The Cell-Cell Communication Signal Indole Controls the Physiology and Interspecies Communication of Acinetobacter baumannii.

Many bacteria utilize quorum sensing (QS) to control group behavior in a cell density-dependent manner. Previous studies have demonstrated that Acinetobacter baumannii employs an N-acyl-L-homoserine lactone (AHL)-based QS system to control biological functions and virulence. Here, we report that indole controls biological functions, virulence and AHL signal production in A. baumannii. The biosynthesis of indole is performed by A1S_3160 (AbiS, Acinetobacter baumannii indole synthase), which is a novel indole synthase annotated as an alpha/beta hydrolase in A. baumannii. Heterologous expression of AbiS in an Escherichia coli indole-deficient mutant also rescued the production of indole by using a distinct biosynthetic pathway from the tryptophanase TnaA, which produces indole directly from tryptophan in E. coli. Moreover, we revealed that indole from A. baumannii reduced the competitive fitness of Pseudomonas aeruginosa by inhibiting its QS systems and type III secretion system (T3SS). As A. baumannii and P. aeruginosa usually coexist in human lungs, our results suggest the crucial roles of indole in both the bacterial physiology and interspecies communication. IMPORTANCE Acinetobacter baumannii is an important human opportunistic pathogen that usually causes high morbidity and mortality. It employs the N-acyl-L-homoserine lactone (AHL)-type quorum sensing (QS) system, AbaI/AbaR, to regulate biological functions and virulence. In this study, we found that A. baumannii utilizes another QS signal, indole, to modulate biological functions and virulence. It was further revealed that indole positively controls the production of AHL signals by regulating abaI expression at the transcriptional levels. Furthermore, indole represses the QS systems and type III secretion system (T3SS) of P. aeruginosa and enhances the competitive ability of A. baumannii. Together, our work describes a QS signaling network where a pathogen uses to control the bacterial physiology and pathogenesis, and the competitive ability in microbial community.

INTRODUCTION

It is well known that quorum sensing (QS) is used by many bacteria to coordinate communal behaviors in a cell density-dependent manner (1–3). To date, many kinds of QS signaling molecules have been identified. Among them, N-acyl-L-homoserine lactones (AHLs) are well characterized and employed by many Gram-negative bacteria (4–7). In addition, other kinds of QS signals, such as diffusible signal factor (DSF) family signals, autoinducer-2 (AI-2), 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS), 2-heptyl-4-quinolone (HHQ), methyl 3-hydroxypalmitate, autoinducer-3 (AI-3), bradyoxetin and diketopiperazines, were also identified as being utilized by bacterial cells to communicate with each other (8–14). In recent years, increasing evidence has confirmed the important role of indole in both intraspecies signaling and interspecies communication. Indole regulates spore formation, cell division, plasmid stability, antibiotic tolerance, biofilm formation, motility, and virulence in both indole-producing and non-indole-producing bacteria (15–17). The biosynthesis process for indole was also identified in Escherichia coli, in which biosynthesis of indole is performed by TnaA (18, 19). Indole induces the expression of multidrug exporter genes through the two-component systems BaeSR and CpxAR in E. coli (20). It also affects antibiotic tolerance of Pseudomonas fluorescens through EmhR (21).

Acinetobacter baumannii is an opportunistic Gram-negative pathogen that usually causes pneumonia and infections in the bloodstream and urinary tract, resulting in high morbidity and mortality (22–24). Abuse of antibiotics has led to the high clinical drug resistance of A. baumannii, and carbapenem-resistant A. baumannii has been listed in the WHO’s first-ever list of priority antimicrobial-resistant pathogens (25–26). Previous studies have shown that A. baumannii employs the AHL-type QS system, which consists of the AHL synthase AbaI and the transcriptional regulator AbaR, to regulate biological functions and virulence (27–30). AbaI produces AHL signals, and AbaR functions as a receptor protein for AHL signals. After AbaR binds to AHL signals, the complex binds to the promoter region of target genes and controls gene expression.

