Pseudomonas aeruginosa is oxygen-deprived during infection in cystic fibrosis lungs, reducing the effectiveness of antibiotics

Abstract

Pseudomonas aeruginosa infects the lungs of patients with cystic fibrosis. Sputum expectorated from the lungs of patients contains low levels of oxygen, indicating that P. aeruginosa may be oxygen-deprived during infection. During in vitro growth under oxygen-limiting conditions, a P. aeruginosa reference strain increases expression of a cytochrome oxidase with a high affinity for oxygen, and of nitrate and nitrite reductases that enable it to use nitrate instead of oxygen during respiration. Here, we quantified transcription of the genes encoding these three enzymes in sputum samples from 18 infected patients, and in bacteria isolated from the sputum samples and grown in aerobic and anaerobic culture. In culture, expression of all three genes was increased by averages of 20- to 500-fold in anaerobically grown bacteria compared with those grown aerobically, although expression levels varied greatly between isolates. Expression of the same genes in sputum was similar to that of the corresponding bacteria in anaerobic culture. The isolated bacteria were less susceptible to tobramycin and ciprofloxacin, two widely used anti-pseudomonal antibiotics, when grown anaerobically than when grown aerobically. Our findings show that P. aeruginosa experiences oxygen starvation during infection in cystic fibrosis, reducing the effectiveness of antibiotic treatment.

Pseudomonas aeruginosa bacteria are deprived of oxygen during infection in patients with cystic fibrosis, making them more resistant to antibiotics.

Introduction

Pseudomonas aeruginosa is one of the most important opportunistic pathogens, causing a wide range of acute and chronic infections (Pendleton et al. 2013, Reynolds and Kollef 2021). Chronic P. aeruginosa infections in the lungs of people with cystic fibrosis (CF) greatly reduce the life expectancy of these individuals and have been intensively studied. In people with CF, there is a build-up of thick mucus in the lungs, making them prone to chronic infection by P. aeruginosa (Ratjen and Doring 2003, Murray et al. 2007). Acquisition of P. aeruginosa is associated with progressive clinical decline, leading to respiratory failure and reduced life expectancy (reviewed in Parkins et al. 2018). About half of adults with CF are infected by P. aeruginosa (UK Cystic Fibrosis Registry 2020, Cystic Fibrosis Foundation 2021). Infections are treated with a range of antibiotics such as ciprofloxacin (a fluoroquinolone), meropenem (a carbapenem), and tobramycin (an aminoglycoside) (Langton and Smyth 2017, McKinzie et al. 2019, Smith and Rowbotham 2022). Antibiotic therapy relieves the symptoms, but once an infection is established, existing treatment regimens usually fail to eradicate these bacteria (Kirkby et al. 2009). Antibiotic sensitivity testing of P. aeruginosa isolated from patients is not a good predictor of the clinical effectiveness of antibiotic treatment (Hurley et al. 2012, LiPuma 2022).

Antibiotic effectiveness is significantly influenced by the physiology of the target bacteria (Stokes et al. 2019). The physiology of P. aeruginosa has been well studied under laboratory conditions but is likely to be very different in the environment of the CF lung (La Rosa et al. 2019). Analysis of gene expression provides a powerful tool for understanding bacterial physiology under different growth conditions. The overall physiology of P. aeruginosa in the CF lung has been examined by analysing the expression of Pseudomonas genes in RNA extracted directly from sputum samples from CF patients using genome-wide transcriptomics, revealing significant differences from growth in standard laboratory media (Cornforth et al. 2018, Rossi et al. 2018). A similar approach, targeted to specific genes and facilitating patient–patient comparison, has been used to analyse genes involved in the metabolism of micronutrients and in antibiotic resistance (Konings et al. 2013, Nguyen et al. 2014, Martin et al. 2018, Mastropasqua et al. 2018, Madden et al. 2022). The resulting data show that there can be significant differences in P. aeruginosa gene expression in CF compared to laboratory culture, and that expression of genes associated with metal metabolism and antibiotic resistance varies greatly between patients.

