In vitro and in vivo screening for novel essential cell-envelope proteins in Pseudomonas aeruginosa.

The Gram-negative bacterium Pseudomonas aeruginosa represents a prototype of multi-drug resistant opportunistic pathogens for which novel therapeutic options are urgently required. In order to identify new candidates as potential drug targets, we combined large-scale transposon mutagenesis data analysis and bioinformatics predictions to retrieve a set of putative essential genes which are conserved in P. aeruginosa and predicted to encode cell envelope or secreted proteins. By generating unmarked deletion or conditional mutants, we confirmed the in vitro essentiality of two periplasmic proteins, LptH and LolA, responsible for lipopolysaccharide and lipoproteins transport to the outer membrane respectively, and confirmed that they are important for cell envelope stability. LptH was also found to be essential for P. aeruginosa ability to cause infection in different animal models. Conversely, LolA-depleted cells appeared only partially impaired in pathogenicity, indicating that this protein likely plays a less relevant role during bacterial infection. Finally, we ruled out any involvement of the other six proteins under investigation in P. aeruginosa growth, cell envelope stability and virulence. Besides proposing LptH as a very promising drug target in P. aeruginosa, this study confirms the importance of in vitro and in vivo validation of potential essential genes identified through random transposon mutagenesis.

The Gram-negative bacterium Pseudomonas aeruginosa is currently regarded as one of the most dreaded opportunistic pathogens in hospitals, and belongs to the “ESKAPE pathogens” group due to its propensity to “escape” antibiotic treatments1. P. aeruginosa is also the main cause of chronic lung infection and mortality in individuals with cystic fibrosis (CF), which is the most common fatal genetic disease in the Caucasian population2. A hallmark of P. aeruginosa infections is the life-threatening severity and poor responsiveness to currently available antibiotic therapies3. Besides its intrinsic resistance to many widely used antibiotics, that is mostly due to low membrane permeability and active drug efflux, P. aeruginosa also managed to acquire resistance via additional mechanisms, including target modification, increased expression of efflux pumps and acquisition of new drug-resistance genes by horizontal gene transfer45. As a result, pan-resistance to currently used antibiotics in clinical P. aeruginosa isolates has already been reported56. Despite the growing concern of clinicians about the very limited number of therapeutic options to fight multi-resistant P. aeruginosa strains, only a very small number of new anti-Pseudomonas drugs are currently in late stage of pre-clinical or clinical development78. In this scenario, there is an impelling need for the identification of novel drug targets and the discovery of new-generation antibiotics active against P. aeruginosa.

A whole genome in silico study has predicted that about 38% of the P. aeruginosa genome (corresponding to more than 2,100 genes) encodes for proteins that are located in the cell envelope or secreted in the extracellular milieu9. Gram-negative bacteria have a typical diderm cell organization, with a cell envelope consisting of a peptidoglycan-containing periplasmic space sandwiched between the cytoplasmic (or inner) membrane and the outer membrane, which is an asymmetric bilayer with an inner leaflet of phospholipids and an outer leaflet of lipopolysaccharide (LPS)10. Periplasmic and outer membrane proteins are mainly translocated through the cytosplasmic membrane via the general secretory (Sec) or twin-arginine translocation (TAT) pathways, that recognize target proteins by means of specific N-terminal signal peptides11.

Cell envelope proteins are involved in different cellular processes, including cell-wall assembly and stability, nutrient uptake, energy production, adherence, motility, environmental sensing, virulence, and antibiotic resistance. In the last decade, several proteomics studies focused on the characterization of the P. aeruginosa membrane and periplasmic compartments (reviewed in12), revealing that about one third of the P. aeruginosa outer membrane and periplasmic proteins cannot be assigned to any functional class on the basis of sequence homology13. Taken together, the relevance of the cell envelope for the physiology of Gram-negative bacteria and the poor functional characterization of the cell envelope sub-proteomes of P. aeruginosa suggest that the extracytoplasmic compartments of P. aeruginosa cells could represent a promising reservoir of still-unexplored protein functions to be investigated as potential targets for drug development.

