Indole is a signalling molecule, produced by a number of Gram-positive and Gram-negative bacteria both in nature as well as clinical environments. Here, we explored the effect of bacterial indole and one of its main derivatives on the virulence of the fungal pathogen Candida albicans.
Methods and Results
We found that indole and its derivate indole-3-acetonitrile (IAN) did not affect the viability of C. albicans. Interestingly, indole and IAN repressed C. albicans biofilm formation as well as the attachment of C. albicans to intestinal epithelial HT-29 cells and inhibited the ability of the yeast to make filaments that are the main virulence factor of C. albicans. In addition, we used the heterologous model host Caenorhabditis elegans to demonstrate in vivo that the presence of indole or IAN attenuates C. albicans infection (P = 0·0188 and P < 0·0001 for indole and IAN, respectively, compared to worms exposed to C. albicans DAY185 alone) and decreases fungal colonization in the nematode gut. Importantly, quantitative real-time polymerase chain reaction (qRT-PCR) results showed that in C. albicans, indole and IAN strongly stimulated the transcription of NRG1.
Indole and IAN attenuates fungal virulence by regulating the transcription of NRG1, a transcriptional factor that influences filamentation and biofilm formation in C. albicans.
Significance and Impact of the Study
Our findings indicate that the bacterial signalling molecules indole and its derivatives play an inter-kingdom role in dynamic network of microbiota and directly modulate the virulence of fungal C. albicans via NRG1.
In polymicrobial communities, bacteria and fungi directly and indirectly influence each other's growth, survival and physiology through multi-factorial pathways (Hogan and Kolter 2002). In particular, the virulence of bacteria or fungi can be influenced by substances produced by other microbes and secreted molecules mediate many types of interactions between bacteria and fungi (Peleg et al. 2010). However, despite the ubiquitousness of bacterial–fungal competition, little is known about the identities of these signalling molecules and their effects. In the microbe world, studies that investigate these interactions are particularly valuable, because a complex microbial community that involves a number of signalling networks is present in nature as well as in clinical environments including human intestinal tract (Hsiao et al. 2008; Lee and Lee 2010).
A number of studies demonstrated that extracellular signalling molecules from bacteria can mediate quorum sensing (QS) and QS molecules produced by one organism and can affect entire cell populations (Miller and Bassler 2001). Importantly, many bacterial genes, including virulence factors, are stimulated by these small signalling molecules produced by individual cells of the same or different species (Peleg et al. 2010). Several species including Pseudomonas aeruginosa employ their own signalling molecules (autoinducer-1 or AI-1, a family of related homoserine lactones) for intraspecies communication purposes, whereas other bacterial species (e.g. Escherichia coli and Salmonella spp.) use a common signalling molecule, AI-2, for interspecies and universal communications (Surette and Bassler 1998; Rasmussen et al. 2005). However, there are a few reports on the cross-kingdom effects of the universal signalling molecules.
A variety of both Gram-positive and Gram-negative bacteria produce indole as an intracellular signalling molecule (Lee and Lee 2010). It has been reported that indole can control the virulence of pathogenic bacteria (Lee et al. 2011) as well as modulate the host immune system (Bansal et al. 2010). In addition, several indole derivates such as indole-3-acetonitrile (IAN), which is a plant-associated growth hormone originated from broccoli, cauliflower and cabbage, could inhibit the biofilm formation and virulence factor productions of Ps. aeruginosa and enterohaemorrhagic E. coli (EHEC) O157:H7 (Lee et al. 2011).