A key factor in the physiology of P. aeruginosa during chronic infections in CF is the amount of oxygen available to the bacteria. In CF patients, pockets of lung tissue can become occluded and relatively inaccessible to air (Filkins and O’Toole 2015). Furthermore, plugs of mucus in the lungs are depleted for oxygen (Cowley et al. 2015). Pseudomonas aeruginosa is present within these plugs and is presumed to be existing under oxygen-poor conditions (Worlitzsch et al. 2002, Hassett et al. 2009). During infection in CF, P. aeruginosa grows in biofilms and this mode of growth can also contribute to reduced oxygen availability (Jensen et al. 2017). In the research laboratory, oxygen deprivation results in large-scale changes in gene expression that profoundly affect the physiology of the bacteria, allowing them to adapt to and survive in oxygen-deficient environments (Schobert and Jahn 2010, Trunk et al. 2010, Tata et al. 2016). They are in stark contrast to the oxygen-replete growth conditions commonly used to grow P. aeruginosa and test their antibiotic sensitivity. Although only limited studies have been carried out, there is evidence that low-oxygen conditions can affect the susceptibility of clinical isolates of P. aeruginosa to tobramycin and in some cases ciprofloxacin (Field et al. 2005, Hill et al. 2005, King et al. 2010, Shewaramani and Kassen 2022).

Genes that have altered expression during anaerobic growth in vitro include genes required for the utilization of nitrate in place of oxygen for respiration, as well as genes that contribute to other aspects of metabolism. In particular, anaerobic growth in the presence of nitrate induces upregulation of nitrate and nitrite reductases, enzymes that enable it to use nitrate ions as oxygen substitutes for energy-generating electron transfer pathways, and of a cytochrome oxidase cbb₃-2 with high affinity for oxygen (Palmer et al. 2007b, Schobert and Jahn 2010, Kamath et al. 2017). Chemical analysis has demonstrated that sputum in CF patients contains sufficient nitrate ions to support the growth of P. aeruginosa in the absence of oxygen (Palmer et al. 2007a), and there is evidence that P. aeruginosa respires by reducing nitrate in the mucus of CF patients (Kolpen et al. 2014, Rossi et al. 2018). In the model laboratory strain PAO1, the absence of oxygen causes the Anr and Dnr regulatory proteins to activate transcription of genes for denitrification and anaerobic respiration (Schreiber et al. 2007, Trunk et al. 2010). These include the narKGHJI operon that encodes nitrate reductase and the nirSMCFLGHJEN operon that encodes nitrite reductase, as well as the ccoN2O2Q2P2 operon that encodes cytochrome oxidase cbb₃-2. Expression of the nitrate and nitrite reductase genes is also upregulated by the NarXL two-component regulatory system, which is activated in the presence of nitrate (Schreiber et al. 2007). Genetic changes undergone by P. aeruginosa during long-term infection in CF facilitate adaptation to the lung environment and can also cause enhanced expression of genes for growth under anoxic conditions (Hoboth et al. 2009, Cullen and McClean 2015).

Although there is evidence that P. aeruginosa is likely to exist under anoxic or even anaerobic conditions during chronic infection in CF, the expression of nitrate and nitrite reductases and the cbb₃-2 high-affinity cytochrome oxidase during infection has not been well studied. Analysis of the transcriptomes of bacteria during infection in five CF patients found that expression of the nir, nar, and cco2 operons was upregulated relative to bacteria grown in laboratory culture (Rossi et al. 2018). Whether upregulation is a general phenomenon, and the extent to which expression varies between isolates and patients, has not been explored. In addition, understanding of the effects of oxygen deprivation on antibiotic susceptibility is incomplete. The aims of this research were to investigate the effects of anaerobiosis on the expression of oxygen-regulated genes in clinical isolates of P. aeruginosa, to determine the extent to which P. aeruginosa experiences anoxic conditions during infection in the lungs of patients with CF, and to quantify the effects of anaerobiosis on the effectiveness of three key anti-pseudomonal antibiotics.