Two large-scale transposon mutagenesis projects, focused on two different strains of P. aeruginosa (PAO1 and PA14), led to the identification of 335 genes shared by the two strains that were not disrupted in more than 60,000 defined transposon insertion mutants1415. A probabilistic calculation of 60,000 random insertions over 6.5 megabases (i.e. the average size of the P. aeruginosa genome) suggested that a gene 327 bp in length should have a 95% probability of being disrupted by random transposon insertion15, indicating that undisrupted genes larger than 327 bp are good candidates as potential essential genes.

Based on this information, we retrieved eight candidate essential genes of P. aeruginosa (undisrupted in the two transposon mutagenesis studies described above) that (i) are longer than 327 bp, (ii) have not previously been confirmed as essential genes in P. aeruginosa, (iii) are present in all P. aeruginosa genomes sequenced so far, and (iv) encode proteins with a predicted cell-envelope localization (Table 1). The role of the selected proteins in P. aeruginosa growth in vitro, cell envelope stability and pathogenicity in different animal models of infection was investigated in order to evaluate their suitability as potential targets for the development of novel anti-P. aeruginosa drugs.

Table 1

Selected candidate essential genes of P. aeruginosa PAO1 investigated in this study.

PAO1 ID	Gene name	PA14 ortholog	Signal peptide1	Transmembrane helix (TMH)2	Subcellular localization3	Protein function	Gene length (bp)
PA0517	nirC	PA14_06730	Yes	No	Periplasm	C-type cytochrome precursor	360
PA1645		PA14_43230	Yes	No	Non-cytoplasmic4	Hypothetical protein	408
PA1981		PA14_38880	Yes	No	Extracellular	Hypothetical protein	648
PA2614	lolA	PA14_30310	Yes	No	Periplasm	Periplasmic chaperone LolA	627
PA3786		PA14_15110	No	1 (N-terminal, in → out orientation)5	IM	Hypothetical protein	390
PA4460	lptH	PA14_57920	Yes	No	Periplasm	LPS transport protein LptA	527
PA4485		PA14_58210	Yes6	No	OM7	Conserved hypothetical protein	378
PA5126		PA14_67700	No	1 (N-terminal, in → out orientation)5	IM	Hypothetical protein	468

1Predicted by SignalP 4.1 (http://www.cbs.dtu.dk/services/SignalP/).

2Predicted by TMHMM Server 2.0 (http://www.cbs.dtu.dk/services/TMHMM/).

3Predicted by PSORTb (http://www.psort.org/psortb/). Abbreviations: OM, outer membrane; IM, inner membrane.

4According to PSORTb, this protein has an equal probability of having an IM, periplasmic, OM or extracellular localization.

5This protein is predicted to be anchored to the IM by means of a single N-terminal TMH, with the bulk of the protein exposed to the periplasmic space.

6Putative lipoprotein according to LipoP 1.0 Server (http://www.cbs.dtu.dk/services/LipoP/).

7The OM localization of this putative lipoprotein was predicted by ref. 30.

Results

Generation of deletion and conditional mutants, and in vitro growth assays

With the aim of characterizing novel P. aeruginosa proteins involved in cell envelope biogenesis and/or homeostasis, we focused our attention on eight genes that were previously proposed to be essential due to the inability to obtain transposon insertion mutants in more than 60,000 defined transposition events1415, and that encode proteins with a predicted cell envelope localization (Table 1). Five of these genes encode hypothetical proteins whose function cannot be predicted on the basis of sequence homology (www.pseudomonas.com), while the remaining genes encode a putative c-type cytochrome or homologues of the Escherichia coli LolA and LptA proteins, which are responsible for the transport of lipoproteins or LPS across the periplasm to the outer membrane, respectively1617. Due to the presence of another gene named lptA in the P. aeruginosa genome, which encodes a lysophosphatidic acid acyltransferase, the LptA homologue of P. aeruginosa has recently been renamed LptH18. A further candidate gene which fulfilled our selection criteria, PA3988 (encoding an homologue of the outer membrane protein LptE of E. coli), was not included in the screening since it is predicted to be involved in the same LPS transport pathway of LptH17.