QS signals play important roles in microbial ecology through interspecies communications, in addition to their critical role in intraspecies signaling (8, 9, 31). A previous study showed that indole produced by the gut microbiota suppressed the expression of the virulence genes of the enteric pathogens enterohemorrhagic E. coli (EHEC) (17). Furthermore, indole is not only an extracellular biofilm signal for E. coli but also represses QS-regulated virulence factors, including pyocyanin, PQS, pyochelin and pyoverdine of P. aeruginosa (32–34). As an interspecies communication signal, exogenous indole inhibits the persister cell waking in P. aeruginosa (35). It also reduces growth and motility but increases biofilm formation and enhances antibiotic tolerance of Agrobacterium tumefaciens, which was found not to synthesize indole (36).

In this study, we showed that indole produced by AbiS in A. baumannii repressed the QS system and type III secretion system (T3SS) of P. aeruginosa and then enhanced the competitive advantages of A. baumannii against P. aeruginosa. In-frame deletion of abiS disrupted biofilm formation, motility and virulence in A. baumannii. Furthermore, we also found that indole controls the AHL QS system in A. baumannii. As A. baumannii and P. aeruginosa usually coexist in the human lung, our data suggest that in addition to the important role of indole in the physiology and pathogenesis of A. baumannii, it also engages in interspecies competition, the importance of which in survival and infection is evident.

RESULTS

The ethyl acetate extract of A. baumannii interferes with the QS system and T3SS of P. aeruginosa.

As A. baumannii and P. aeruginosa usually share the same niche in the human lung, we then investigated whether there is competition or a collaborative relationship between the two pathogens. We first extracted low-molecular-weight compounds from a 1 L liquid culture of A. baumannii using ethyl acetate and then concentrated the extract to 1 mL with methanol. The results showed that the 1% ethyl acetate extract exerted an inhibitory effect on the motility and cytotoxicity of P. aeruginosa (Fig. 1a and b). As P. aeruginosa has evolved both QS systems and T3SS for the regulation of important biological functions and pathogenicity (10, 34, 37), we continued to test the effect of the extract and found that exogenous addition of the ethyl acetate extract of A. baumannii could reduce the transcriptional expression of both the master regulator-encoding gene of the T3SS and the regulator-encoding genes of QS systems (Fig. 1c and d), that is, lasR, rhlR and pqsR, but did not obviously affect the bacterial growth of P. aeruginosa (Fig. S1 in the supplemental material), suggesting that A. baumannii produces some compounds that interfere with the physiology and pathogenicity of P. aeruginosa.

FIG 1

Effects of ethyl acetate extract of A. baumannii on P. aeruginosa. Effect of the 1% A. baumannii extract on the motility (a) and cytotoxicity (b) of P. aeruginosa PAO1, the transcriptional expression of both the master regulator-encoding gene of the T3SS (c) and the regulator-encoding gene of the QS systems (d). Extract was dissolved in methanol, and the same volume of methanol used as the solvent for the compounds was used as a control. The data are the means ± standard deviations of three independent experiments. The significance of result a, b, c was determined by unpaired t test. The significance of result d was determined by two-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

The major active components of A. baumannii are indole and its derivatives.

To identify the active components of A. baumannii that interfere with the physiology and pathogenesis of P. aeruginosa, we isolated and purified the active fractions from 140 L A. baumannii ATCC17978 culture supernatants using high-performance liquid chromatography (HPLC). There was a total of four peaks with active components, which were collected and purified for determination of the chemical structures (Fig. S2a in the supplemental material). Among these peaks, approximately 6.2 mg of the purified compound of peak 4 was purified. ESI-MS spectrometry analysis of the active compound revealed a molecular ion [M+H]+ with an m/z ratio of 118.2 (Fig. S2b), consistent with the molecular formula of C8H7N. There were six protons in the aromatic region and one active hydrogen in the 1H NMR spectrum (Table 1). The 13C NMR data of the compound showed the presence of eight aromatic carbons (Table 1). The results are consistent with the literature (38), which indicated that the active compound was indole (Fig. S2b in the supplemental material). In addition, three indole derivatives were also isolated from the active ingredients and characterized: peak 1 was identified as indole-3-acetic acid (IAA), peak 2 was identified as 2-(2,2-di [1H-indole-3-yl] ethyl) aniline (di-IEA), and peak 3 was identified as methyl 2-(1H-indole-3-yl) acetate (MIA). The electrospray ionization-tandem mass spectrometry (ESI-MS) spectrometry analysis results of the three active compounds are shown in Fig. S2c, 2d, and 2e. A total of 100 μM indole and its derivatives showed obvious inhibitory activity on the motility and cytotoxicity of P. aeruginosa PAO1 (Fig. 2). We also measured the effect of indole and its derivatives on the growth rate of P. aeruginosa PAO1 and found that exogenous addition of 100 μM indole and its derivatives had little influence on the growth rate of P. aeruginosa PAO1 in both LB and MP culture media (Fig. S3 in the supplemental material).

