Indole and Derivatives Modulate Biofilm Formation and Antibiotic Tolerance of Klebsiella pneumoniae

Abstract

Intercellular communication is a crucial process for the multicellular community in both prokaryotes and eukaryotes. Indole has been recognized as a new member of the signal molecules which enables the regulated control of various bacterial phenotypes. To elucidate the inter-species relationship among enteric microorganisms via indole signaling, Klebsiella pneumoniae (KP) culture was treated with indole solution and examined for the pathogenicity by using various phenotypic tests. Both synthetic and naturally-produced indole preparations had no deteriorating effect on growth and autoaggregative capacity of KP. The results showed that biofilm formation of carbapenem-susceptible K. pneumoniae (KP-S) was clearly induced by indole exposure (≈ 2–10 folds), whereas no significant difference was observed in the resistant counterpart. In addition, the tolerance to β-lactam antibiotics of KP was altered upon exposure to indole and/or derivatives assessed by Kirby–Bauer disk diffusion test. Taken together, our finding indicates the functional role of indole in changing or modulating pathogenic behaviors of other bacteria.

Introduction

The interaction among microorganisms either cooperative or competitive manner has been described [1, 2]. This sophisticated process also known as quorum sensing (QS) is accomplished through many different kinds of signal molecules such as autoinducing peptides (AIP) in Gram positive bacteria, N-acyl-homoserine lactones (AHLs) in Gram negative bacteria, and autoinducer-2 (AI-2) in both Gram positive and Gram negative bacteria [3–5]. Recently, indole has been introduced as another class of QS molecule involved in the regulation of various bacterial physiologies and behaviors [6, 7]. It targets biofilm formation, production of virulence factors, and antibiotic resistance.

l-Tryptophan serves as a substrate of indole biosynthesis through a tryptophanase-catalyzed reaction [6]. Human gastrointestinal (GI) tract contains a substantial amount of this QS molecule owing to the presence of many indole-producing bacteria, especially E. coli [8, 9]. Other microorganisms may accept this signal molecule and cause alteration of their pathogenicity. Interestingly, each bacterium responds differently to indole exposure [6]. In the presence of indole and derivatives, toxin production of non-indole producing pathogens was clearly repressed in Pseudomonas aeruginosa [10, 11] and Staphylococcus aureus [12]. A biofilm-forming ability of bacteria has been found to be either increased or decreased in different microorganisms by the action of indole [6, 7, 13]. A recent review on association of gut microbiota-derived indole and derivatives with host innate immunity and homeostasis suggests the fundamental role of indole in host–microbes interactions [14]. Therefore, it is worth to prove that whether Klebsiella pneumoniae (KP) living non-pathogenically in the human intestine could become more virulent after exposure to indole. Both synthetic and naturally-produced indoles were used.

Materials and Methods

Chemicals, Bacterial Strains and Culture Condition

Escherichia coli BL21(DE3), E. coli BL21(DE3) tn5:tnaA and clinical strains of K. pneumoniae (KP) were recovered from − 20 °C stock culture collection, cultivated in Luria–Bertani (LB) broth and re-checked the species by biochemical tests. KP isolates resistant to meropenem, imipenem or ertapenem by disk diffusion method were designated as resistant KP (KP-R) strains, while the remaining isolates were designated as susceptible KP strains (KP-S) throughout this study. All bacteria were cultured in LB broth at 37 °C unless indicated otherwise. Indole was of analytical grade and purchased from Sigma-Aldrich.

Preparation of Bacteria-Free Culture Supernatant from E. coli Strains

Escherichia coli BL21(DE3) and its isogenic BL21(DE3) tn5:tnaA strains were used as indole producer and non indole producer, respectively. Luria–Bertani (LB) was used as a basic medium for the production of indole and derivatives. Bacterial culture was prepared by inoculation of a fresh single colony into LB broth and incubation at 37 °C for 16–18 h with 120 rpm agitation. Culture supernatant (CS) was collected after centrifugation at 12,000 rpm for 25 min, and was subjected to ultrasonic disintegration for 20 min (VCX 130, Vibra-Cell™, Sonics). The sterility and indole quantification of all prepared CS were determined by the drop plate method and Kovac’s reagent [15, 16], respectively, before using in any of the experiment.