In order to verify the essentiality of the genes of interest, we first attempted to generate unmarked in-frame deletion mutants in the reference strain P. aeruginosa PAO1. Surprisingly, we obtained a deletion mutant in six genes of interest (PA0517, PA1645, PA1981, PA3786, PA4485 and PA5126; Table 2), including all the genes encoding proteins with unpredictable function. This evidence clearly rules out that these genes are strictly essential for P. aeruginosa growth in vitro, as confirmed by the finding that all the deletion mutants showed growth yields and kinetics overall comparable to those of the wild type strain, both in Mueller-Hinton (MH) broth (Fig. 1a) and in MH agar plates (Supplementary Fig. S1), as well as in other complex (LB) or minimal media (M9 supplemented with different carbon sources) (Supplementary Fig. S2).

Table 2

Bacterial strains used in this study.

Strain	Genotype and/or relevant characteristics	Reference/source
P. aeruginosa
PAO1 (ATCC15692)	Prototroph	American type culture collection
PAO1 ΔPA0517	PAO1 with an in-frame deletion of the PA0517 gene	This study
PAO1 ΔPA1645	PAO1 with an in-frame deletion of the PA1645 gene	This study
PAO1 ΔPA1981	PAO1 with an in-frame deletion of the PA1981 gene	This study
PAO1 ΔPA3786	PAO1 with an in-frame deletion of the PA3786 gene	This study
PAO1 ΔPA4485	PAO1 with an in-frame deletion of the PA4485 gene	This study
PAO1 ΔPA5126	PAO1 with an in-frame deletion of the PA5126 gene	This study
PAO1 araCPBADlolA ΔlolA	PAO1 with an arabinose-inducible copy of lolA (PA2614) inserted into the attB neutral site and an in-frame deletion of the endogenous copy of lolA	This study
PAO1 araCPBADlptH ΔlptH	PAO1 with an arabinose-inducible copy of lptH (PA4460) inserted into the attB neutral site and an in-frame deletion of the endogenous copy of lptH	This study
E. coli
S17.1λpir	thi pro hsdR hsdM+ recA RP4-2-Tc::Mu-Km::Tn7 λpir, GmR	41
DH5αF’	recA1 endA1 hsdR17 supE44 thi-1 gyrA96 relA1 Δ(lacZYA-argF)U169[ϕ80 dlacZΔM15], NalR	42

Figure 1

Role of each protein of interest in P. aeruginosa growth in vitro.

(a) Growth curves of the wild type strain PAO1 and the PA0517, PA1645, PA1981, PA3786, PA4485 and PA5126 deletion mutants in MH broth at 37 °C in microtiter plates at 200 rpm. The generation time (Tg) of each strain is reported in the figure. Results are the mean of three independent experiments, with standard deviations (SD) being <10% of the values. (b,c) Growth of PAO1 and lptH and lolA conditional mutants in MH broth at 37 °C in microtiter plates at 200 rpm in the absence or in the presence of 0.5% arabinose (+ara), after a 1:1000 dilution from overnight cultures in MH supplemented with 0.5% arabinose. Growth was measured as OD600 (panel b) or CFU/ml (panel c). Results are the mean (±SD) of three independent experiments.

In contrast, several attempts to generate deletion mutants in lptH and lolA failed, suggesting that these two genes are important for P. aeruginosa growth and/or viability, at least under our experimental conditions. To confirm this hypothesis, we generated lptH and lolA conditional mutants carrying an arabinose-inducible copy of each gene of interest in a neutral site of the genome and an in-frame deletion in the endogenous gene (Table 2), and assessed their ability to grow in MH broth in the presence or in the absence of arabinose. The growth of both conditional mutants was strongly impaired in MH broth, but was restored to wild type levels in the presence of arabinose (Fig. 1b), confirming the relevance of both LolA and LptH for P. aeruginosa growth as well as the reliability of our conditional mutants. Notably, while the lptH conditional mutant did not grow at all in the absence of arabinose, the lolA conditional mutant showed some residual growth (Fig. 1b). This was confirmed by measuring the number of colony-forming units (CFU) over time. As shown in Fig. 1c, the lolA conditional mutant incubated in MH without arabinose was able to perform about six generations before stabilizing at a cell density about 8.5 fold lower than the wild type. In contrast, lptH mutant cells only completed a couple of generations in the absence of arabinose before beginning to die (Fig. 1c). Comparable growth profiles were obtained in different complex or minimal media (Supplementary Fig. S2). Overall, these results indicate that, in our experimental conditions, LptH is essential for both P. aeruginosa growth and cell viability, while LolA only appears to be important for growth.

