Broad-spectrum in vitro activity of macrophage infectivity potentiator inhibitors against Gram-negative bacteria and Leishmania major.

BACKGROUND: The macrophage infectivity potentiator (Mip) protein, which belongs to the immunophilin superfamily, is a peptidyl-prolyl cis/trans isomerase (PPIase) enzyme. Mip has been shown to be important for virulence in a wide range of pathogenic microorganisms. It has previously been demonstrated that small-molecule compounds designed to target Mip from the Gram-negative bacterium Burkholderia pseudomallei bind at the site of enzymatic activity of the protein, inhibiting the in vitro activity of Mip.
OBJECTIVES: In this study, co-crystallography experiments with recombinant B. pseudomallei Mip (BpMip) protein and Mip inhibitors, biochemical analysis and computational modelling were used to predict the efficacy of lead compounds for broad-spectrum activity against other pathogens.
METHODS: Binding activity of three lead compounds targeting BpMip was verified using surface plasmon resonance spectroscopy. The determination of crystal structures of BpMip in complex with these compounds, together with molecular modelling and in vitro assays, was used to determine whether the compounds have broad-spectrum antimicrobial activity against pathogens.
RESULTS: Of the three lead small-molecule compounds, two were effective in inhibiting the PPIase activity of Mip proteins from Neisseria meningitidis, Klebsiella pneumoniae and Leishmania major. The compounds also reduced the intracellular burden of these pathogens using in vitro cell infection assays.
CONCLUSIONS: These results indicate that Mip is a novel antivirulence target that can be inhibited using small-molecule compounds that prove to be promising broad-spectrum drug candidates in vitro. Further optimization of compounds is required for in vivo evaluation and future clinical applications.

Introduction

Targeting non-essential virulence factors is a rapidly emerging approach in antimicrobial drug development.[1] In contrast to antibiotics that target essential bacterial pathways, causing bactericidal or bacteriostatic effects, this strategy aims to disarm the bacteria. This enables the host to utilize its own immune defence mechanisms to eliminate the infection, resulting in reduced selection pressure.[2] One such antivirulence target is the macrophage infectivity potentiator (Mip) protein. Mip was first characterized in Legionella pneumophila and shown to be required for optimal intracellular infection of macrophages.[3] Subsequently, Mip orthologues have been associated with virulence in several other Gram-negative bacteria (Burkholderia pseudomallei,[4]Chlamydia trachomatis,[5]Neisseria gonorrhoeae[6] and Neisseria meningitidis[7]). A Mip protein has also been shown to participate in cell invasion of the protozoan intracellular parasite Trypanosoma cruzi.[8] The role of Mip in the pathogenesis of a wide range of pathogenic organisms makes this protein a promising target for novel drug development.

Mip is an immunophilin, belonging to the FK506-binding protein (FKBP) subfamily. Some orthologues (e.g. from B. pseudomallei) consist solely of the FKBP domain; others (e.g. from L. pneumophila) have an additional dimerization domain.[9] These proteins exhibit peptidyl-prolyl isomerase (PPIase) activity.[10] This enzymatic activity can be inhibited by the potent fungal macrocyclic compounds rapamycin and FK506.[11] However, the engagement of rapamycin and FK506 with the mTOR complex and calcineurin, respectively, results in immunosuppressive effects,[12,13] rendering them unsuitable for use in antimicrobial therapy. NMR analyses of L. pneumophila Mip protein complexed with rapamycin revealed that binding to Mip is predominantly mediated via the pipecoline moiety of the drug to the PPIase domain of Mip.[14] This study led to the design, synthesis and testing of small-molecule pipecolic acid derivatives that displayed L. pneumophila Mip inhibition without exerting immunosuppressive effects.[15]