Similar to bacteria, the pathogenic fungus Candida albicans produces at least two signalling molecules called farnesol and tyrosol that influence growth, virulence and other physiological pathways (De Sordi and Muhlschlegel 2009). Although inter-kingdom signalling molecules play an important role in competition between bacteria and fungi, or the underlying molecular pathways involved (Hogan and Kolter 2002; Peleg et al. 2008), their roles in micro-organism world are still unclear. Here, we provide direct evidence that bacterial indole and IAN influence the virulence of fungal C. albicans, by altering the attachment as well as the filament and biofilm formation by C. albicans. Also, using qRT-PCR studies, we explored whether indole directly influences fungal virulence via NRG1, the transcriptional factor that influences filamentation and biofilm formation in C. albicans. Finally, we used the established model host Caenorhabditis elegans (reviewed in (Irazoqui et al. 2010)) to study the role of these molecules in eukaryotic-prokaryotic competition in vivo.
Materials and methods
The C. albicans strains used in this study were DAY185 (ura3Δ::λimm434/ura3Δ::λimm434 ARG4:URA3::arg4::hisG/arg4::hisG his1::hisG::pHIS/his1::hisG) as the wild-type strain (Davis et al. 2000) and the efg1Δ/cph1Δ double mutant (ura3Δ::λimm434/ura3Δ::λimm434 cph1Δ::hisG/cph1Δ::hisG efg1Δ::hisG/efg1Δ::hisG-URA3-hisG) (Lo et al. 1997). Yeast strains were grown in yeast-peptone-dextrose (YPD) or on brain heart infusion (BHI; Difco, Detroit, MI, USA) agar containing kanamycin (50 μg ml−1) at 30°C (for the C. elegans solid killing assay). The E. coli OP50 (Breger et al. 2007) was routinely cultured in Luria-Bertani (LB) medium (Difco) at 37°C.
In the nematode assays, we used the C. elegans wild-type strain N2 Bristol, the strain CF512 fer-15(b26);fem-1(hc17), for C. elegans killing assay, and the strain AU37 glp-4(bn2);sek-1(km4) for C. elegans filamentation assay. These strains have been described in previous papers (Murphy et al. 2003; Kim and Mylonakis 2011) and nematodes were cultured and maintained on E. coli OP50 using standard procedures (Brenner 1974; Breger et al. 2007).
Silicone pad biofilm assay
The effect of indole and IAN on C. albicans biofilm growth was evaluated using a silicone pad assay, as described previously (Kim and Mylonakis 2011). In brief, the biofilm mass was calculated by subtracting the original weight of the silicone pad from its postincubation weight at 60 h and adjusted for the weight of control silicone pads not exposed to fungal cells. Spider medium was used in the C. albicans biofilm assays.
Candida albicans attachment assay
The human intestinal epithelial HT-29 cell line was obtained from the Korea Cell Line Bank (KCLB, Seoul, Korea). Cells were routinely cultured in RPMI 1640 medium (Gibco BRL, Rockville, MD, USA) supplemented with 10% heat-inactivated foetal bovine serum (FBS). Candida albicans attachment experiments were performed as described previously (Kim et al. 2009) with some modifications. Prior to the attachment assay, the HT-29 monolayers were washed three times with PBS to remove culture medium and nonattached cells. The prepared C. albicans strains (3 × 102 cells ml−1), with 1 mmol l−1 indole or IAN, were added to the monolayers at 37°C in an atmosphere of 5% CO2. After allowing 6 h, the monolayers were washed six times with PBS to remove nonattached bacteria, and the attached cells were collected using a cell scraper. Serial dilutions of the mixture were plated onto YPD agar containing kanamycin and incubated at 37°C for 48 h.
Filamentation assay using Caenorhabditis elegans
The C. elegans coinfection assays for C. albicans filamentation were performed as described in (Kim and Mylonakis 2011). Briefly, young adult nematodes (n = 20) were infected for 4 h on lawns of C. albicans DAY185 and then transferred into wells (six-well microplate). Each well contained 2 ml of liquid assay medium (20% BHI and 80% M9) containing 0·2 mmol l−1 indole or IAN. The plates were then incubated at 25°C and examined for 5 days for the formation of penetrative filaments.