Materials and methods

Subjects and sampling

All subjects were adults with CF attending Dunedin (New Zealand), Christchurch (New Zealand), or Hobart (Tasmania) hospitals. All of the patients were chronically infected with P. aeruginosa. Patients’ age range was 20–51 years, median 27 years (Table S1).

Sputum samples were collected under the approval of the New Zealand Health and Disability Ethics Committees (NTY/10/12/106) and the Southern Tasmanian Health and Medical Research Ethics Committee (H9813). Written informed consent was obtained for all study participants. Two portions of expectorated sputum were collected from each patient, with P. aeruginosa isolated from one portion and RNA extracted from the other, as described previously (Martin et al. 2018). Pseudomonas aeruginosa was isolated using cetrimide agar (Oxoid). For isolation of RNA, sputum was expectorated directly into RNAlater (QIAGEN) and stored at 4°C. A total of 22 sputum samples were collected. Pseudomonas aeruginosa reference strains PAO1 and ATCC27853 (Table S1) were included for comparison in some analyses.

Growth of bacteria and antibiotic sensitivity testing

For gene expression analysis, bacteria were grown at 37°C in synthetic cystic fibrosis medium (SCFM) broth (Palmer et al. 2007a) supplemented with potassium nitrate (0.4%). For aerobic growth, cultures were shaken at 200 rpm. For anaerobic growth broth, cultures and agar plates were incubated in a BD GasPak EZ incubation chamber from which the oxygen had been removed using a GasPak EZ gas generating sachet (BD Diagnostics, MD, USA) following the manufacturer’s instructions. Growth medium was solidified with agar (1.5%) as required.

Antibiotic sensitivity testing was carried out as described previously (Wiegand et al. 2008, Martin et al. 2018). Briefly, overnight cultures of P. aeruginosa that had been grown in LB broth were diluted to ∼10⁶ cfu/mL, and 5 μL portions were inoculated as spots onto plates of Mueller–Hinton Agar (Becton Dickinson) containing potassium nitrate (0.4%) and doubling dilutions of antibiotic. After the spots had dried, the plates were incubated at 37°C for 24 hours under aerobic or 48 hours under anaerobic conditions. The minimum inhibitory concentration (MIC) was defined through visual examination as the lowest concentration of antibiotic that inhibited the growth of the bacteria. MIC testing was carried out at least three times for each isolate with each antibiotic under both aerobic and anaerobic conditions. Clinical breakpoints for resistance are ciprofloxacin, ≥4 mg/L; meropenem, ≥8 mg/L; and tobramycin, ≥8 mg/L (CLSI 2018).

RT-qPCR

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was carried out as described previously (Martin et al. 2018, Konings et al. 2013). RNA was extracted from sputum samples that had been collected into RNAlater and from bacteria grown in SCFM broth supplemented with KNO₃ and incubated overnight at 37°C. Bacteria were collected into 1 mL of a 2:1 mixture of RNAprotect (QIAGEN) and Kings B broth. Following centrifugation, the bacterial cell pellet was resuspended in 200 μL of TE buffer containing lysozyme (1 mg/mL) and incubated for 5 min at room temperature to lyse the bacteria. RNA was then extracted using the RNeasy kit (QIAGEN) in conjunction with the RNase-Free DNase on-column set (QIAGEN) and PerfeCTa DNase I (Quantabio).