Cytotoxicity and Effect of Indole and Derivatives on Growth of K. pneumoniae

Indole toxicity to KP isolates was examined by adding indole solution (final concentration of 10–5000 μM dissolved in 0.5% MeOH) into 0.5 McFarland bacterial culture in LB broth. Cultures were incubated at 37 °C for 18 h, and then the cell density was measured by spectrophotometer at 600 nm (PowerWave™ XS Microplate Reader, USA). Indole concentration with no effect on bacterial growth was subsequently used in further experiments. Besides, growth of KP isolates supplemented with synthetic indole solution or CS from indole-producing E. coli strain was intermittently monitored spectrophotometrically at 600 nm at 37 °C for 48 h. Treatment of KP cells with 0.5% MeOH or CS from indole-deficient E. coli strain was performed as a control. This experiment was done in triplicate.

Evaluation of Biofilm Formation

Method for assessing biofilm formation was derived from the previous report with a slight modification [17, 18]. In brief, 0.5 McFarland KP culture was prepared and then mixed with indole solution or CS from indole-producing E. coli strain in 96-well microtiter plate. Incubation was done at 37 °C for 48 h. Cell density (OD_600nm) was recorded prior to the staining of bacterial biofilm with 1% crystal violet (CV) for 30 min at room temperature. Planktonic cells and unbound dye were removed by decanting and washing with tap water until a clear solution was seen. 95% ethanol was added to extract biofilm CV and then measured absorbance at 595 nm. Normalization with bacterial optical density at 600 nm could avoid the effect of growth rate and cell density [17]. Normalized biofilm formation was calculated by OD_595nm/OD_600nm formula. Values were an average of 6 replicates.

Effect of Indole and Derivatives on Bacterial Aggregation

Determination of bacterial autoaggregation was carried out as reported elsewhere with minor modification [19]. Briefly, indole solution or CS from indole-producing E. coli was added into 0.5 McFarland KP culture. A mixture was cultivated at 37 °C for 24 h and subsequently held at 4 °C for 24 h to facilitate cell sedimentation. A 200-μL portion of cell suspension was taken before and after vigorous agitation for cell turbidity measurement at 600 nm to obtain OD_final and OD_initial, respectively. The calculation of autoaggregation percentage was done as the following formula: 100 × [1 − (OD_final/OD_initial)]. KP culture treated with 0.5% MeOH or CS from indole-deficient E. coli strain was considered as a vehicle control.

Changes in Antibiotic Tolerance by Indole and Derivatives

Kirby–Bauer disk diffusion assay was adopted from previous works with minor modification to examine the effect of indole and its derivatives on antibiotic susceptibility [20–22]. Mueller–Hinton agar (MHA) was made with the addition of indole solution or CS from E. coli. Overnight culture of KP isolates was adjusted to 0.5 Mcfarland turbidity and then spread evenly over the surface of MHA. Antibiotic-impregnated disks with defined concentration were placed onto the agar surface followed by incubation at 35 ± 2 °C for 16–18 h. Antibiotics tested in this work were 30 µg ceftazidime (CAZ), 30 µg ceftriaxone (CRO), 10 µg ampicillin (AMP), 10 µg amoxicillin–clavulanic acid (AMC), 30 µg cefoxitin (FOX), 10 µg imipenem (IPM) and 10 µg meropenem (MEM). Inhibition zone around each disk was measured, and values from the triplicate experiment were analyzed.

Statistical Analysis

Paired samples t test was employed to analyze the statistical differences between indole- and MeOH-treated cultures or CS from indole-positive E. coli- and CS from indole-negative E. coli-treated cultures. Statistically significant differences (p value < 0.05 or 0.01) were shown as asterisk (*).

