Introduction

Pseudomonas aeruginosa is one of the most common and persistent opportunistic pathogens affecting patients with cystic fibrosis (CF). Chronic P. aeruginosa infections lead to progressive deterioration of patient lung function, causing morbidity and mortality in CF patients [1]. P. aeruginosa is able to survive and thrive in these patients by growing as biofilms, which provide tolerance to both inflammatory defense mechanisms of the host and antibiotic therapies [2, 3]. Several clinical studies have shown that long-term treatment of CF patients with chronic P. aeruginosa infections with the macrolide antibiotic azithromycin (AZM) leads to improved lung function and increased body weight [4–7]. The median concentration of AZM in the sputum of patients receiving high-dose therapy (250 mg AZM per day) is 9.5 μg/mL, well below the minimum inhibitory concentration (MIC) for P. aeruginosa (128 to 512 μg/mL) [4, 6].

It has been suggested that one way AZM treatment improves patient lung function is by inhibiting P. aeruginosa exchange of molecules involved in cell-to-cell interactions [8, 9]. This results in a reduction of exoproducts and pathogenicity. These exoproducts include specialized metabolites (quorum sensors, virulence factors, and natural products) which are key components in the interactions between P. aeruginosa and the host. P. aeruginosa utilizes a hierarchical signaling pathway to control specialized metabolite production where the transcriptional factor pair RhlI-RhlR is subordinate to the LasI-LasR pair [10, 11]. LasI produces N-3-oxo-dodecanoyl-L-homoserine lactone (3-oxo-C12-HSL), whereas RhlI produces N-butanoyl-L-homoserine lactone (C4-HSL). These HSLs bind to the transcriptional activators LasR and RhlR, respectively, and activate target promoters. Both las and rhl have been implicated in regulating the production of a third signaling metabolite, 2-heptyl-3-hydroxy-4-quinolone [Pseudomonas quinolone signal (PQS)] [12–14]. The HSLs and PQS have been shown to regulate the expression of genes required for specialized metabolite production, including those for the siderophores pyochelin and pyoverdine, as well as the phenazine, quinolone, and rhamnolipid molecular families [15]. Herein, we describe the application of agar-based microbial matrix-assisted laser desorption ionization (MALDI) imaging mass spectrometry (IMS) in combination with liquid chromatography-tandem mass spectrometry (LC-MS/MS) to investigate the effect of AZM on specialized metabolite production of three strains of P. aeruginosa: FLR01, a non-mucoid clinical isolate, and two common laboratory strains, PAO1 and PA14.

Methods

General

P. aeruginosa strain PA14 was provided by the D. Hung lab (Harvard Medical School, MA, USA) and originated in the lab of F. M. Ausubel (Massachusetts General Hospital, MA, USA) [16]. Strain PAO1 was provided by the S. Noble lab (University of California, San Francisco, CA, USA) and originated in the lab of C. Manoil (University of Washington, Seattle, WA, USA) [17]. Non-mucoid clinical isolate FLR01 was provided by the F. Rohwer lab (San Diego State University, CA, USA). All chemicals used for LB and ISP2 media were purchased from Sigma-Aldrich (St. Louis, MO, USA). Azithromycin was purchased from Sigma-Aldrich. LC-MS grade organic solvents were purchased from J.T. Baker (Center Valley, PA, USA).

P. aeruginosa Culture

P. aeruginosa strain PA14, PAO1, and FLR01 were cultured to stationary phase (OD600 = 1.2) overnight in LB liquid medium from a single colony. The cultures were diluted to a 20% glycerol/water stock and stored at –80°C (~2.0 × 108 CFU/mL). P. aeruginosa (1 μL) was inoculated directly from glycerol stocks on ISP2 agar medium (10 mL) containing 0, 2, 4, 6, or 8 μg/mL azithromycin in 100 o.d. × 25 mm Petri dishes (Fisherbrand). Cultures were incubated at 30°C for 48 h.

MALDI IMS

MALDI IMS samples were prepared as previously described [18, 19]. Briefly, for each sample a region of agar including the colony biofilm was excised from the culture and laid on top of a Bruker MALDI MSP 96 anchor plate. To measure qualitative concentration differences in metabolite production between samples, cultures of the same strain cultured under different AZM concentration were analyzed on the same MALDI plate. A photograph was taken of the agar sections transferred to the MALDI plates. Universal matrix (Sigma-Aldrich) was applied manually using a 53 μm molecular sieve. Samples were dried at 37°C overnight. All colonies were subjected to MALDI-TOF IMS in positive reflectron mode using 500 μm spatial resolution in both X and Y dimensions by a Bruker Daltonics (Billerica, MA, USA) Microflex.