Cell envelope stability

Since all the genes of interest encode proteins that are predicted to be located in the cell envelope or, in the case of PA1981, secreted into the extracellular milieu (Table 1), the possible role of these proteins in cell envelope biogenesis and/or stability was assessed through a sodium dodecyl sulphate (SDS) sensitivity assay. Deletion mutants were grown in MH, washed once with saline, and resuspended in saline in the presence of increasing concentrations of SDS. Since the expression of the LptH and LolA proteins was found to be essential for the growth of the P. aeruginosa lolA and lptH conditional mutants in vitro (Fig. 1), a previously-developed culturing strategy was employed to obtain P. aeruginosa cells depleted of each protein of interest19. Briefly, the lptH and lolA conditional mutants were grown in MH in flasks for 14 h in the presence of 0.5% (lptH mutant) or 0.1% arabinose (lolA mutant), and then two successive refreshes (1:30 dilution) were performed in the absence of arabinose. As soon as a growth defect was observed in the conditional mutants with respect to wild type cultures (dashed box in Fig. 2a), cells were collected and tested for SDS sensitivity. All deletion mutants showed an SDS sensitivity profile comparable to that of the wild type strain PAO1 (Fig. 2b). On the other hand, LptH- and LolA-deficient cells were much more sensitive to the lytic effect of SDS, although the defect in SDS resistance was more pronounced in the lptH conditional mutant compared to the lolA conditional mutant (Fig. 2b). Notably, SDS resistance was restored to wild type levels in both mutants when they were cultured in the presence of 0.5% arabinose (Supplementary Fig. S3). In the whole, this experiment demonstrates that LptH and, to a lesser extent, LolA are the only proteins here investigated with a relevant role in cell envelope stability.

Figure 2

Effect of depletion of each protein of interest on cell envelope stability.

(a) Growth of PAO1 and the lptH and lolA conditional mutants at 37 °C in MH broth at 200 rpm in flasks after two successive subcultures in the absence of arabinose, in order to obtain cells depleted of the LptH or LolA protein. Bacteria were cultured for 14 h at 37 °C and 200 rpm in MH supplemented with 0.1% (PAO1 and the lolA conditional mutant) or 0.5% arabinose (lptH conditional mutant) (not shown in the figure) and then diluted 1:30 in fresh medium without arabinose (time 0). After 3 h of growth, cultures were diluted again 1:30 in fresh medium and incubated at 37 °C until the appearance of a growth defect in the conditional mutants with respect to the wild type. (b) Lytic effect of different SDS concentrations (0–5%), measured as decrease in cell suspension turbidity (OD600), on PAO1 wild type cells, the PA0517, PA1645, PA1981, PA3786, PA4485 and PA5126 deletion mutant cells, and the LptH- or LolA-deficient conditional mutant cells (lptH and lolA, respectively) cultured as shown in panel a. The graphs are representative of three independent experiments giving similar results.