Another attractive feature of Mip as an antimicrobial target is the high sequence identity of PPIase domains across species, particularly at the FK506 binding site.[16] This formed the foundation of testing of previously designed pipecolic acid derivatives against L. pneumophila Mip against other bacterial species, such as B. pseudomallei, the causative agent of melioidosis. Indeed, a select number of pipecolic acid esters from a library of compounds were demonstrated to also be efficient at inhibiting the activity of the B. pseudomallei Mip (BpMip) protein.[17] Similarly, the broad-spectrum activity of Mip inhibitors designed to inhibit the BpMip protein was found to be effective against the Mip of neisserial and chlamydial species in vitro.[7]

Previously, the detailed design and characterization of small-molecule pipecolic acid derivative compounds that can inhibit the activity of BpMip by means of a structural biology approach has been reported.[18] These compounds displayed a low micromolar affinity to BpMip and a reduction in the in vitro activity of BpMip in both enzymatic and cell-based infection assays. This study expands on this work to reveal three lead small-molecule compounds targeting BpMip. Binding activity of all three compounds was verified using surface plasmon resonance (SPR) spectroscopy. The determination of crystal structures of BpMip in complex with these compounds, together with molecular modelling and in vitro assays, demonstrated that two of the three compounds have broad-spectrum antimicrobial activity against a range of pathogens. These include Gram-negative bacteria such as N. meningitidis and Klebsiella pneumoniae, as well as the obligate intracellular parasite Leishmania major, which belongs to the same Kinetoplastida class of organisms as T. cruzi.

Materials and methods

Recombinant protein production and purification

Recombinant BpMip (strain K96243) was produced and purified as described previously.[4] The genes encoding Mip from K. pneumoniae (KpMip; SSGCID ID KlpnA.00130.a.B2.GE39724) and L. major Friedlin (LmMip; SSGCID ID LemaA.00130.a.B1.GE34425) were cloned into a BG1861 expression vector (pET-14b derivative).[19] The gene encoding Mip from N. meningitidis (NmMip; NMB1567) was cloned into the pET28a(+) expression vector (Novagen) with the addition of a C-terminal 6 × His tag. The protein sequences of each construct are listed in Table S1, available as Supplementary data at JAC Online. Plasmid DNA was transformed into Escherichia coli strain BL21 (DE3) pLysS. KpMip, LmMip and NmMip were recombinantly expressed and purified by HisTrap, followed by size exclusion chromatography. The details are in the Supplementary methods.

Co-crystallization and structure determination of BpMip complexes

Co-crystallization and structure determination of B. pseudomallei complexes was performed as described in the Supplementary methods and by Norville et al.[17]

B. pseudomallei SPR assay

SPR experiments were performed as described in the Supplementary methods.

PPIase assay

The PPIase assay was performed as described in the Supplementary methods and by Vivoli et al.[20]

Modelling

Homology modelling was performed using YASARA v19.12.14 (YASARA Biosciences).[21] Each protein was modelled using the hm_build macro, by homology to the structures of Mips from L. pneumophila (PDB ID: 1FD9), T. cruzi (1JVW; not included for the NmMip) and B. pseudomallei (this study), and FkpA from E. coli (PDB ID: 1QVH). Sequences were aligned using YASARA and curated by hand. Template structures were superimposed using PyMOL v1.8.4.2. NmMip was modelled as a dimer. Each model was then refined using the md_refine macro. Ten independent refinements were performed for each model, and the best scoring model across all refinements was selected in each case. Docking was performed using the dock_play macro, with the active-site location defined from superposition with the B. pseudomallei structures with compounds from this study.

B. pseudomallei cytotoxicity assays

The cytotoxicity assays were performed as described in the Supplementary methods and by Begley et al.[18]

N. meningitidis intracellular survival assays

Infection of Detroit 562 epithelial cells was carried out as described in the Supplementary methods and by Reimer et al.[7]