TABLE 1

1H NMR (400 MHz) and 13C NMR (101 MHz) data of indole and its derivatives

No.	Indolea		di-IEAa		MIAa		IAAb
	δ C	δ H	δ C	δ H	δ C	δ H	δ C	δ H
NH-1		7.76		7.86		8.11
2	124.3	7.15	122.0	7.12	123.3	7.12	124.6	7.13
3	102.5	6.54	119.7		108.3		108.9
3a	127.9		127.0		127.3		128.6
4	119.9	7.67	119.2	6.91–7.01	118.9	7.60	119.4	7.53
5	120.8	7.03	119.8	7.47	119.7	7.04	119.8	7.00
6	122.0	7.20	121.9	6.91–7.01	122.2	7.18	122.4	7.08
7	111.2	7.27	111.2	7.28	111.4	7.27	112.2	7.32
7a	135.8		136.7		136.2		138.0
8			34.5	4.84	31.2	3.77	31.9	3.71
9			37.2	3.39	172.8		176.4
10			126.1		52.1	3.68
11			127.0	6.91–7.01
12			115.8	6.53
13			130.4	6.91–7.01
14			118.8	6.61
15			144.8
NH-1’				7.86
2’			122.0	7.12
3′			119.7
3’a			127.0
4’			119.2	6.91–7.01
5′			119.8	7.47
6’			121.9	6.91–7.01
7’			111.2	7.28
7’a			136.7

The solvent was CDCl3.

The solvent was CD3OD.

FIG 2

Effects of indole and indole derivatives on the biological functions of P. aeruginosa. Exogenous addition of 100 μM indole, 100 μM di-IEA, 100 μM MIA and 100 μM IAA on the motility (a) and cytotoxicity (b) of P. aeruginosa PAO1. Compounds was dissolved in DMSO, and the same volume of DMSO used as the solvent for the compounds was used as a control. The data are the means ± standard deviations of three independent experiments. The significance was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

Indole and its derivatives interfere with the T3SS and QS systems of P. aeruginosa.

As indole and its derivatives inhibited the motility and cytotoxicity of P. aeruginosa, we continued to investigate the mechanisms by which these compounds affect P. aeruginosa. We first evaluated the effect of indole and its derivatives on the T3SS system of P. aeruginosa. The real-time quantitative reverse transcription-PCR (RT-qPCR) and PexsCEBA-lacZ fusion reporter analyses showed that the addition of 100 μM indole and IAA decreased the transcriptional expression levels of exsCEBA (Fig. 3). However, di-IEA and MIA had no inhibitory activities (Fig. 3).

FIG 3

Effects of indole and indole derivatives on the T3SS system of P. aeruginosa. (a) The effects of 100 μM indole, 100 μM di-IEA, 100 μM MIA, and 100 μM IAA on the expression of the master regulator-encoding gene of the T3SS were evaluated by RT-qPCR (OD600 = 1.0). (b) Inhibitory effect of 100 μM indole, 100 μM di-IEA, 100 μM MIA, and 100 μM IAA on the expression of the T3SS system, as determined by using PexsCEBA-lacZ transcriptional fusion reporter strains (OD600 = 3.0). The β-galactosidase activity of PexsCEBA-lacZ in the P. aeruginosa PAO1 wild-type strain was arbitrarily defined as 100% and used to normalize the β-galactosidase activity of PexsCEBA-lacZ in the P. aeruginosa PAO1 strains supplemented with indole and its derivatives. Compounds was dissolved in DMSO, and the same volume of DMSO used as the solvent for the compounds was used as a control. The data are the means ± standard deviations of three independent experiments. The significance was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

Then, we evaluated the effect of indole and its derivatives on the QS systems of P. aeruginosa. The results of fluorescence RT-qPCR and promoter-lacZ fusion reporter assays showed that the expression levels of lasI, rhlI, pqsA, lasR, rhlR and pqsR were significantly reduced with the addition of indole (Fig. S4a and c in the supplemental material). Consistent with these results, the production of C4-HSL, 3-oxo-C12-HSL and PQS was decreased significantly in the presence of indole (Fig. S4b). The three derivatives showed obvious inhibitory activities on the QS system, except exogenous addition of 100 μM IAA, which exhibited no detectable effect on the transcriptional expression levels of lasR and pqsR (Fig. S4d–f).