Pathogenicity in Galleria mellonella

Since laboratory cultures could not reflect bacterial growth during infection, each deletion or conditional mutant was also tested in a model of infection based on the larvae of the insect G. mellonella, which represents a convenient and easy-to-handle animal model to screen the pathogenicity of P. aeruginosa mutants20. All the deletion mutants showed lethality curves similar to that of the wild type strain (Fig. 3 and Supplementary Fig. S4), indicating that none of them was relevantly impaired in pathogenicity in this model of infection. This was confirmed by a less-than-2-fold change between the lethal dose 90% (LD90) values of mutants and wild type, with the only exception of PAO1 ΔPA1981, whose LD90 was 2.5-fold higher than that of PAO1 (Table 3). In agreement with the results of the growth assays, the lptH and lolA conditional mutants were found to be significantly impaired in infectivity in G. mellonella larvae, although a remarkable difference in the pathogenic behaviour of the two strains was observed. Indeed, while the lptH conditional mutant showed an LD90 of 1.6 × 107 cells, in line with that recently determined in the same model of infection for P. aeruginosa conditional mutant cells lacking the essential periplasmic protein TolB19, the LD90 of the lolA conditional mutant was about 60 cells, less than 25-fold higher than that of the wild type (Table 3). To verify that the observed lethality of the lolA mutant was related to active growth in vivo rather than to increased toxicity of LolA-depleted cells, we infected five G. mellonella larvae with about 500 cells of the wild type PAO1 and the lolA conditional mutant, and determined the number of viable cells in the hemolymph of dead larvae on selective plates supplemented or not with 0.5% arabinose. The number of viable cells per larva determined on arabinose-containing plates was comparable between the wild type and the lolA conditional mutant [1.6 (±0.3) × 1010 and 2.1 (±0.4) × 1010, respectively)], while the number of CFU on plates without arabinose was below the detection limit of the assay (250 cells) for lolA-infected larvae. Overall, these results indicate that, although impaired in virulence with respect to the wild type, the lolA conditional mutant is still able to grow and cause infection in G. mellonella, suggesting that LolA depletion could be less important for P. aeruginosa growth in vivo than that observed in selected laboratory media.

Figure 3

Pathogenicity of selected P. aeruginosa mutants in the G. mellonella infection model.

Survival curves, generated by the GraphPad Prism software, of G. mellonella larvae infected with different doses of P. aeruginosa PAO1, the lolA and lptH conditional mutants, and a representative deletion mutant (PAO1 ΔPA3786). The survival curves for the remaining deletion mutants are shown in Supplementary Fig. S4.

Table 3

Lethal dose 90% (LD90) in G. mellonella larvae for P. aeruginosa PAO1 and the isogenic deletion or conditional mutants analysed in this study1.

Strain	LD90	R2
PAO1	2.7	0.71
PAO1 ΔPA0517	2.3	0.78
PAO1 ΔPA1645	3.2	0.88
PAO1 ΔPA1981	6.8	0.73
PAO1 ΔPA3786	5.0	0.83
PAO1 ΔPA4485	4.8	0.78
PAO1 ΔPA5126	4.1	0.79
PAO1 araCPBADlolA ΔlolA	63.4	0.85
PAO1 araCPBADlptH ΔlptH	1.6 × 107	0.73

1The LD90 and R2 values were determined using the GraphPad Prism software and the survival curves shown in Fig. 3 or Supplementary Fig. S4.

Pathogenicity in a mouse model of pulmonary infection

Considering the different behaviour of the lptH and lolA conditional mutants in the G. mellonella infection model, we assessed their pathogenicity in a mouse model of acute lung infection21, in order to further verify their suitability as anti-P. aeruginosa drug targets. A preliminary experiment revealed that a dose corresponding to 107 cells of our wild type strain was necessary to cause 100% lethality in this infection model (Supplementary Fig. S5). We then compared the pathogenicity of the wild type and each conditional mutant at two different infecting doses (107 and 108 cells). As expected, all mice infected with 107 cells of the wild type strain died within 48 h post infection (Fig. 4a). In contrast, at the same infecting dose, the lolA conditional mutant only caused lethality in 25% of mice, while the lptH conditional mutant was completely avirulent (Fig. 4a). Moreover, a significant difference in body weight recovery after the infection was observed between mice infected with the two conditional mutants (Fig. 4b), suggesting that the challenge with the lptH conditional mutant had a less deleterious effect on mice as compared to that with the lolA conditional mutant. This was also confirmed by the observation that very high infecting doses (108 cells) caused 100% mortality within 36 h from the challenge in the case of the lolA conditional mutant, while mice infected with the lptH conditional mutant showed a significantly delayed mortality curve (Fig. 4a).

Figure 4

Pathogenicity of lptH and lolA conditional mutants in a mouse lung infection model.