Aliquots of RNA were reverse transcribed to make cDNA using the qScript cDNA Synthesis Kit (Quantabio). For RT-qPCR, gene-specific primers were narI (5′-ACC ACC TGA TCA GCA CCG AG-3′ and 5′-GAG AGG CCG AGG ATC AGC TG-3′), nirS (5′-GAC CAA GGT CGC CGA GAT CA-3′ and 5′-GTT CCG GGT GGT AGG TCT GG-3′), ccoN2 (5′-CTG GCA ACT GGT GAT CCT GC-3′ and 5′-AGC TCG AGG TTG TTC ACC AC-3′), and the reference genes clpX and oprL described previously (Konings et al. 2013, Nguyen et al. 2014). The use of reference genes corrected for differences in bacterial load across different samples. Primer amplification efficiencies were determined using serial dilutions of genomic DNA templates from the reference strain PAO1 and were between 1.8 and 2.0 for all primer pairs (with all templates). Quantitative PCR was carried out using the SYBR Green I master mix in conjunction with the LightCycler 480 platform (Roche). All reactions were carried out in duplicate. The presence of correct products was confirmed by melt curve analysis and by agarose gel electrophoresis, and amplification efficiencies of each reaction were confirmed using LinRegPCR (Ruijter et al. 2009). For each sample, the expression of the genes of interest was calculated relative to the geometric mean of clpX and oprL using the LightCycler software (Konings et al. 2013). RT-qPCR for samples derived from sputum was first carried out with psd7 primers (5′-CAA AAC TAC TGA GCT AGA GTA CG-3′ and 5′-TAA GAT CTC AAG GAT CCC AAC GGC T-3′; Matsuda et al. 2007) that can detect low amounts of P. aeruginosa RNA. Samples in which the psd7 crossing point was >20 were excluded from further analysis as having insufficient RNA for RT-qPCR of the genes of interest.

Statistical analyses

Levels of gene expression were compared between pairs of outcomes for one isolate or sputum sample for each patient using Wilcoxon matched-pairs signed-ranks tests to determine if between-condition differences tended to be positive or negative. For comparison of gene expression in sputum and in laboratory bacteria, values were from bacteria isolated from the same sputum sample as the RNA. The same tests were used to compare MICs of aerobically and anaerobically grown bacteria. Analyses were performed using R 4.2.3, and two-sided P < .05 was considered statistically significant in all cases. Some values were unavailable for some sputum samples and all available data were used for each particular analysis.

Results

Effects of oxygen on expression of oxygen-starvation genes

Expression of the narI, nirS, and ccoN2 genes is induced under anaerobic conditions in model laboratory strains of P. aeruginosa (Palmer et al. 2007b, Schreiber et al. 2007, Trunk et al. 2010, Kamath et al. 2017; Table S2), and here we refer to these three genes as oxygen-starvation genes. To the best of our knowledge, the effect of anaerobiosis on the expression of these genes has not been systematically examined in isolates from patients. The effect of anaerobiosis on the expression of these genes was therefore measured in 25 isolates of P. aeruginosa from 18 different CF patients, as well as in reference strains PAO1 and ATCC27853. The bacteria were grown with and without oxygen in SCFM, which has an equivalent chemical composition to that of sputum (Palmer et al. 2007a), and that was supplemented with nitrate, which allows respiration when oxygen is unavailable. Results for one isolate from each of the patients are shown in Fig. 1, with values for each isolate in Table S2.

Figure 1.

narI, nirS, and ccoN2 are upregulated in clinical isolates of P. aeruginosa during anaerobic growth. Pseudomonas aeruginosa isolated from 18 patients with CF were grown under aerobic (in red) and anaerobic (in green) conditions and gene expression was measured. The expression value following normalization to the clpX and oprL reference genes is shown for each isolate, along with medians and interquartile ranges for each gene and condition. (A) narI, (B) nirS, and (C) ccoN2. FCM, fold change of the medians.

As expected, anaerobic growth of strain PAO1 induced expression of all three genes narI (57-fold), nirS (600-fold), and ccoN2 (21-fold). In clinical isolates, all three genes were also significantly more highly expressed (P ≤ .001) under anaerobic conditions. The nirS gene was the most highly induced (104-fold change of the medians) with the narI and ccoN genes having lower increases (9- and 11-fold, respectively). However, there was at least an 18-fold, and up to 530-fold, strain–strain variation in expression levels of the three genes under both aerobic and anaerobic conditions (Table S2). Under aerobic conditions, all the clinical isolates had higher expression of nirS and ccoN2, and all but one of narI, than PAO1 (Table S2). Under anaerobic conditions also, expression of all three genes in PAO1 was lower than the average expression in the clinical isolates. Expression in reference strain ATCC27853 tended to be nearer the average of the clinical isolates.