Biosynthesis of indole is performed by AbiS in A. baumannii.

Previous studies have confirmed the important role of indole in cell–cell communication. To further study the functions of indole in A. baumannii, we first looked for the enzyme involved in the biosynthesis of indole and its derivatives. When we selected TnaA from E. coli for homology analysis in A. baumannii, we found no homologous protein in A. baumannii. But we noticed that EstC in Burkholderia cenocepacia was noted to be a methyl indole-3-acetate methyltransferase. Given the fact that EstC is also an esterase in some other bacteria, so we supposed that EstC is possibly involved in the production of indole. Homologs of estC were then searched for in the genome sequence of A. baumannii ATCC17978, and A1S_3160 was identified as one such homolog by using the National Center for Biotechnology Information Basic Local Alignment Search Tool (BLAST) program (39). The A1S_3160 protein has an alpha/beta hydrolase fold (Fig. 4a), and mutation of this protein completely abolished indole production (Fig. S5) but resulted in only a slight decrease in the growth rate of the bacterial cells in both LB medium and MP medium (Fig. S6). We named this enzyme indole synthase (AbiS).

FIG 4

The biosynthesis of indole is performed by AbiS in A. baumannii. (a) Domain structure analysis of AbiS (top). Genomic organization of the abiS region in A. baumannii ATCC17978 (bottom). (b) Time-course analysis of indole accumulation in the A. baumannii wild-type strain, the ΔabiS deletion mutant strain and the ΔabiS(abiS) complemented strain. The data are the means ± standard deviations of three independent experiments.

To investigate whether the biosynthesis of indole is related to cell density, we measured the time course of indole production by determining indole concentrations at various times (Fig. 4b). The yield of indole was quantitatively calculated by the correlation equation determined by the standard curve (Fig. S7 in the supplemental material). BLAST searches revealed that the AbiS homologs are widely conserved in many other bacterial genera, including Acinetobacter, Marinobacter, Klebsiella, and Salmonella (Table S1).

The indole biosynthetic pathway in A. baumannii is different from that in E. coli.

A previous study indicated that TnaA can reversibly convert tryptophan to indole, pyruvate and ammonia (19). To date, this is the sole indole production pathway identified in bacteria (15). BLAST searches revealed that the TnaA homologs were also conserved in many bacterial genera (Table S2 in the supplemental material). Moreover, it was found that some bacterial species have both TnaA and AbiS homologs (Table S3). To determine whether the biosynthetic pathway of indole in A. baumannii is the same as or different from that in E. coli, we knocked out tnaA in E. coli K12 and overexpressed AbiS and TnaA in the E. coli tnaA mutant, respectively. Then the strains mentioned above were cultured to an OD600 of 1.2 in both 0.5 × LB medium and MP+ medium. It was reported that LB contains 0.5–0.6 mM tryptophan (40), so we then cultured these strains in 0.5 × LB firstly. The indole yield of E. coli wild-type strain was measured as 208.98 μM (Fig. 5a). It was found that deletion of tnaA completely abolished the production of indole, which was restored to the wild-type strain level in the complemented strain (Fig. 5a). Intriguingly, in trans expression of AbiS in the tnaA deletion mutant only slightly rescued the indole production to 21.76 μM (Fig. 5a). We also found that exogenous addition of 300 μM tryptophan can significantly increase indole production in E. coli to 477.97 μM. Meanwhile, the addition of 300 μM tryptophan had no effect on the indole yield of ΔtnaA(abiS) strain (Fig. 5a). To further confirm the different pathways conducted by AbiS and TnaA, we cultured the E. coli wild-type, tnaA deletion mutant, tnaA(tnaA) complement, and tnaA(abiS) strains in the tryptophan-deficient MP+ medium. It was found that both the indole yields of E. coli wild-type and complement strains were remarkable lower than those in 0.5 × LB, while the ΔtnaA(abiS) strain produced the similar amount of indole in both tryptophan-rich and tryptophan-deficient media (Fig. 5b), suggesting that tryptophan is not a precursor for the biosynthesis of indole by AbiS.

