocasein. Briefly, equal volumes of azocasein and culture supernatants were reacted at 37 • C for 30 min and then 600 µL of 10% trichloroacetic acid was added to stop the proteolysis. Reaction tubes were kept for 30 min at −20 • C to precipitate unreacted azocasein and centrifuged. An equal volume of 1 M NaOH was added to supernatants (700 µL), and absorbances were read at 440 nm as previously described (Coulthurst et al., 2006).

Lipase Assay
The effects of indole and indole derivatives on extracellular lipase production were evaluated by incubating 1 volume of supernatant from a S. marcescens control with 9 volumes of substrate buffer [1 volume of buffer A containing 3 mg/mL of pnitrophenyl palmitate in isopropyl alcohol, 9 volumes of buffer B containing 1 mg/mL of gummi arabicum and 2 mg/mL sodium deoxycholate in 50 mM Na 2 PO 4 buffer (pH-8.0)] for 30 min in the dark at room temperature. After incubation, reaction tubes were centrifuged at 12,000 rpm for 10 min and lipase activity was terminated by adding 1 volume of 1 M Na 2 CO 3 . Absorbances was measured at 405 nm as previously described (Patel et al., 2018).

Fimbria Activity Assay
Effects of indole and indole derivatives on S. marcescens fimbria activity were accessed using Saccharomyces cerevisiae (Sigma, product no. YSC2), as previously described (Shanks et al., 2007). Yeast agglutination was measured spectrophotometrically by adding 1.5 mL of PBS containing 0.5 mL of S. cerevisiae (2% w/v in PBS) and 0.4 mL of S. marcescens cells in PBS (OD 600 , 0.5). To achieve a uniform mixture of Saccharomyces cerevisiae and Serratia marcescens cell suspension, the reaction tubes were gently vortexed for 5 s and the initial OD 600 was measured. After 10 min of incubation at room temperature, 100 µL of upper phase was transferred to a 96 well plate and OD 600 was measured. During 10 min incubation, fimbria present on the S. marcescens cell binds with S. cerevisiae and agglutinates S. cerevisiae. Formation of agglutination indicates the presence of fimbria on S. marcescens cells and a decrease in the fimbria results in the reduction of aggregate formation. Presence of visible aggregates of agglutinated cells affected the OD 600 measurement and hence vigorous vortexing for 30 s was done to disturb the agglutinated cells before the reading of the second OD 600 values. Percentage agglutination was calculated using 100 × (1-OD 600 before vortexing/OD 600 after vortexing).

H 2 O 2 Sensitivity Assay
Disk diffusion assays were performed by spreading control and treated cells on LB agar plates, placing 6-mm sterile paper disks on the agar, loading disks with 10 µL of 30% H 2 O 2 , and then incubating plates for 24 h at 30 • C. Zones of inhibition were measured as previously described (Shanks et al., 2007).

Extracellular Polymeric Substance (EPS) Extraction and FTIR
Extracellular polymeric substance (EPS) extraction was carried as previously described (Badireddy et al., 2008). Briefly, 100 mL of S. marcescens control and 6-fluoroindole or 7-methylindole (1 mM) treated cultures for 24 h were centrifuged at 7,000 rpm for 15 min at 4 • C to collect cells. Cell-free culture supernatants were stored at −20 • C for cell-free EPS extraction. Collected cell pellets were washed with wash buffer (10 mM Tris/HCl pH 8.0, 10 mM EDTA), resuspended in 100 mL of isotonic extraction buffer (10 mM Tris/HCl pH 8.0, 10 mM EDTA, 2.5 mM NaCl), incubated at 4 • C for 12 h, vortexed for 5 min, and centrifuged at 5,000 rpm for 15 min to collect supernatants containing cell bound EPS. Cell-free culture supernatants containing cell free EPS and isotonic buffer containing cell bound EPS were pooled, mixed with 3 volumes of ice-cold ethanol, and kept at −20 • C for 18 h to precipitate EPS. Precipitated EPS was collected by centrifugation at 10,000 rpm for 10 min at 4 • C, vacuum dried, and analyzed by FTIR spectrometry (Spectrum Two TM FTIR, PerkinElmer, Massachusetts, United States) at 400-4,000 cm −1 .