(a) Mortality and (b) body weight curves for mice (n = 8) infected with 107 or 108 cells of PAO1, the lptH or the lolA conditional mutant, previously cultured in TSB supplemented with 0.5% arabinose. Data were pooled from two independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. The asterisks in panel b refer to the differences between mice infected with 107 cells of the lptH and lolA conditional mutants.

Discussion

The rise of antibacterial resistance among bacterial pathogens leads to a growing need for the identification and development of novel antibacterial agents. Notably, the recent discovery of new promising antimicrobials with a completely new mechanisms of action2223 indicates that we have only explored a fraction of the microbial targets that could be used for antibiotic drug discovery.

Rational development of antibacterial drugs with novel mechanisms of action requires the identification of new molecular targets, that may emerge from a better understanding of cellular processes essential for cell survival and/or pathogenicity. In this view, large-scale systematic analysis of gene essentiality represented an important step towards the characterization of novel potential drug targets. High density whole-genome transposon mutagenesis followed by sequence-based identification of insertion sites was the most frequently used approach to predict gene essentiality in bacterial species (http://www.essentialgene.org/). However, transposon mutagenesis suffers from intrinsic biases that can lead to misannotation of some essential or non-essential genes24. Moreover, since gene essentiality prediction by random transposon mutagenesis relies on the inability to obtain mutants in a given gene, this approach does not allow to get information about the function of each putative essential protein or its suitability as a potential antibiotic target.

The present study was specifically aimed at verifying the in vitro and in vivo essentiality of selected P. aeruginosa cell-envelope proteins, in order to propose them as novel potential drug targets. The rationale for focusing on cell envelope-located or surface-exposed proteins is linked to the fact that they are expected to be more accessible to drugs, and drug binding to cell envelope targets could avoid or delay later extrusion by efflux pumps, which represent key components of both intrinsic and acquired antibiotic resistance in P. aeruginosa25. Among the eight genes here investigated, only two were found to be required for P. aeruginosa growth in vitro, namely lptH and lolA (Fig. 1), responsible for the transport of LPS and lipoproteins to the outer membrane, respectively. In agreement with the predicted role of these proteins in outer membrane biogenesis, LptH- and LolA-depleted cells also showed increased sensitivity to the detergent SDS (Fig. 2). In contrast, the remaining genes (PA0517, PA1645, PA1981, PA3786, PA4485 and PA5126) appeared to be dispensable for P. aeruginosa growth in vitro and for cell envelope stability (Figs 1 and 2). This finding is in line with very recent Tn-seq studies reporting that transposon insertion mutants in these genes were able to grow in different growth media2627. Here we also demonstrate that the deletion of each of these genes has no relevant impact on P. aeruginosa pathogenicity in the G. mellonella model of infection (Fig. 3, Supplementary Fig. S4 and Table 3), clearly ruling out the corresponding proteins as potential targets for conventional antimicrobial drugs.

Notably, the G. mellonella infection model also revealed a different pathogenic behaviour of the lolA and lptH conditional mutants. Indeed, while LptH depletion almost completely abrogated P. aeruginosa infectivity in G. mellonella larvae, LolA depletion only resulted in a 25-fold increase in the LD90 (Fig. 3 and Table 3). A similar trend was also observed in a mouse model of lung infection, although the high infecting doses required to establish the infection in this model likely mitigated the differences between the two conditional mutants, as well as between conditional mutants and the wild type strain (Fig. 4). One could hypothesize that small amounts of arabinose possibly present in G. mellonella hemolymph or in the mouse lung could trigger sufficient lolA expression to promote growth of the arabinose-dependent conditional mutant. While to the best of our knowledge arabinose levels in G. mellonella larvae have not been investigated to date, arabinose concentration was found to be ≤0.02 mg/ml in mouse serum28 (corresponding to 0.002%), and should reasonably be even lower in the lung. This arabinose concentration is far below that required to induce growth of the lolA conditional mutant in vitro (Supplementary Fig. S6), indirectly ruling out that the observed pathogenicity in vivo may be due to arabinose-dependent restoration of LolA expression. This is further corroborated by the different pathogenic behaviour in animal models observed for the lptH conditional mutant, which carries the same arabinose-dependent regulatory element and shows an in vitro response to increasing arabinose concentrations overall comparable to the lolA mutant (Supplementary Fig. S6).