FIG 5

Analysis of indole biosynthesis performed by AbiS. (a) Analysis of the production of indole in the E. coli wild-type, ΔtnaA, ΔtnaA(tnaA) and ΔtnaA(abiS) in 0.5 × LB medium, and in the ΔtnaA(abiS) and E. coli wild-type strains in 0.5 × LB medium supplemented with 300 μM tryptophan. (b) Analysis of the production of indole in the E. coli wild-type, ΔtnaA, ΔtnaA(tnaA), and ΔtnaA(abiS) strains in MP+ medium. The data are the means ± standard deviations of three independent experiments. The significance was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

Deletion of abiS impairs biological functions and pathogenicity in A. baumannii.

It was determined that indole and its derivatives interfere with the QS systems and T3SS of P. aeruginosa through interspecies communication. We continued to study whether indole plays an important role in the regulation of biological functions and pathogenicity in A. baumannii. As deletion of abiS completely abolished the production of indole, we tested the phenotypic changes in biofilm formation, motility and cytotoxicity in an in-frame deletion mutant of abiS. Deletion of abiS decreased biofilm formation and motility by 25.49% and 61.91%, respectively (Fig. 6a and b), while it attenuated cytotoxicity by 52.66% when A549 cells were incubated with the A. baumannii strains at 8 h postinoculation (Fig. 6c).

FIG 6

Effects of abiS and indole on the biological functions of A. baumannii. The virulence-related phenotypes of biofilm formation (a), motility (b) and cytotoxicity (c) in the A. baumannii wild-type, abiS mutant, abiS complemented strains and abiS mutant with addition of different concentrations of indole were examined. Compounds was dissolved in DMSO, and the same volume of DMSO used as the solvent for the compounds was used as a control. The data are the means ± standard deviations of three independent experiments. The significance was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

Interestingly, both the in trans expression of abiS and the addition of exogenous indole restored these phenotypes of the abiS deletion mutant to the wild-type strain level (Fig. 6). Further investigation was conducted to analyze the symptoms of mouse lungs infected by the mutant strains because the lung is the most important niche for A. baumannii. At 7 d postinfection, the macroscopic pathological findings indicated that, compared with the normal group (Fig. S8a in the supplemental material), a significant reduction in the number of alveoli and obvious injury of the epithelial cell lining of bronchioles were observed in the groups infected by the wild-type (Fig. S8b) and complemented (Fig. S8d) strains. In addition, some alveolar walls were severely damaged or broken, and the microvasculature was congested and bled when the mice were infected with the wild-type and complemented strains, while only mild injury was observed when the mice were infected with the abiS mutant strain (Fig. S8c).

Indole controls the AHL QS system and the expression of various genes in A. baumannii.

The QS system of A. baumannii consists of an AHL synthase (AbaI) and a transcription regulator (AbaR), which is activated upon binding to AHL. N-(3-hydroxydodecanoyl)-L-homoserine lactone (3-OH-C12-HSL) is the major AHL signal found in A. baumannii ATCC17978 (29). The RT-qPCR analysis and the promoter-lacZ fusion reporter assay results showed that mutation of abiS caused a decrease in the expression level of abaI (Fig. 7a and b). We also found that the production of 3-OH-C12-HSL was reduced in the abiS mutant strain by 54.52% (Fig. 7c). These results demonstrated that indole positively regulated the AHL QS system in A. baumannii.