RNA Isolation and Transcriptomic Studies
For transcriptomic analyses, 25 mL of S. marcescens at an initial turbidity of 0.05 at OD 600 was inoculated into LB broth in 250 mL Erlenmeyer flasks and incubated for 6 h at 30 • C with agitation at 250 rpm in the presence or absence of 6-fluoroindole (0.5 mM). To prevent RNA degradation, RNase inhibitor (RNAlater, Ambion, TX, United States) was added to cells immediately after incubation. Total RNA was isolated using a hot acidic phenol method (Amin-ul Mannan et al., 2009), and RNA was purified using a Qiagen RNeasy mini Kit (Valencia, CA, United States).
Quantitative reverse transcriptase PCR (qRT-PCR) was used to determine the expressions of 10 QS-related genes list all bmsA (biofilm), carA (prodigiosin production), fimA (type 1 fimbriae), flhD (motility), luxS (quorum sensing), pigA (prodigiosin production), pigC (prodigiosin production), SmaI/R (LuxIR-type quorum sensing system), and rpoS (motility and biofilm). The specific primers and housekeeping gene (16S rRNA) used for qRT-PCR are listed in Supplementary  Table 1. The expression of 16S rRNA was not affected by 6-fluoroindole. The qRT-PCR method used was as described by Kim Y.-G.et al. (2016), and was performed using SYBR Green master mix (Applied Biosystems, Foster City, United States) and an ABI StepOne Real-Time PCR System (Applied Biosystems). At least two independent cultures with four repetitions were used.

Statistical Analysis
Most assays were conducted with two independent cultures with six repetitions while motility, EPS quantification, and qRT-PCR assays were performed with two independent cultures with four repetitions. Results are expressed as means ± standard deviations. The student's t-test was used to determine the significances of intergroup differences, and statistical significance was accepted for P < 0.05.
Growth curve analysis was used to evaluate the effects of these six derivatives and indole on the growth of S. marcescens. The obtained results revealed 6-fluoroindole and 7-methylindole had slight bacteriostatic activity but not in the presence of indole up to 1 mM (Figures 1C,F,I).

Effects of Indole Derivatives on Protease and Lipase Productions
Proteases secreted by bacterial pathogens play important roles in the establishment of infections and in systemic dissemination by degrading host defense proteins (Saint-Criq et al., 2018). Indole and 3-indoleacetonitrile, 5-fluoroindole, 6-fluoroindole, 5-methylindole, and 7methylindole at 1 mM were found to inhibit extracellular protease production effectively and dose-dependently by 25-60% (Figures 3D-F and Supplementary Figure 8).