Although the inability to obtain a lolA deletion mutant does not allow to definitely conclude that LolA, even at low intracellular levels, is not required for P. aeruginosa virulence, the residual pathogenicity of the lolA conditional mutant in animal models suggests that P. aeruginosa cells are viable and can grow in vivo with an impaired transport of lipoproteins to the outer membrane. By combining all the whole-genome transposon mutagenesis data available so far1415262729, we found that only one of the 135 putative outer membrane lipoproteins of P. aeruginosa30 was predicted to be essential for in vitro growth in all experimental conditions tested. This protein corresponds to the E. coli LolB homologue (PA4668), which is responsible for lipoprotein anchoring into the outer membrane3132. However, the above-mentioned transposon mutagenesis studies also predicted lolA as an essential gene in P. aeruginosa1415262729, in agreement with our in vitro results in laboratory media (Fig. 1 and Fig. S2), suggesting that the growth conditions tested in those studies could not be suitable to appreciate the residual growth and infectivity of LolA-depleted cells that we observed in vivo. Although further studies are clearly required to characterize at the biochemical and ultrastructural level the outer membrane of LolA-depleted cells during in vivo growth, as well as to decipher the in vivo factor(s) associated with the restored growth of the lolA conditional mutant during infection, from a clinical point of view our findings argue against LolA as an ideal target for anti-P. aeruginosa drug development, since LolA inhibition could only partially affect P. aeruginosa growth and persistence during infections. Notably, while the essential role of E. coli LolA for in vitro growth is well known since many years33, and its mechanism of action has been investigated in great detail34, the infectivity of E. coli cells depleted of LolA has never been assessed in animal models. Thus, considering that our in vivo results provide evidence of partially-retained infectivity of P. aeruginosa cells impaired in lipoproteins transport, it would be interesting to verify the pathogenic role of LolA both in the model bacterium E. coli and in other Gram-negative bacteria in which lolA homologues have been predicted to be essential by in vitro large-scale mutagenesis studies (e.g., Vibrio, Haemophilus, Acinetobacter, Burkholderia, Helicobacter; http://www.essentialgene.org/).

In contrast, LptH depletion in P. aeruginosa resulted in growth inhibition and loss of cell viability in vitro (Fig. 1), as well as in almost complete abrogation of the ability to cause infection in different animal models (Table 3 and Fig. 4). Notably, an inhibitor of LPS transport in P. aeruginosa has already been reported35. This peptidomimetic blocks LPS translocation across the outer membrane by inhibiting the activity of the outer membrane protein LptD36, and was demonstrated to prevent P. aeruginosa lethality in a mouse septicemia model35. The in vitro and in vivo results obtained here with the lptH conditional mutant, together with the recently-disclosed three-dimensional structure of P. aeruginosa LptH18 and the above-mentioned pharmacological demonstration of LPS transport as a suitable drug target in P. aeruginosa35, provide strong support for further investigation of the druggability of the P. aeruginosa LptH protein.

More generally, our study highlights that experimental validation of putative essential genes, based on both in vitro and in vivo evidence, is important to confirm the results from high-throughput transposon mutagenesis, and crucial to propose novel potential molecular targets for antibacterial drug discovery programs.

Methods

Bacterial growth conditions

Bacterial strains used in this study are listed in Table 2. Unless otherwise stated, growth assays were performed at 37 °C in microtiter plates in MH, LB (Acumedia) or M9 minimal medium37 supplemented with 50 μM FeCl3 and either 20 mM succinate or 20 mM glucose as the carbon source. When required, arabinose was added to growth media at the indicated concentration. Bacterial growth was assessed by measuring the OD600 of the bacterial cultures in a Victor plate reader (Wallac). When indicated, bacterial growth was also evaluated by determining the number of CFU/ml.