Caenorhabditis elegans solid killing assay
The C. elegans assays were performed using the protocols described by (Pukkila-Worley et al. 2011). In brief, 10 μl of an overnight C. albicans DAY185 culture was spread onto a BHI agar (that containing 0·2 mmol l−1 indole or IAN as needed). Then, young adult fer-15;fem-1 worms (n = 30) were infected by placing them on the lawns (three assay plates were used per condition). Animals were incubated at 25°C and scored as live or dead on a daily basis by gently touching them with a platinum wire. Worms that crawled onto the walls of the tissue culture plates were eliminated from the analysis.
Measuring fungal colonization in the intestinal tract of Caenorhabditis elegans
The numbers of fungal cells in the worms intestine were determined as previously described (Garsin et al. 2001) with minor modifications. More specifically, at day 5, each worm was removed to a new sterile tube containing M9 medium with 1% Triton X-100 and was mechanically disrupted using a pestle (Kontes, Vineland, NJ, USA). The concentration of yeast cells was evaluated by diluting cells from 100 to 107 via 10-fold serial dilution steps in 0·85% NaCl solution; these dilutions were then plated on YPD agar containing kanamycin (50 μg ml−1), ampicillin (100 μg ml−1) and streptomycin (100 μg ml−1) (Kim et al. 2010). Plates were incubated at 30°C for 48 h.
Quantitative real-time PCR (qRT-PCR)
The qRT-PCR assays were performed using the Applied Biosystems 7300 Real-Time PCR system (Applied Biosystems, Foster, CA, USA). Following mechanical disruption with zirconia/silica beads (Biospec, Bartlesville, OK, USA) and a pestle in Trizol solution (Invitrogen, Carlsbad, CA), total RNA samples from C. albicans were purified using a RNeasy Mini kit according to the manufacturer's instructions (Qiagen, Valencia, CA, USA). After isolating RNA, 50 ng of total RNA was used for the qRT-PCR reaction using the Power SYBR® Green RNA-to-CT™1-Step kit (Applied Biosystems). The primers sequences used were as follows: HWP1 (5′-CTCCAGCTGGCTCAAGTGGT-3′ and 5′-TGGCAGATGGTTGCATGAGT-3′), NRG1 (5′-CACCTCACTTGCAACCCC-3′ and 5′-GCCCTGGAGATGGTCTGA-3′), LIP2 (5′-GGCCTGGATTGATGCAAGAT-3′ and 5′-GGCCTGGATTGATGCAAGAT-3′), ALS3 (5′-ACTTCCACAGCTGCTTCCAC-3′ and 5′-TGCAGATGGAGCATTACCAC-3′) and 18s rRNA (5′-GTGCCAGCAGCCGCGGTA-3′ and 5′-TGGACCGGCCAGCCAAGC-3′). Relative expression levels were calculated using the method (Livak and Schmittgen 2001). Expression levels of the control gene 18S rRNA were used to normalize the expression data for C. albicans.
Candida elegans survival is presented using the Kaplan–Meier method and the significance of differences in survival was determined using the log-rank test (STATA6; STATA, College Station, TX, USA). Differences between experiments were determined using Student's t-test. Each result shown is representative of at least two independent experiments. A P-value of 0·05 in all replicates experiments was considered statistically significant.
Indole and IAN dramatically inhibit biofilm formation by Candida albicans
Initially, to investigate whether indole signalling molecules influence the viability of C. albicans DAY185, growth curves were measured in the presence or absence of indole and IAN in YPD. As expected, there was no significant difference in the growth curves of C. albicans exposed to 1 mmol l−1 indole or IAN compared to the C. albicans alone (data not shown); hence, we indicated indole and IAN do not affect the viability of C. albicans. And then, we evaluated whether indole and IAN influenced C. albicans biofilm formation using the standard silicone pad assay. As shown in Fig. 1, C. albicans biofilm formation was dramatically repressed (by approximately 80% compared to control) by the presence of indole and IAN. These observations confirm that the bacterial signalling molecules indole and IAN influence fungal biofilm formation and/or fungal attachment.