K. pneumoniae internalization and survival assays

Murine macrophage RAW 264.7 cells (ATCC® TIB-71™) were seeded onto 24-well tissue culture plates using 1 × 106 cells in DMEM supplemented with 10% FBS, 4 mM GlutaMAX and penicillin-streptomycin (10 000 U/mL). K. pneumoniae strain ST628 was harvested in 10 mL of L-15 medium supplemented with 10% FBS and GlutaMAX to an OD590 of 0.6–0.7. The K. pneumoniae culture was incubated with each inhibitor (50 μM) or DMSO control at room temperature for 1 h. RAW 264.7 cells were infected with treated K. pneumoniae at an moi of approximately 500:1 for 15 min at 37°C following centrifugation of the culture onto the cells at 453 × g for 2 min. Post-infection, the bacteria were removed by washing cells three times with PBS and treating with 1 mL of L-15 supplemented with 1 mg/mL kanamycin for 1 h at 37°C to kill any extracellular bacteria. Infected RAW 264.7 cells were lysed using 0.01% (v/v) Triton X-100 (Sigma–Aldrich) in PBS and incubated for 15 min at room temperature. Intracellular bacteria were enumerated by viable count. All results are presented as the mean of seven independent experiments containing two technical repeats. Results are expressed as the percentage of intracellular counts following inhibitor treatment relative to the vehicle control (DMSO).

L. major intracellular survival assays

Murine macrophage RAW 264.7 cells (ATCC® TIB-71™) were seeded onto 24-well tissue culture plates using 5 × 105 cells in RPMI 1640 medium (Gibco) supplemented with 10% FBS, 4 mM GlutaMAX, 100 U/mL penicillin (Gibco) and 100 mg/mL streptomycin (Gibco) and incubated for 24 h at 36°C and 5% CO2 for 24 h to enable adherence prior to infection. L. major V121 axenic amastigote parasites were generated as per Zilberstein and Nitzan Koren et al.[22] with modifications (See Supplementary methods and the study by Zilberstein and Nitzan Koren[22]). Axenic amastigote parasites were centrifuged and resuspended at a density of 5 × 106/mL in supplemented RPMI 1640 medium and incubated with each inhibitor (50 μM) or DMSO control at 36°C and 5% CO2 for 1 h. The medium from the cultured RAW cells was removed and 1 mL of treated parasites was added and incubated further at 36°C and 5% CO2 for 3 h. Post-infection, the medium was removed from the RAW cells, washed twice with PBS and then cultured again with 1 mL of supplemented RPMI with 50 μM inhibitor for 24 h. Cells were harvested from the plate by scraping and 150 μL added to chamber funnels and centrifuged at 36 × g for 3 min onto poly-l-lysine-coated slides using a Centurion Scientific K3 cytospin. Slides were fixed in ethanol and stained using the Differential Quick stain (Amber Scientific) and visualized at 1000× magnification to check for infection. For each well, 300 randomly chosen cells were examined for infection, with percentage infected cells and the average number of parasites inside cells recorded. All results are presented as the means of nine independent experiments containing two technical repeats. Results are expressed as the percentage of cells infected with parasite.

Inhibitor cytotoxicity assays

Inhibitor-induced cytotoxicity of J774A.1 macrophage cells, Detroit 562 epithelial cells and RAW264.7 macrophage cells was measured after a 24 h incubation using the Roche LDH Cell Cytotoxicity kit (#11644793001), following the manufacturer’s instructions. All results are presented as the means of five independent experiments for J774A.1 cells and three experiments for Detroit 562 and RAW264.7 cells containing two technical repeats. A positive control (Triton X-100) and negative control (media only) were included in every assay in triplicate. Results are expressed as the percentage of lactate dehydrogenase (LDH) measured in Triton X-100 for that assay.

In vivo pharmacokinetic evaluation

In vivo pharmacokinetic evaluation was performed as described in the Supplementary methods.

Statistical analyses

All statistical analyses were carried out using Mann–Whitney U-test and analysed using GraphPad Prism v8.