Extraction Procedure

General chemical extraction of the samples was performed. The agar from entire single plates including the colony biofilms was sliced into small pieces and extracted with 10 mL of ethyl acetate followed by 10 mL of methanol. The solvent was separated from agar pieces by filtration and concentrated in vacuo. Combined ethyl acetate-methanol extracts were resuspended in 1 mL of methanol and centrifuged prior to analysis.

LC-MS/MS Analysis

Prior to LC-MS/MS analysis, samples were diluted 10-fold. Mass spectrometry was performed using a Bruker Daltronics Maxis qTOF mass spectrometer equipped with a standard electrospray ionization source. The mass spectrometer was tuned by infusion of Tuning Mix ES-TOF (Agilent Technologies, Santa Clara, CA, USA) at a 3 μL/min flow rate. For accurate mass measurements, a wick saturated with hexakis(1H,1H,3H-tetrafluoropropoxy)phophazene ions (Synquest Laboratories, Alachua, FL, USA; m/z 922.0098) located within the source was used for lock mass internal calibration. Samples were introduced by a Thermo Scientific UltraMate 3000 Dionex ultraperformance liquid chromatograph (UPLC) using a 20-μL injection volume. Ethyl acetate-methanol extracts were separated using a Phenomenex (Torrance, CA, USA) Kinetex 2.6 μm C18 (30 × 2.10 mm) UPLC column. A linear water-acetonitrile gradient (from 98:2 to 2:98 water/acetonitrile) containing 0.1% formic acid was utilized. The flow rate was 0.5 mL/min. The mass spectrometer was operated in data-dependent positive-ion mode, automatically switching between full-scan MS and MS/MS acquisitions. Full-scan MS spectra (m/z 50 to 2000) were acquired, and the top 10 most intense ions in a particular scan were fragmented using collision-induced dissociation at 35 eV for +1 ions and 25 eV for +2 ions in the collision cell.

Data Analysis

IMS data analysis was performed using Bruker Daltronics FlexImaging v3.0. Visualization of ion intensity was optimized to highlight differences between samples being compared. Structural verification of ions putatively identified in IMS was performed manually by comparing the exact mass from the LC-MS/MS data and the corresponding MS/MS spectra with previously reported structural characterization [20, 21]. LC-MS/MS data analysis was performed using Bruker Daltronics DataAnalysis v4.1 (Build 362.7). Lock mass internal calibration using hexakis(1H,1H,3H-tetrafluoropropoxy)phosphazene ions (Synquest Laboratories; m/z 922.0098) was applied. Extracted ion chromatograms (EICs) using the exact mass of a metabolite of interest were created. The MS/MS spectra from these EICs were manually compared with previously reported data. XCMS Online was used for peak matching, non-linear retention time alignment, and quantification of mass spectral chromatographic peak areas across all samples [22]. To visualize these data, peak areas were normalized to culture growth (CFUs), fold-changes were calculated versus wild-type, log2 scaled, and visualized as a heat map in R using a previously described script [23].

Results and Discussion

Macrolides, including AZM, have been reported to suppress the production of exoproducts including exopolysaccharide, exoenzymes, and specialized metabolites through inhibition of P. aeruginosa’s signaling cascade [8, 9]. As interference with cell-to-cell signaling has been postulated as a potential beneficial effect of AZM treatment, we aimed to analyze the effect of AZM on specialized metabolite production of P. aeruginosa colony biofilms. Three P. aeruginosa strains, including the laboratory strains PAO1 and PA14 and a clinical isolate from a CF patient treated with AZM, FLR01, were grown in the presence of five different concentrations of AZM (0, 2, 4, 6, and 8 μg/mL) below the average level of AZM in the sputum of CF patients undergoing AZM therapy [4]. To visualize the effect of AZM on metabolite production, MALDI-IMS was used [18, 19]. In order to compare relative signal intensity between samples, IMS data for a single strain were collected in one experiment while data for the different strains were collected separately. As shown in Fig. 1, a wide variety of specialized metabolites were detected, including the phenazines pyocyanin (PYO), phenazine-1-carboxamide (PCN), and phenazine-1-carboxylic acid (PCA); quinolones 2-heptyl-4-quinolone (HHQ), PQS, 4-hydroxy-2-heptylquinoline-N-oxide (HQNO), 2-nonyl-4-quinolone (NHQ) and 2-nonenyl-3-hydroxy-4-quinolone (C9:db PQS), and 4-hydroxy-2-nonenylquinoline-N-oxide (C9:db NQNO) among others; various rhamnolipids as exemplified by 3-[3-[(2R,3R,4R,5R,6S)-4,5-dihydroxy-6-methyl-3-[(2S,3R,4R,5R,6S)-3,4,5-trihydroxy-6-methyloxan-2-yl]oxyoxan-2-yl]oxydecanoyloxy]decanoic acid (Rha-Rha-C10-C10) and the siderophores pyochelin and pyoverdine [20, 21].