Indole and IAN reduce the ability of Candida albicans attachment
In this series of experiments, we explored whether the bacterial signalling molecules indole and IAN affect the attachment of the fungal pathogen C. albicans to HT-29 intestinal epithelial cells. The presence of indole and IAN did not alter the viability of the HT-29 cell monolayer over a period of 6 h (data not shown), but indole and IAN significantly reduced the attachment of C. albicans to HT-29 cells (Fig. 2a).
Indole and IAN inhibit the filamentation and attenuate virulence of Candida albicans in Caenorhabditis elegans in vivo by decreasing the fungal burden in the nematode intestine
Moreover, we evaluated C. albicans filamentation in the presence of indole and IAN in vivo using the established model C. elegans. Corroborating the biofilm and attachment results, filamentation in the presence of the indole and IAN was significantly less than that of the C. albicans control (P = 0·025 and P = 0·010 for indole and IAN, respectively, compared to worms exposed to C. albicans DAY185 alone) (Fig. 2b).
Using a solid killing assay, we showed that infection of C. elegans fer-15(b26);fem-1(hc17) with C. albicans DAY185 was lethal within 7 days, whereas nematodes infected with the attenuated efg1Δ/cph1Δ double mutant strain of C. albicans were not killed. Interestingly, the presence of indole and IAN significantly enhanced the resistance of C. elegans to infection with C. albicans (Fig. 3a; P = 0·0188 and P < 0·0001 for indole and IAN, respectively, compared to worms exposed to C. albicans DAY185 alone). Similar results were obtained with N2 wild-type worms (data not shown). Because C. albicans colonizes the nematode intestinal tract and causes a persistent lethal infection (Breger et al. 2007), we evaluated the possibility that indole and IAN decreased the number of persistent C. albicans in the C. elegans intestine by assessing the number of colony forming units (CFU) of C. albicans in the worm intestines. We found that worms treated with indole or IAN had significantly fewer C. albicans cells present in their intestines after 5 days compared to control worms (Fig. 3b).
Indole and IAN specifically stimulate the transcription of NRG1
Based on the findings above that indole and IAN are involved in the regulation of several biofilm-associated factors, we examined the expression levels of C. albicans genes involved in adherence as well as filament and biofilm formation. As shown on Fig. 4, qRT-PCR assays showed that NRG1 expression was upregulated by indole and IAN (4·3 ± 1·7-fold for indole and 7·6 ± 1·8 for IAN, respectively), whereas there was no significant difference in the expression levels of HWP1, LIP2 and ALS3. Therefore, our qRT-PCR results suggest that, at least in part, indole and its derivatives may inhibit fungal virulence by altering the expression of NRG1.
Bacterial indole signalling is important for controlling microbial biofilm formation by pathogenic E. coli (Lee et al. 2011) and Vibrio cholerae (Mueller et al. 2009). Although we previously showed that bacterial signalling molecules can influence the attachment ability of pathogenic bacteria (Kim et al. 2007), a few study has investigated the role of these signalling molecules in the adherence of yeast cells. In the current study, we confirmed that indole is an interactive regulator between bacteria and fungi, and showed that the anti-fungal activity of indole and its derivative IAN is related to the regulation of NRG1. In this work, we provide multiple lines of evidence that indole and IAN directly regulate fungal virulence: (i) indole and IAN did not influence fungal growth, (ii) indole signalling decreased fungal virulence in vitro as shown in biofilm, attachment and filamentation assays (Figs 1 and 2) and in vivo as shown in the C. elegans killing assay (Fig. 3) and (iii) NRG1 expression was altered by indole signalling (Fig. 4). Hence, indole plays a critical role in specific bacteria–fungi interaction networks as a cross-kingdom signalling molecule.