Construction of plasmids, deletion and conditional mutants

Plasmids and primers used in this study are listed in Supplementary Tables S1 and S2, respectively. In-frame deletion of each gene of interest was obtained using the sacB-based suicide vector pDM4 as previously described38. pDM4 derivatives were generated by cloning ca. 500-bp long DNA fragments corresponding to the upstream and downstream genomic regions of each gene of interest in the sequencing vector pBluescript II (Stratagene). After DNA sequencing check, the upstream and downstream fragments were excised from pBluescript II and sub-cloned into pDM4. The restriction sites used for cloning are listed in Supplementary Table S2.

Conditional mutagenesis was performed using a recently-described strategy19. Briefly, a mini-CTX1 derivative carrying the coding sequence of each gene of interest under the control of an arabinose-dependent regulatory element, which includes the araC gene, the PBAD promoter and a modified ribosome binding site to reduce the basal level of expression39, was inserted into a neutral site of the P. aeruginosa genome. Then, in-frame deletion of the endogenous genes was performed by using specific pDM4 derivatives under permissive conditions (i.e. growth in the presence of 0.5% arabinose). Mini-CTX1 derivatives were generated by replacing the tolB gene in mini-CTX1-araCPBADtolB19 with the coding sequence of each gene of interest. Gene deletion and insertion events were verified by PCR and DNA sequencing.

Detergent sensitivity assay

Sensitivity to the lytic effect of SDS was assessed as previously described19, by determining the turbidity (OD600) of bacterial cell suspensions in saline after 5-min incubation at room temperature in the presence of increasing SDS concentrations (0–5%).

G. mellonella infection assay

P. aeruginosa strains were grown in MH with 0.5% arabinose, and serial dilutions of bacterial cell suspensions in saline were injected into G. mellonella larvae as described20. Larvae were incubated at 30 °C for one week to monitor mortality. Each strain was tested in at least three independent experiments. The LD90 was determined using the GraphPad Prism software as previously described40. The number of viable cells in dead larvae was determined by plating serial dilutions of the larval hemolymph in saline on MH agar plates supplemented with 100 μg/ml of ampicillin (to which P. aeruginosa is intrinsically insensitive19) and/or 0.5% arabinose.

Mouse lung infection model

C57Bl/6 mice (20–22 gr) were purchased from Charles River. Mice were housed in filtered cages under specific-pathogen conditions and permitted unlimited access to food and water. Prior to animal experiments, the parental strain P. aeruginosa PAO1 and the lptH and lolA conditional mutants were grown for 3 h to reach exponential phase in TSB with 0.5% arabinose. Next, bacterial cells were pelleted by centrifugation (2,700 × g for 15 min), washed twice with sterile PBS and the OD of the bacterial suspension was adjusted by spectrophotometry at 600 nm. Mice were anesthetized and the trachea directly visualized by a ventral midline incision, exposed and intubated with a sterile, flexible 22-g cannula attached to a 1 ml syringe according to established procedures21. A 50 μl inoculum of 107 or 108 CFU of PAO1, lptH or lolA were implanted via the cannula into the lung, with both lobes inoculated. After infection, mortality and body weight were monitored over one week. The animals were handled in compliance with European Communities Council Directive 86/609 for the care of laboratory animals and ethical guidelines for research in animals. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the San Raffaele Scientific Institute (Milan, Italy) and adhered strictly to the Italian Ministry of Health guidelines for the use and care of experimental animals.

Bioinformatics predictions

Protein subcellular localization, presence of signal peptide for protein export and of transmembrane helices (TMHs) were predicted with the PSORTb Subcellular Localization Prediction Tool (www.psort.org/psortb/), the SignalP 4.1 Server (http://www.cbs.dtu.dk/services/SignalP/) and the TMHMM Server 2.0 (http://www.cbs.dtu.dk/services/TMHMM/), respectively.

Statistical analysis

Two-way ANOVA with Bonferroni’s multiple comparison test was used to compare change in body weight. Survival curves for the mouse infection assay were analyzed using the log-rank Mantel–Cox test. Statistical analysis was performed with the software GraphPad.

Additional Information

How to cite this article: Fernández-Piñar, R. et al. In vitro and in vivo screening for novel essential cell-envelope proteins in Pseudomonas aeruginosa. Sci. Rep. 5, 17593; doi: 10.1038/srep17593 (2015).