FIG 7

Effects of abiS on the AHL QS system of A. baumannii. The expression of the AHL signal synthase-encoding gene was evaluated by RT-qPCR (a) and by assessing the β-galactosidase activity of the promoter-lacZ transcriptional fusions in the wild-type and abiS mutant strains (OD600 = 3.0) (b). The production of the AHL signal in the A. baumannii wild-type strain was arbitrarily defined as 100% and used to normalize the amount of that signal in the abiS mutant and complemented strains (c). The data are the means ± standard deviations of three independent experiments. The significance of result b was determined by unpaired t test. The significance of result c was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

To determine the comprehensive regulatory roles of indole in controlling bacterial physiology and pathogenicity, we analyzed and compared the transcriptomic profiles of the wild-type strain and the ΔabiS mutant strain by using RNA sequencing. Differential gene expression analysis showed that 85 genes were upregulated and 140 genes were downregulated (log2 fold change ≥1.0) in the ΔabiS mutant strain compared with the wild-type strain. These differentially expressed genes were associated with a range of biological functions, including transport and metabolism, cell envelope biogenesis, cell motility and secretion, coenzyme metabolism, defense mechanisms, lipid metabolism, signal transduction mechanisms, transcription and translation ribosomal structure and biogenesis (Fig. S9 and Table S4 in the supplemental material). Further analysis showed that some virulence-related genes, such as epsA (A1S_0051), pgaA (A1S_2162), and plD (A1S_2099), were also downregulated in the abiS deletion mutant strain (Fig. S10).

Null mutation of AbiS decreases the competitive capability of A. baumannii against P. aeruginosa in the mixed culture.

As indole interferes with the QS systems and T3SS of P. aeruginosa and plays a vital role in the regulation of the physiology and pathogenicity of A. baumannii, we investigated whether there exist any competitive interactions between these two bacterial species, and if so, whether indole plays a role in interspecies competition. To this end, we cocultured the P. aeruginosa PAO1 strain with fluorescence labeling in the absence or presence of A. baumannii strains. The results showed that the GFP mean fluorescence intensity (MFI) of P. aeruginosa was 6.7 × 106 when it was grown alone (Fig. 8a). The value decreased to 2.83 × 106 when P. aeruginosa was cocultured with the A. baumannii wild-type strain. The GFP MFI of P. aeruginosa cocultured with the deletion mutant ΔabiS was approximately 148% that of P. aeruginosa cocultured with the A. baumannii wild-type strain at 48 h postinoculation (Fig. 8a). The results showed that deletion of abiS caused a drastic reduction in the competitive capability of A. baumannii against P. aeruginosa. We also found that deletion of abiS caused an approximately 37% reduction in mCherry fluorescence intensity when the cells were cocultured with P. aeruginosa at 48 h postinoculation compared with the A. baumannii wild-type strain (Fig. 8b). These results suggest that the abiS gene-encoded products or functions might play a key role in enhancing the competitive fitness of A. baumannii in the microbial community.

FIG 8

Null mutation of AbiS decreases the competitive capability of A. baumannii against P. aeruginosa in the mixed culture. The green fluorescent protein expression vector was carried by the P. aeruginosa PAO1 wild-type strain. The ATCC17978 wild-type, ΔabiS and ΔabiS(abiS) strains carried the mCherry fluorescent protein expression vector. The ATCC17978 wild-type, ΔabiS and ΔabiS(abiS) strains were cocultured with P. aeruginosa PAO1 at a ratio of 4:1 (vol/vol) at OD600 = 0.1. (a) Green mean fluorescence intensity of the P. aeruginosa PAO1 wild-type strain carrying the green fluorescent protein expression vector. (b) Red mean fluorescence intensity of the ATCC17978 WT, ΔabiS and ΔabiS(abiS) strains carrying the mCherry fluorescent protein expression vector. The data are the means ± standard deviations of three independent experiments. The significance was determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns = no significance).

DISCUSSION

In microbial communities, many bacteria use small diffusible signal molecules to sense local environmental conditions and coordinate multicellular behavior (41). In this study, we found that indole and its derivatives play a role in microbial ecology and chose indole as the representative metabolite, as it has been revealed to be a versatile interspecies and interkingdom signaling molecule in prokaryotic and eukaryotic communities (15, 16, 36). Interestingly, we found that indole and its derivatives from A. baumannii interferes with the T3SS and QS systems of P. aeruginosa (Fig. 3, Fig. S4 in the supplemental material), and indole promotes the competitive advantages of A. baumannii against P. aeruginosa (Fig. 8). As P. aeruginosa and A. baumannii are human pathogens that coexist in the lungs, our data not only enhance the important role of indole in interspecies communication but also reveal complicated interactions, including competition and collaboration, among human pathogens.