Results and discussion

Mip inhibitor design and development

A library of small-molecule compounds aiming to inhibit BpMip using the structure of rapamycin, but lacking rapamycin’s immunosuppressive properties, was designed and synthesized (reviewed by Scheuplein et al.[23]). Three compounds from this library, SF235, SF339 and SF354, were selected for this study as their structures supported co-crystallization with Mip. SF235 (Figure 1a) and SF354 (Figure 1b) were synthesized as previously described (see Supplementary methods and the study by Seufert et al.[24]). In brief, a one-pot synthesis was performed starting with an amidation of nicotinic acid and the corresponding benzoic acid, respectively, under Steglich conditions, followed by an esterification with S-pipecolic acid (again using Steglich conditions). After N-deprotection, the sulphonamide was formed with benzyl sulfonyl chloride. To obtain SF339 (Figure 1c), an amidation with 3-((tert-butoxycarbonyl)amino)propanoic acid was carried out using 3,4,5-trimethoxyaniline, EDC·HCl and HOBt. The protection group was then cleaved with trifluoroacetic acid in CH2Cl2. The following amidation was performed using S-pipecolic acid, EDC·HCl and HOBt. The Boc-protection group was cleaved using trifluoroacetic acid in CH2Cl2. Subsequently, the sulphonamide was obtained by reaction of the free amine, DIPEA and equimolar amounts of benzyl sulfonyl chloride.

Figure 1.

Structures of small-molecule compounds designed to inhibit BpMip. The synthesis of SF235 (a) and SF354 (b) have been described previously.[15,24] The synthesis of SF339 is shown (c).

Co-crystallization studies of BpMip with Mip inhibitors

Co-crystallization trials were set up with BpMip and compounds SF235, SF339 and SF354. All these yielded crystals that diffracted to high enough resolution for structure determination (Figure 2). The N-terminal 6 × HisSmt tag remained attached to the protein in all three cases, with the majority of the Smt residues observed in each structure (for complete structural data see Table S2). The structure of SF235 contains four molecules per asymmetric unit, whereas the structures of SF339 and SF354 contain two molecules per asymmetric unit. In all cases the crystal lattice produces face-to-face pairs of molecules with intermolecular ligand–ligand interactions stabilizing the lattice. The pipecolic acid scaffold binding conformation previously observed[18] was retained by all three compounds with polar contacts made between the amide nitrogen of Ile63 and the carbonyl ester of all three compounds, and between the hydroxyl group of Tyr89 and one of the sulfonyl oxygens. In all three structures, we also observed the terminal phenyl ring participating in a π-edge stacking interaction with Phe43 (Figure 2).

Figure 2.

The active site of BpMip showing the interactions between the protein and the small molecules. SF235 (a) in green, SF354 (b) in magenta, SF339, Pose A (c) in cyan, and SF339, Pose B (d) in salmon. The protein is coloured light grey in all panels. Non-carbon atoms are coloured as follows: red, oxygen; blue, nitrogen; yellow, sulphur.

SF339 adopts two poses in the structure. This results in additional interactions not observed with SF235 or SF354. In both poses, while one oxygen from the sulfonyl group is contacting Tyr89 as described above, the other sulfonyl oxygen makes a hydrogen bond to the hydroxyl of Tyr33 (Figure 2c and d). The amide adjacent to the tri-methoxy phenyl group of SF339 exists in two conformations. In one conformation (Figure 2c), the carbonyl oxygen points towards Tyr89, forming an additional hydrogen bond, which is not observed in the other conformation. Additionally, we observed an interaction between the amide nitrogen of Gly97 and one of the tri-methoxy oxygens in both poses. The extra interactions of SF339 may help to explain the lower K and IC50 values compared with SF354 (Table 1). However, despite these additional interactions, the PPIase inhibitory activity is not necessarily reproduced in a cell-based infection model.

Table 1.

IC50 and K values of Mip inhibitors SF235, SF339 and SF354 against recombinant BpMip protein

Compound	IC50 (μM)		Ki (μM)
	Mean	SE	Mean	SE
SF235	0.42	0.18	0.29	0.06
SF339	0.26	0.10	0.45	0.09
SF354	0.58	0.08	0.98	0.29

SE, standard error.