Figure 1
figure 1

Microbial IMS images of selected metabolites produced by P. aeruginosa strains PAO1, PA14, and FLR01 treated with increasing concentrations of AZM grown on ISP2 media for 48 h. Optical images are displayed in the top row. All other images are overlays of falsely colored m/z distributions over optical images. Both PAO1 and PA14 showed decreased metabolite production with increasing AZM concentration likely due to decreased growth. Strain FLR01 metabolite production was detected at all concentrations of AZM

Full size image

With increasing concentration of AZM, both PAO1 and PA14 showed decreased growth (Fig. 1) as measured by cell forming units (CFUs). Typically, antibiotic sensitivity is measured against exponentially growing P. aeruginosa with measured MIC values of AZM up to 512 μg/mL [6]. The growth inhibition of the colony biofilms of AZM-sensitive strains (PAO1 and PA14) supports previous finding that P. aeruginosa biofilms respond to AZM at much lower concentrations, similar to P. aeruginosa liquid cultures grown to stationary phase [24–26]. It is believed that macrolides exhibit antipseudomonal activity at clinically achievable levels in a time-dependent manner against both biofilm and stationary phase P. aeruginosa by a combination of disruption of the cell membrane and inhibition of protein synthesis [27, 28]. Not surprisingly, in the IMS data for PAO1 and PA14, a decrease in observed metabolite production coincides with a reduction in CFUs. However, specialized metabolite production was detected for all concentrations of AZM administered to strain FLR01.

To more quantitatively assess specialized metabolite production, entire agar plates of replicate colony biofilms were extracted and subjected to LC-MS/MS. XCMS Online was used for peak matching, non-linear retention time alignment, and quantification of mass spectral chromatographic peak areas across all samples [22]. To visualize specialized metabolite production as a heat map, the log2 scaling of the ratio of metabolite peak areas (AZM treated to not treated) normalized to cell growth was calculated. As shown in Fig. 2, the metabolite abundance normalized to growth shows an overall decrease for lower concentrations of AZM (2 and 4 μg/mL) for PAO1 and PA14. However, in general, metabolite abundance increases for higher concentrations of AZM. This may be partially explained by the role of antibiotics as signaling metabolites [29–31]. Recent research suggests that low concentrations of antibiotics can induce specific changes in the behavior of bacteria susceptible to those antibiotics. For example, subMICs of the antibiotics tobramycin, ciprofloxacin, and tetracycline led to increased transcription levels of genes involved in biofilm formation in strain PAO1 [29, 31]. Interestingly, FLR01 does not respond to AZM in the same manner. This may be due to the fact that FLR01 shows some resistance to AZM and does not need to upregulate virulence genes to enhance survival.

Figure 2
figure 2

Heat map for the LC-MS/MS profiles of selected specialized metabolites produced by P. aeruginosa strains PAO1, PA14, and FLR01 treated with increasing concentrations of AZM grown on ISP2 media for 48 h. Blue indicates increased metabolite production with respect to untreated biofilms, whereas red indicates decreased production. Both PAO1 and PA14 showed decreased metabolite production with lower concentrations of AZM, but increased metabolite production with higher concentrations. Metabolite production remained relatively consistent for strain FLR01 regardless of AZM concentration. * Denotes not detected. # Pyoverdine is not extracted under our conditions

Full size image

Conclusion

In this work, we described the application of microbial IMS in combination with LC-MS/MS to unravel the effect of AZM treatment on specialized metabolite production of P. aeruginosa biofilms. While previous research has suggested that AZM may act as a cell-to-cell signaling inhibitor at levels below the MIC, our results show that when corrected for a decrease in cell counts, specialized metabolite production increases with increasing concentration of AZM in AZM-sensitive strains. Further investigation of the effect of common antibiotic classes at sub-inhibitory levels on P. aeruginosa specialized metabolites, both in single culture and co-culture, will help unravel the chemical response of pathogens affecting immunocompromised patients during treatment.