Previous studies showed that indole has diverse signaling roles, including in modulation of biofilm formation, virulence, stress responses and increase of epithelial-cell tight-junction resistance and decrease of inflammation indicators (17, 20, 32, 33, 42–50). In this study, in addition to playing an important role as an interspecies communication signal, indole also contributed to the physiological function of A. baumannii. Deletion of indole synthase impaired biofilm formation, motility and virulence of A. baumannii (Fig. 6). Previous studies showed that there is an AHL-based QS system, AbaI/AbaR, playing a vital role in the regulation of biofilm formation, motility and virulence in A. baumannii (27–30). As the indole signal also controls these AHL-regulated functions in A. baumannii, we then investigated whether there is a relationship between the two kinds of signaling systems. It was observed that both the expression of abaI and the production of 3-OH-C12-HSL were significantly reduced in the indole-null mutant (Fig. 7), suggesting a complex hierarchy of signaling systems in A. baumannii where indole is upstream of the AHL QS signaling system. However, the detailed regulatory mechanism of indole in A. baumannii still needs further investigation.

It has been previously reported that the indole synthase tryptophanase (TnaA) exists in E. coli. To date, this is the only indole synthase identified in bacteria (15); this enzyme reversibly converts tryptophan to indole, pyruvate and ammonia (19). In contrast to TnaA, AbiS contains an alpha/beta hydrolase fold. Indole production by TnaA in E. coli is determined by exogenous tryptophan concentrations (40).To determine the biosynthesis pathways of indole performed by AbiS, we studied the following aspects. First, overexpression of AbiS in E. coli K12 tnaA mutant strain rescued the production of indole in both 0.5 × LB and MP+ culture medium (Fig. 5), suggesting that the same precursors that could be converted to indole by AbiS were present in E. coli as in A. baumannii. Second, we found that exogenous addition of 300 μM tryptophan could not affect the production of indole in the Δ tnaA(abiS) strain but significantly increased the production of indole in the E. coli wild-type strain in 0.5 × LB (Fig. 5a), suggesting that tryptophan is not a precursor for the biosynthesis of indole by AbiS. Moreover, the ΔtnaA(abiS) strain produced the similar amount of indole in both tryptophan-rich and tryptophan-deficient media (Fig. 5b). Together, these findings suggested that the indole biosynthetic pathway performed by AbiS in A. baumannii is different from that in E. coli.

In addition, a BLAST search with the homologs of the two different indole synthases, AbiS and TnaA, revealed that their homologs are highly conserved in many bacterial species (Tables S1 and S2 in the supplemental material). Homologs of AbiS were found in Acidovorax, Acinetobacter, Alcanivorax, Marinobacter, and Noviherbaspirillum (Table S3), while homologs of TnaA were found in Acinetobacter, Aeromonas, Citrobacter, Edwardsiella, and Haemophilus (Table S2). More interestingly, some bacterial genera, including Acinetobacter, Aeromonas, Chromobacterium, Vibrio, and Grimontia, have homologs of both AbiS and TnaA (Table S3), suggesting that they may have two different biosynthetic pathways to produce indole. In general, our data suggest that the indole/AbiS system is a unique signaling mechanism that is present in various bacterial species. Our findings will trigger further investigation of the roles and mechanisms of this signaling system in diverse bacterial genera.

MATERIALS AND METHODS

Ethics statement.

This study was approved by the ethics committee of School of Pharmaceutical Sciences (Shenzhen), Sun Yat-sen University (SYSU-20200404), and all participants gave informed consent.

Bacterial strains and growth conditions.