Results and Discussion

Appropriate Indole Concentration and Effect on KP Growth

Multispecies bacterial community exerts a sophisticated cell-to-cell process called quorum sensing (QS) to maintain dynamic equilibrium. Indole and derivatives, the product of tryptophan degradation catalyzed by tryptophanase enzyme, have increasingly gained interest as a potential new class of QS molecule with diverse controlling behaviors including biofilm formation, motility and antibiotic tolerance [6, 7]. Millimolar level of indole naturally found in human intestine may trigger changes in bacterial virulence or physiology of non indole-producing bacteria including K. pneumoniae (KP) [6, 7, 23]. Effect of exposure to indole and derivatives on phenotypic behaviors of KP was investigated in this work. For avoiding cytotoxicity as well as using a physiological amount of indole, KP isolates were cultivated in LB medium supplemented with different indole concentration (10–5000 μM). Inhibition on the growth of all KP isolates was not observed at indole concentrations ≤ 500 μM (Fig. 1) which is close to a natural level in the human gut [8, 9]. Accordingly, this neutral effect on the growth of other non-indole producing microorganisms have also reported such as in Staphylococcus aureus [12], Pseudomonas aeruginosa [10], and P. putida [24]. The growth of KP-R isolate reduced considerably at 1 mM indole, whereas KP-S clearly unaffected (Fig. 1). Meanwhile, the toxicity to both KP isolates at high indole concentration (5 mM) correlated with previous studies. A significant decrease in growth caused by indole treatment has been observed in Acinetobacter baumannii (≥ 3.5 mM) [25] or P. putida (≥ 2 mM) [24]. Damage of cell membrane and suppression of cell division may be attributable to the harmfulness of indole [26, 27]. In order to investigate the indole effect without affecting on the bactericidal activity, further phenotypic experiments were therefore performed by using 500 μM of indole concentration.

Fig. 1.

Cytotoxicity of synthetic indole to KP isolates. Indole was added at different concentrations (0–5000 μM) to 0.5 McFarland KP culture as described in materials and methods. As a control, 0.5% MeOH was incubated with KP culture. The plot elucidates that indole concentrations of 1000 and 5000 μM are toxic to KP isolates. Mean, and SD values of triplicate individual experiments are shown

The bacterial-free CS prepared from E. coli BL21(DE3) (an indole-producer) and BL21(DE3) tn5:tnaA (an indole-deficient counterpart) strains was also evaluated. With our conditions used, CS from indole-producing E. coli strain contained 255 ± 45 μM indole and derivatives quantified by Kovac’s reagent and was used in all specified experiments. This work showed different quantity of indole from other studies, and this could be due to the different E. coli strains and culturing conditions as discussed in [6]. Exposure of KP to indole at physiological concentration [500 μM synthetic indole (Fig. 2a, b) or 255 μM natural indole and derivatives (Fig. 2c, d)] had no significant effect on bacterial growth in comparison with the control. The specific growth rates of KP cultures were 0.461–0.566 h⁻¹ (MeOH) versus 0.409–0.546 h⁻¹ (indole) and 0.390–0.504 h⁻¹ (indole-negative CS) versus 0.361–0.537 h⁻¹ (indole-positive CS). Thus, the exposure of indole and derivatives at physiological concentration did not significantly disturb the growth and survival of KP isolates.

Fig. 2.

Growth of KP isolates in the presence of indole. a Synthetic indole-treated KP-S; b synthetic indole-treated KP-R; c indole-containing CS-treated KP-S; d indole-containing CS-treated KP-R. Indole treatment (filled circle) has no effect on growth of KP isolates in comparison with its corresponding control (blank circle). The data represent mean and SD values from three independent experiments