In a small number of cases, expression was increased <2-fold, or was not increased, by anaerobic growth. Notably, isolate U254 had lower expression of narI and nirS during anaerobic growth. The genome sequence of this isolate, and of 14 other isolates for which genome sequences were available (Table S1), was examined for differences from strain PAO1 in the sequences of the Anr, Dnr, NarL, and NarX transcriptional regulator proteins. The Anr sequence was invariant across all of the analysed genomes, whereas some variants were present in Dnr, NarX, and NarL (Table S3). The Dnr protein in strain U254 contained two unique sequence variants V35I and E205K. These variants may affect protein function and explain the lack of induction and relatively high level of expression of narI and nirS in this isolate when grown aerobically. There was no clear relationship between the other sequence variants and levels of gene expression.

Overall, these data show that anaerobic conditions induce increased expression of narI, nirS, and ccoN2 in most clinical isolates, but the level of gene expression and the extent of induction vary greatly between isolates. The narI gene had a lowest average level of expression under anaerobic conditions.

Expression of oxygen-starvation genes during infection

Analysis of RNA extracted from sputum provides a tool to investigate gene expression, and hence physiology, of P. aeruginosa as it occurs during infection in patients. We tested the hypotheses that P. aeruginosa are oxygen-deprived during infection in CF and expression of the narI, nirS, and ccoN2 genes is upregulated. RNA was extracted from sputum freshly collected from patients and gene expression was measured. Fifteen samples had quantifiable expression for at least one of the genes of interest (Table S2). Expression values from sputum were compared with those of the corresponding bacteria during in vitro culture (Fig. 2; Table S2). Expression of the narI and nirS genes was significantly higher (50–100-fold difference in the medians) for the sputum samples when compared with aerobically grown bacteria (Fig. 2 A–C), consistent with the hypothesis that P. aeruginosa are under anoxic conditions during infection. Expression of ccoN2 was also higher (1.7-fold) in the sputum samples than in the corresponding aerobically grown bacteria, although the difference was not statistically significant. Conversely, gene expression in sputum was not significantly different from that of anaerobically grown bacteria for any of the three genes (Fig. 2D–F). These data are consistent with P. aeruginosa being oxygen-deprived during chronic infection in CF, with upregulation of narI and nirS allowing use of nitrate for respiration.

Figure 2.

Expression of anoxically induced genes is higher in sputum than in aerobically grown bacteria. Expression of narI, nirS, and ccoN2 was measured in sputum samples from patients and compared with expression in laboratory-grown bacteria. The expression value following normalization to the clpX and oprL reference genes is shown for each sample, along with medians and interquartile ranges. (A–C) Comparison of sputum values (red) with those of aerobically grown bacteria (green). (D and E) Comparison of sputum values (red) with anaerobically grown bacteria (green). (A and D) narI; (B and E) nirS; and (C and F) ccoN2. FCM, fold change of the medians.

Effect of anaerobiosis on antibiotic susceptibility

There is evidence that growth under conditions of oxygen deprivation reduces the effectiveness of some antibiotics against P. aeruginosa (Field et al. 2005, Hill et al. 2005, King et al. 2010, Shewaramani and Kassen 2022). To determine whether this was true for the isolates in this study, susceptibility to three widely used anti-pseudomonal antibiotics was measured under aerobic and anaerobic conditions. The antibiotics tested were ciprofloxacin (a fluoroquinolone), meropenem (a carbapenem), and tobramycin (an aminoglycoside) all of which are used in the treatment of CF patients infected with P. aeruginosa. There was a wide range of MICs for all three antibiotics with the isolates, under both aerobic and anaerobic conditions (Fig. 3; Table S2). Pairwise comparisons showed that the median MIC for ciprofloxacin increased by 2-fold, and that for tobramycin increased by 4-fold, under anaerobic conditions. These changes were statistically significant (P = .003 and P < .001 or ciprofloxacin and tobramycin, respectively). In contrast, MICs for meropenem tended to be lower under anaerobic conditions, although this difference was not statistically significant.