The bacterial strains used in this study are listed in Table S5. All the bacterial strains and plasmids used in this study have been sequenced. A. baumannii, P. aeruginosa, and E. coli strains were cultured in Luria-Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, 5 g/L NaCl; pH 7.0–7.5) or LB agar (LB medium containing 15 g/L agar) at 37°C. MP minimal medium (1 L): FeSO4·7H2O, 1.25 × 10−4 g; (NH4)2SO4, 0.5 g; MgSO4·7H2O, 0.05 g; KH2PO4, 3.4 g. The pH was adjusted to 7.0. MP+ medium: MP minimal medium added with 1% of a 14 amino acid mixture (14 amino acid mixture contained alanine, arginine, aspartic acid, cysteine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, serine, threonine, and valine). The antibiotics were added to the medium according to the experimental needs, and the following antibiotics were used in this work: gentamicin (50 μg/mL), tetracycline (10 μg/mL), apramycin (100 μg/mL), kanamycin (50 μg/mL) or ampicillin (100 μg/mL). Indole, di-IEA, MIA, IAA (HPLC ≥ 99%) were dissolved in DMSO to a final concentration of 100 mM, and the solutions were added to the medium in the experiments, respectively.

Purification and structural analysis.

A. baumannii ATCC17978 cells were cultured overnight in LB medium, and the supernatant was extracted with an equal volume of ethyl acetate. After the ethyl acetate was concentrated by evaporation, the extract was dissolved in methanol and subjected to HPLC analysis on a C18 reverse-phase column (Atlantis T3 Column, 5 μm, 4.6 mm × 250 mm) and then eluted with acetonitrile-water (from 10:90 to 70:30 vol/vol) at a flow rate of 1 mL/min. The active fractions were detected, concentrated, and purified by HPLC using a semipreparative C18 reverse-phase column. Peaks were monitored using an UV detector at 210 nm.

The 1H and 13C nuclear magnetic resonance (NMR) spectra were recorded on an AVANCE III HD 400 (Temperature 298.0 K, Bruker, Billerica, MA, USA) operating at 400 MHz for 1H or 101 MHz for 13C. UHPLC-ESI-MS/MS was performed in a Shimadzu LC-30A UHPLC system (Shimadzu Corporation, Kyoto, Japan) with a Waters C18 column (1.8 μm, 150 × 2.1 mm) and a Shimadzu 8060 QQQ-MS mass spectrometer with an ESI source interface (Shimadzu Corporation, Kyoto, Japan).

Construction of in-frame deletion mutants.

Gene knockout was achieved by a highly efficient CRISPR-Cas9-based genome engineering platform in A. baumannii as previously described (51). The CRISPR-Cas9 system is a genome editing tool developed in recent years with strong DNA site-directed cleavage capability. By coupling the RecAb recombination system and the CRISPR-Cas9 genome cleavage system, a two-plasmid genome-editing system, pCasAb/pSGAb, was used for gene deletion. The 80-nt ssDNA was used for double-strand break (DSB) repair in A. baumannii ATCC 17978. Meanwhile, the target gene was integrated into the chromosome to obtain the complemented strains by using pWH1266. The resulting constructs were introduced into A. baumannii ATCC17978 deletion mutants using electroporation. The primers used for this study are listed in Table S6.

Construction of reporter strains and measurement of β-galactosidase activity.

The promoter fragments were inserted upstream of the promoterless lacZ gene in pME2-lacZ. The lacZ fusion constructs were transformed into wild-type and abiS deletion mutant strains of A. baumannii ATCC17978 by electroporation to obtain the relevant reporter strains. Two hundred μL of bacterial cultures (OD600 = 3.0) were collected to isolate the cells to determine the β-galactosidase activity as previously described (52). Each experiment was repeated three times in parallel.

Competition assays in mixed culture.

The green fluorescent protein expression vector was used and transformed into the P. aeruginosa PAO1 wild-type strain. The mCherry fluorescent protein expression vector was introduced into the ATCC17978 wild-type strain, ΔabiS deletion mutant strain and ΔabiS(abiS) complemented strain by electroporation. Then, the A. baumannii strains (OD600 = 0.1) were cocultured with P. aeruginosa (OD600 = 0.1) at an initial ratio of 4:1 at 37°C with shaking at 220 rpm for 48 h. Then, the mixed cultures were analyzed by a Spectra Max i3x multifunctional enzyme labeling instrument (Molecular Devices, CA, USA).

Statistical analysis.

Statistical analyses were performed using Prism 8 software (GraphPad). Unpaired t test between two groups, one-way analysis of variance (ANONA) or two-way analysis of variance among multiple groups were used to calculate P values. Statistical significance is indicated as follows: ns = no significance; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. All results were calculated from the average of at least three replicates.

Data availability.

Data supporting the findings of this study are available within the paper and its supplementary information files.