Induced Biofilm Formation of KP by Indole and Derivatives

One of the most characterized traits regulated by indole signaling is biofilm formation [6, 7, 13]. Bacteria produce extracellular polymeric matrix forming plenty of three-dimensional cavities where they live inside. This complex structure, called biofilms, can act as a protective shelter against exogenous stress, inappropriate conditions and host immunity [28–30]. The present study emphasized the functional role of indole in modulating biofilm forming capability. Indole could significantly increase biofilm formation of KP-S isolates, measured by microtiter plate-based CV assay [17, 18], comparing with the corresponding control (p < 0.05) as seen in Fig. 3. Synthetic indole (tenfold) had a higher impact than natural indole (twofold) prepared from E. coli. This might be due to lower indole concentration in CS (255 μM indole and derivatives in CS) or the inhibitory effect of other indole derivatives in CS. One of the indole derivatives from Rhodococcus sp. BFI 332 (indole-3-acetaldehyde) has been shown to suppress the biofilm formation of E. coli O157:H7, whereas indole-3-acetic acid did not [31]. Figure 3 showed that the levels of biofilms in KP-R isolate treated with indole and the control were not statistically different. It has been demonstrated that biofilm biomass assessed by CV staining correlated well with biofilm volume and mean thickness assessed by SEM and CLSM, respectively [32, 33]. Hence, our study provided the clear evidence that indole signaling could modulate biofilm formation in several bacteria including KP.

Fig. 3.

Biofilm formation of KP isolates after indole exposure. The graph represents average CV absorbance of two KP isolates normalized by growth upon treatment of indole preparations as indicated in the figure. The graph clearly demonstrates that biofilm formation of indole-treated KP-S isolate (filled bar) is significantly increased in comparison with that of its control (blank bar). Mean, and SD values of 6 individual replicates are shown. The statistical difference between indole-treated experiment and its corresponding control (p < 0.05) are shown as an asterisk (*)

Since biofilm formation involves several steps including multiple cell aggregations, surface attachment, production of exopolysaccharides or other polymeric substances, biofilm maturation, and biofilm detachment [28, 30]. Effect of indole exposure on bacterial autoaggregation was also assessed in this study. As illustrated in Fig. 4, there was an increase, though not statistically significant, in autoaggregation level of synthetic indole-treated KP-S cells compared with the control. Treatment with synthetic indole or indole-positive CS did not affect the aggregative behavior of KP-R. Bacterial biofilms form initially by the production of extracellular polymeric substances (EPS) assisting cell clumping, surface attachment, and biofilm maturation. The positive correlation of biofilm forming ability to EPS production and bacterial aggregation has been well-documented [19, 34]. The finding of the present study suggested that the increased biofilms of KP may result from enhanced autoaggregative property induced by indole exposure. However, other mechanism such as enhanced exopolysaccharide production by indole signaling, which has been previously revealed in V. cholerae [35, 36], might also involve in this process.

Fig. 4.

Autoaggregation of indole-treated KP isolates. The graph represents average bacterial aggregability of KP isolates following indole treatment as labeled in the figure. The graph shows that there is no significant difference in autoaggregation of KP isolates between indole-containing solution (filled bar) and its control (blank bar). Mean, and SD values of triplicate individual replicates are shown

Effect of Indole and Derivatives on Antibiotic Susceptibility

In this work, the disk-diffusion susceptibility test was done to assess the antibiotic resistance of KP culture with or without indole supplementation. The result showed that indole could alter antibiotic susceptibility of KP isolates (Table 1). In KP-S strain, synthetic indole could slightly enhance the tolerance to ampicillin, and the inhibition zone of cells treated by CS containing indole and derivatives was smaller in meropenem and ceftriaxone indicating higher antibiotic resistance. Meanwhile, KP-S cultures treated with indole-positive CS were more susceptible to imipenem, amoxicillin, and cefoxitin. It is worth to note that lesser effect was observed in KP-R strain. It was demonstrated that KP-R strain following exposure to indole became more resistant or susceptible to imipenem or meropenem, respectively. As previously reported, indole and derivatives could modulate antibiotic tolerance in many pathogenic bacteria including P. aeruginosa [10, 11], P. putida [37, 38], and S. enterica serovar Typhimurium [39–41].