Figure 3.

MICs under aerobic and anaerobic conditions. Pseudomonas aeruginosa isolated from 18 patients with CF were grown under aerobic (red) and anaerobic (green) conditions and MICs were determined for (A) ciprofloxacin, (B) meropenem, and (C) tobramycin. The MIC for each isolate is shown (y-axis), along with means and interquartile ranges. FCM, fold change of the medians.

Discussion

Our findings show that isolates of P. aeruginosa from patients with CF have a wide range of expression of terminal oxidases needed for anaerobic and anoxic respiration. However, expression of all three of the narI, nirS, and ccoN2 genes is higher in the absence of oxygen. This finding is consistent with the regulatory mechanisms characterized in the PAO1 reference strain (Schreiber et al. 2007, Trunk et al. 2010), although expression is generally higher in the clinical isolates. Expression levels of narI and nirS in infecting bacteria demonstrate that the bacteria experience anoxic, or even anaerobic, conditions during infection and can use nitrate instead of oxygen as a terminal electron acceptor. Anaerobic conditions reduce the effectiveness of tobramycin and ciprofloxacin in vitro and by implication in vivo.

Expression of narI, nirS, and ccoN2 under both aerobic and anaerobic conditions varied by over 100-fold between isolates in most cases. Nonetheless, all isolates were able to grow under anaerobic conditions, indicating that amounts of nitrate and nitrite reductases were not rate-limiting for growth. Increased production of the oxygen-dependent cbb₃-2 cytochrome oxidase under anaerobic conditions, indicated by increased expression of ccoN2, would not contribute directly to respiration as oxygen is completely absent, but this protein influences anaerobic growth, denitrification, and cell morphology under these conditions (Hamada et al. 2014). Expression of narI, nirS, and ccoN2 in PAO1 was, with one exception, lower than for all the clinical isolates under aerobic conditions and was also below the mean of the clinical isolates under anaerobic conditions. The wide range of expression between clinical isolates, the different extent to which anaerobic growth increased gene expression in isolates with identical Anr, Dnr, and NarXL proteins, and the higher expression than strain PAO1 all show that other factors contribute to levels of gene expression. Nonetheless, anr and dnr genes were present in all of the isolates studied here, and in most cases, expression of the three target genes was increased under anaerobic conditions, indicating that the regulatory mechanisms characterized in strain PAO1 are likely functional in clinical isolates. However, there were some exceptions. Notably, isolate U254, which had unique sequence variants in the dnr gene, had the highest expression of both narI and nirS under aerobic conditions and did not show increased expression of these genes under anaerobic conditions, indicating that they are constitutively expressed independent of the presence of oxygen. It may be that ongoing anoxic or anaerobic conditions during chronic infection have resulted in bacterial adaptation and genetic changes that result in constitutive expression of these genes in U254, and increased expression in the other isolates, relative to PAO1.

Only some sputum samples yielded sufficient RNA to give valid measurements of gene expression, with the amount of P. aeruginosa in other samples likely too low to enable sufficient yield of RNA. However, the data clearly showed that narI, nirS, and ccoN2 are all expressed, and hence nitrate and nitrite reductases and the cbb₃-2 cytochrome oxidase are all made during infection. Furthermore, expression of the narI and nirS genes was significantly higher in sputum than during aerobic growth in laboratory culture, and similar to growth under anaerobic conditions. This finding shows that P. aeruginosa are starved of oxygen during chronic infection in CF patients, consistent with measurements showing that there is little oxygen in sputum (Worlitzsch et al. 2002, Cowley et al. 2015). Increased expression of nitrate and nitrite reductases would allow the utilization of nitrate, which is present in sputum (Palmer et al. 2007b, Jones et al. 2000), as a terminal electron acceptor in place of oxygen. The wide range of expression levels in the sputum samples is consistent with different levels of expression for different isolates in laboratory culture but may also reflect differences in lung environment and disease severity that occur between patients (Collaco et al. 2010).