Medically and environmentally important interactions between bacteria and fungi are common in nature (Peleg et al. 2010). Cell-to-cell communication, which is mediated by signalling molecules, is critical in polymicrobial communities. Although the intra- and inter-species-specific behaviours of these molecules have been investigated (Miller and Bassler 2001), only a few studies have examined cross-kingdom signalling. Reen and colleagues (Reen et al. 2011) showed that PQS and HHQ, which are 4-quinolone signalling molecules produced by Pseudomonas aeruginosa, significantly repressed the biofilm formation by fungi (C. albicans) as well as bacteria (Bacillus subtilis). These molecules are not naturally present in the human intestine and are toxic. However, a variety of both Gram-positive and Gram-negative bacteria (more than 80 species) can produce indole, and indole signalling is involved in spore formation, drug resistance, virulence, plasmid stability and biofilm formation in multi-species bacterial communities (Lee and Lee 2010). We showed that indole and IAN directly modulate the virulence of C. albicans. It is reasonable to assume that a high concentration of these signalling molecules is physiologically present in the host intestine. For example, strains of E. coli can produce up to 0·6 mmol l−1 indole, which is not toxic to the host (Chant and Summers 2007).
Consistent that filament formation is critical for C. albicans virulence (Lo et al. 1997) involved in biofilm formation (Nobile and Mitchell 2005; Richard et al. 2005), we showed indole and IAN repressed C. albicans filamentation using the established in vivo model C. elegans (Breger et al. 2007; Pukkila-Worley et al. 2009; Tampakakis et al. 2009; Kim and Mylonakis 2011). Here, we employed the low, nontoxic concentration of 0·2 mmol l−1 indole or IAN for the C. elegans killing assay as indicated previously (Anyanful et al. 2005) as 0·5 or 1 mmol l−1 concentrations of indole or IAN were toxic to C. elegans (data not shown). Equally interesting, under same conditions, we showed indole or IAN significantly attenuates the toxicity of C. albicans. Taken together, these studies demonstrate that bacterial indole and IAN inhibit fungal attachment and filamentation as well as biofilm formation, and prolong the survival of nematodes exposed to fungal C. albicans by decreasing the burden of C. albicans infection in the intestine, and decreasing C. albicans virulence.
Based on the findings presented in this report, indole can be considered to be an inter-kingdom signalling molecule as well as an inter-genus and inter-species signalling molecule. Previously, it has been well established that several genes included HWP1 (encodes a filament-specific cell wall protein) (Nobile et al. 2006), NRG1 (encodes a transcriptional repressor and regulates filament formation and virulence) (Uppuluri et al. 2010), LIP2 (encodes a biofilm-associated lipase) (Xie et al. 2011) and ALS3 (encodes filament-specific adhesion) (Zhao et al. 2006) are important role in filament and biofilm formation of C. albicans. Unexpectedly, these factors were not involved in response to indole or IAN treatment. Interestingly, this effect is by affecting the transcription of NRG1, a gene that encodes a DNA-binding protein. Much has been learned that NRG1 critically appears to influence all these aspects of C. albicans pathogenesis (Braun et al. 2001; Murad et al. 2001). As recently shown (Uppuluri et al. 2010), NGR1 gene expression is involved in C. albicans biofilm formation and dispersal. Based on previous results, we speculated that indole cross-kingdom signalling would affect NRG1 transcription; we found that NRG1 transcription was induced by indole and IAN, resulting in a significant reduction in fungal filamentation and biofilm formation. As noted above, in C. albicans, biofilm formation is playing critical roles in their virulence (Richard et al. 2005) and is involved in two major processes, (i) attachment to biotic or abiotic surface and (ii) morphological transition from yeast to filaments are major requirements for biofilm formation (Blankenship and Mitchell 2006).
To our knowledge, this is first report that bacterially produced indole controls fungal virulence by regulating the transcription of NRG1 gene, a transcriptional factor. Our findings support the idea that indole-related signalling molecules may play an important role in inter-kingdom regulatory networks that are relevant to human health.
This study was supported by research funds of Chonbuk National University in 2012.