Table 1.

Antibiotic susceptibility of KP isolates after indole treatment

Antibiotic	KP-S				KP-R
	MeOH	Synthetic indole	Indole negative CS	Indole positive CS	MeOH	Synthetic indole	Indole negative CS	Indole positive CS
IPM	27.0 ± 0.0	27.3 ± 0.6	28.0 ± 0.0	30.3 ± 0.6*	10.7 ± 1.2	12.7 ± 1.2	17.3 ± 0.6	15.7 ± 0.6*
CAZ	27.7 ± 0.6	27.7 ± 0.6	27.7 ± 0.6	26.7 ± 0.6	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0
AMC	22.3 ± 0.6	23.0 ± 0.0	20.0 ± 0.0	22.7 ± 0.6*	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0
FOX	23.3 ± 0.6	24.0 ± 1.0	24.3 ± 0.6	26.3 ± 0.6*	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0
AMP	8.0 ± 0.0	7.0 ± 0.0*	9.0 ± 0.0	9.3 ± 0.6	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0
CRO	30.0 ± 0.0	27.0 ± 5.2	33.7 ± 0.6	32.3 ± 0.6*	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0	6.0 ± 0.0
MEM	26.3 ± 1.2	28.7 ± 1.2	27.3 ± 0.6	20.0 ± 0.0**	6.0 ± 0.0	7.7 ± 0.6*	6.0 ± 0.0	7.7 ± 0.6

Abbreviation for antibiotics: IPM imipenem, CAZ ceftazidime, AMC amoxicillin–clavulanic acid, FOX cefoxitin, AMP ampicillin, CRO ceftriaxone, MEM meropenem

*Indicates significant difference at p value < 0.05

**Indicates significant difference at p value < 0.01

The previous experiments have shown that the efflux pump system is one of the cellular mechanisms involved in indole-enhanced antibiotic resistance. Genes encoding efflux transporters were highly upregulated in the presence of indole assessed by quantitative PCR (qPCR). In Gram-negative bacteria, it is the AcrB gene responsible for the efflux system [42, 43]. Mutation or reduced expression of this gene leads to enhanced resistance to a variety of antibiotics, including β-lactams, quinolone, tetracycline, and erythromycin. Indole has been demonstrated to confer an effect on AcrB gene expression resulting in altered antibiotic tolerance in P. aeruginosa [10, 11] as well as Salmonella enterica Typhimurium [39–41, 44]. Altered tolerance to antibiotics induced by indole was also observed in KP. Contribution of the similar efflux pump system, like AcrAB [45], may take place to modulate antibiotic resistance of K. pneumoniae in response to indole. Indole-induced expression of two drug exporter genes including acrD and mdtABC has been observed in E. coli through two-component signal transduction systems [46]. Moreover, indole has also been shown to induce ttgAB encoding two genes of RND-type multidrug efflux operons and an ampC encoding β-lactamase in P. putida [37, 38]. Taken together, the role of indole in driving antibiotic tolerance may differ by individual. In addition, several mechanisms, including efflux pump system or β-lactamase may be attributable to indole-induced antibiotic tolerance of bacteria. Therefore, further study on the expression of genes involved in potential antibiotic tolerance processes, e.g., efflux pump system and β-lactamase, using qPCR is required for understanding the molecular mechanism underlying the indole-enhanced antibiotic resistance of KP.

Conclusions

Indole signaling was demonstrated in KP to involve in regulating biofilm formation, antibiotic tolerance, and some virulence phenotypes. Our study indicates that it is the indole, not derivatives, which seems to play a major role in this context. More importantly, this supports the interconnection through QS molecules between different microorganisms in the multi-species community as well as host and commensal bacteria in human body [6, 7, 14]. Further phenotypic and molecular investigations are essential to elucidate the more explicit mechanism of action of indole and derivatives in the regulation of virulence. These would help in better understanding of indole signaling and finding the potential inhibitors for bacterial virulence [47–49].

Compliance with Ethical Standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.