It has previously been reported that the MIC for tobramycin increases by 2-fold when P. aeruginosa is grown anaerobically (Field et al. 2005). For the isolates in this study, there was a more marked increase in tobramycin MIC (median change 4-fold) when the bacteria were grown anaerobically. In addition, the median MIC for ciprofloxacin was increased by 2-fold under anaerobic conditions for the isolates in this study. The reduced effectiveness of these antibiotics may be due at least in part to reduced entry into anaerobically growing bacterial cells. Transfer of tobramycin across the cytoplasmic membrane is an energy-dependent process dependent on the membrane potential (Webster and Shepherd 2022). The use of nitrate as a terminal electron acceptor in place of oxygen may result in a lowered membrane potential, as has been reported in Escherichia coli (Tran and Unden 1998), which would reduce the uptake of tobramycin. Ciprofloxacin enters the cytoplasm of gram-negative bacteria by diffusion (Cramariuc et al. 2012), although whether this process is influenced by the membrane potential is not yet known. Antibiotic efficacy may also be related to the production of reactive oxygen species in bacteria (Dwyer et al. 2014), which would be reduced in anoxic or anaerobic environments. Anaerobic growth did not reduce susceptibility to meropenem, which is active in the periplasm and does not need to cross the cytoplasmic membrane to have an antibacterial effect. It seems likely that anaerobic growth will also reduce the effectiveness of other antibiotics that require an energy gradient to access the cytoplasm of P. aeruginosa. It also seems likely that anoxic conditions will affect the antibiotic susceptibility of other important pathogens that can grow in both oxygen-rich and oxygen-poor environments, such as E. coli and Staphylococcus aureus, although studies of this are very limited (Harrell and Evans 1977, Gilpin et al. 2021).

The reduced effectiveness of antibiotics under anaerobic conditions, coupled with the evidence that P. aeruginosa are oxygen-starved during infection, has important implications for treatment. Antibiotic susceptibility testing is carried out under aerobic conditions, and anoxic conditions in infections may partially explain the discord between the effectiveness of antibiotics in vitro and in treating infections (Hurley et al. 2012). It will be of interest to combine a comparison of aerobic and anaerobic growth with other approaches that aim to better represent conditions of infection in CF, such as the use of artificial sputum medium (Palmer et al. 2007a), or determination of the minimum antibiotic concentration required to inhibit biofilm formation (MBIC; Macia et al. 2014). Pseudomonas aeruginosa can also chronically infect the lungs of patients with chronic obstructive pulmonary disease or non-CF bronchiectasis (Garcia-Clemente et al. 2020). In these infections also, the bacteria may be oxygen-deprived and conventional antibiotic susceptibility testing may overestimate the susceptibility of infecting bacteria to antibiotic treatment. More accurate predictions of the likely effectiveness of antibiotics in treating infections when bacteria exist under anoxic conditions might be obtained by carrying out MIC testing under anaerobic conditions. Terminal oxidases have been explored as targets for the development of novel antimicrobials (Lee et al. 2021), and the importance of nitrate and nitrate oxidases during infection suggests that these enzymes could also be potential targets for novel antibacterials.

In conclusion, our findings demonstrate that P. aeruginosa shows high expression of nitrate and nitrite reductases during infection in patients with CF, demonstrating that the bacteria exist under anoxic conditions, although the level of expression varies greatly between isolates. The altered physiology resulting from oxygen starvation reduces the effectiveness of representative antibiotics from two key classes used in treatment, indicating that anoxic conditions experienced by P. aeruginosa during infection can reduce the effectiveness of these antibiotic treatments.

Conflict of interest

The authors declare no conflict of interest.

Funding

The authors would like to thank CureKids (grant number 3574) and Cystic Fibrosis New Zealand for their support.