Effects of Indole Derivatives on Fimbria-Mediated Yeast Agglutination and on Sensitivity to H 2 O 2
In S. marcescens, the transcriptional regulator OxyR is required for the regulation of oxidative stress response and the initial stages of biofilm formation via the modulation of the expression of type I fimbria (Shanks et al., 2007). Furthermore, QS inhibitors such as phenol, 2,4-bis(1,1-dimethylethyl) (Padmavathi et al., 2014), and phytol (Srinivasan et al., 2016) have been shown to inhibit fimbrial expression in S. marcescens, and hence, we evaluated the effect of selected indole derivatives on fimbria-mediated yeast agglutination and sensitivity to H 2 O 2 . We found S. marcescens cells in the presence of 5-fluoroindole, 6-fluoroindole, 5-fluoro-2-methylindole, 5-methylindole, 7methylindole, or indole at 1 mM S. cerevisiae agglutination by 36-68% (Figures 3J-L and Supplementary Figure 10). Also, we found the sensitivities of 6-fluoroindole, 7-methylindole, or indole treated S. marcescens cells to H 2 O 2 were greater than that of non-treated controls ( Figure 4A).
Frontiers in Microbiology | www.frontiersin.org or HHL mostly restored QS activity in the presence of 6fluoroindole (6FI) and 7-methylindole (7MI) at 0.25 mM while two AHLs could not complement the presence of the higher amount of two indoles at 0.5 mM (Figure 6B). This result also supports the previous finding that violacein inhibition by indole could be counteracted by the exogenous C10-AHL (Hidalgo-Romano et al., 2014).
The pig-gene cluster encodes for a group of enzymes responsible for the biosynthesis of prodigiosin in S. marcescens, and the majority of these genes are involved in the conversion of 2-octenal to monopyrrole MAP (2-methyl-3-amylpyrrole) and the bipyrrole moiety MBC (4-methoxy-2-2 -bipyrrole-5carbaldehyde). pigC encodes for an enzyme that condenses MBC and MAP to produce prodigiosin (Pan et al., 2020). A significant reduction in prodigiosin levels was observed after exposing with the QS inhibitor hordenine, which downregulated the expressions of pig-genes (Zhou et al., 2019). The expression of the pig-gene cluster is also affected by various transcriptional factors (Pan et al., 2020), and is a potential target of indole derivatives.
Therapeutic agents that reduce the virulence of pathogens are topics of active research in the pharmaceutical industry and in academia. Protease defective strains of P. aeruginosa (Breidenstein et al., 2012) and Vibrio cholera (Rogers et al., 2016) reduce motility, biofilm formation, and virulence, and bioactive compounds like N-mercaptoacetyl-Phe-Tyramide (Cathcart et al., 2011), curcumin (Rudrappa and Bais, 2008;Sethupathy et al., 2016a), and hydroxamic acid (Kany et al., 2018) have been reported to reduce the virulence and biofilm formation by P. aeruginosa by targeting protease. Lipases are important secreted virulence factors that support bacterial and fungal pathogens during the early infection stage by damaging the phospholipid layers of host cells and disrupting innate defenses (Chen and Alonzo, 2019). Bioactive compounds such as alpha-bisabolol (Sethupathy et al., 2016b), vanillic acid (Sethupathy et al., 2017), and phytol (Srinivasan et al., 2016) have been shown to affect protease and lipase production in S. marcescens by interfering with QS. The protease and lipase inhibitory activities of indole derivatives against S. marcescens (Figure 3 and Supplementary Figures 8, 9) warrant further investigation as potential therapeutic agents, and the results of our H 2 O 2 sensitivity assay also suggest the disruption of oxidative stress response in S. marcescens, which is one of the prerequisites of biofilm formation under challenging conditions in vivo.
Our findings indicate that both 6-fluoroindole and 7methylindole reduce yeast agglutination and increase sensitivity to H 2 O 2 probably through the differential regulation of OxyR in S. marcescens.
Previously, the addition of exogenous C10-AHL could restore the production of QS activity (Hidalgo-Romano et al., 2014) and the current studies also showed that the additions of other AHLs (BHL or HHL) mostly restored QS activity in the presence of indole derivatives (Figure 6B). Similarly, 6-gingerol (Kim H.-S.et al., 2015) and quercetin (Gopu et al., 2015) were reported to inhibit violacein production in the presence of AHL by blocking AHL-transcriptional receptor protein complex, which is essential for the violacein biosynthesis. In the majority of Gram-negative pathogens, AHLtranscriptional receptor protein complex formation is essentially required for the activation of QS controlled phenotypes and virulence gene expression (Miller and Bassler, 2001;Mukherjee and Bassler, 2019) and unlike antimicrobial agents, inhibiting AHL-transcriptional receptor protein complex formation by indoles is expected to reduce pressure favoring the development of drug resistance.
While indole and most of the indole derivatives did not have significant antimicrobial activity in S. marcescens (Figure 1 and Supplementary Figure 3), a few indoles such as 5-fluoroindole, 5-iodoindole and 5-methylindole showed significant antimicrobial activity (Supplementary Figure 4). Also, in the C. violaceum strain, most indoles at 0.25 and 0.5 mM markedly affect planktonic cell growth (Supplementary Table 2). Furthermore, it was reported that 5-iodoindole showed strong bactericidal activity against Escherichia coli strains and Staphylococcus aureus  and could rapidly kill Acinetobacter baumannii (Raorane et al., 2020). Therefore, it is important to carefully assess the toxicity of indoles on bacteria and animals.
The present study demonstrates the abilities of indole derivatives to inhibit QS in S. marcescens, and thus, to inhibit prodigiosin pigment production, biofilm formation, swimming motility, swarming motility, and fimbrial activity. Of the indole derivatives tested, 6-fluoroindole and 7-methylindole potently inhibited lipase, protease, and EPS production in S. marcescens. AHL supplementation assay using C. violaceum CV026 confirmed the disruption of QS by indole derivatives. Based on the results obtained, we suggest that indole and indole derivatives interfere with QS in S. marcescens and C. violaceum possibly by preventing AHL molecules binding to their cognate receptors. Based on our observation that indole derivatives differentially inhibit the virulence of and biofilm formation by S. marcescens, we suggest further studies be undertaken to determine the molecular mechanism involved.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Materials, further inquiries can be directed to the corresponding authors.
Supplementary Figure 2 | Concentration-dependent inhibition of prodigiosin production by selected indole derivatives in S. marcescens. Error bars and asterisks represent standard deviation and the statistically significant difference (p < 0.05), respectively. Supplementary Figure 8 | Effects of indole derivatives on protease production by S. marcescens. Error bars and asterisks represent standard deviation and the statistically significant difference (p < 0.05), respectively.
Supplementary Figure 9 | Effects of indole derivatives on lipase production in S. marcescens. Error bars and asterisks represent standard deviation and the statistically significant difference (p < 0.05), respectively.
Supplementary Figure 10 | Effects of indole derivatives on the fimbria-mediated yeast agglutination by S. marcescens. Error bars and asterisks represent standard deviation and the statistically significant difference (p < 0.05), respectively.
Supplementary  Figure 6 of QS inhibition assay.