BpMip binding to Mip inhibitors

Binding activity of the compounds SF235, SF339 and SF354 was investigated using BpMip and SPR spectroscopy. Data were processed using double referencing and analysed using steady-state affinity methods, resulting in binding affinities of 10.6 ± 0.2, 10.9 ± 0.2 and 8.9 ± 0.2 μM, respectively, for the three compounds. Double-referenced experimental data and steady-state affinity analyses are provided in Figure S6. Full steady-state analysis results are provided in Table S3.

Inhibition of BpMip PPIase activity by Mip inhibitors

The in vitro efficacy of SF235, SF339 and SF354 was tested using a high-throughput PPIase assay.[20] The compounds had K values of 290 ± 60, 450 ± 90 and 980 ± 290 nM, respectively (Table 1). The three compounds showed IC50 values between 260 and 580 nM, within the experimental error of K. This is consistent with competitive inhibition as expected from the structures.[25] These are in the same order as the most potent inhibitors that we previously identified (K= 160–1200 nM for compounds effective in cell-based assays).[18,20] This suggests that these inhibitors retain the potency of previous compounds.

Mip inhibitors SF235 and SF354 reduce B. pseudomallei virulence in vitro

Inhibition of B. pseudomallei-induced cytotoxicity was investigated using 50 μM SF235, SF339 and SF354 in a J774.1 macrophage cytotoxicity assay. Following infection with B. pseudomallei, treatment with SF235 and SF354 significantly reduced cytotoxicity by 25% compared with the DMSO-only control, as measured by LDH release from cells (Figure 3; P < 0.01). This reduction was similar to levels observed using a BpMip deletion mutant and previous generations of Mip inhibitors.[17] Interestingly, SF339 was less active, with no statistically significant difference detected in B. pseudomallei-induced cytotoxicity and was excluded from further study. This was despite this compound showing lower K and IC50 values than SF354 (Table 1). This highlights that PPIase inhibitory activity is not necessarily recapitulated in a cell-based infection model. Potentially, SF339 is less stable in cell infection models, resulting in faster degradation, which will be determined in future work. Nevertheless, the activity of Mip inhibitors SF235 and SF354 against B. pseudomallei highlights the potential of these two compounds as novel antibacterials for the treatment of melioidosis. Furthermore, both compounds induce low cytotoxicity when incubated with J774A.1 cells (Figure S1).

Figure 3.

Mip inhibitors SF235 and SF354 reduce B. pseudomallei-induced cytotoxicity in vitro. J774A.1 cells were infected with B. pseudomallei K96243 treated with compounds SF235, SF354 and SF339. Cytotoxicity was measured based on LDH release, and the results are presented as the percent cytotoxicity adjusted to infected cells treated with DMSO (control). **, statistically significant, P < 0.01, n = 5, Mann–Whitney U-test. Mean and each biological replicate are shown.

Computational modelling of inhibitor binding to NmMip, KpMip and LmMip proteins

To understand the potential of SF235 and SF354 as broad-spectrum Mip inhibitors, they were docked into Mip structures from several other pathogens. Mip’s role in N. meningitidis virulence has been previously identified.[7,26] The pathogens L. major and K. pneumoniae have Mips with 47%–50% amino acid sequence identity to B. pseudomallei and N. meningitidis Mips in the PPIase domain (Figure S2). The role of Mip in these pathogens’ virulence has not been investigated. The Mip proteins from N. meningitidis, L. major and K. pneumoniae were selected for modelling, followed by docking of SF235 and SF354 into the modelled enzymes. Homology models of N. meningitidis, L. major and K. pneumoniae Mips demonstrated that all three form a very similar active site to BpMip (Figure 4). SF235 and SF354 were docked into these active sites (Figure S3 and Table 2). The relative energy of docking poses similar to the structure of the compound bound to BpMip was compared with docking poses that did not reflect this structure. For LmMip, SF235 and SF354 docking poses reflecting the experimental structures had clearly higher energies than dissimilar poses. The same was true for SF235 binding to KpMip. For SF354 binding to KpMip and both ligands binding to NmMip, the highest energy docking pose reflects the ligand binding to BpMip, but there are dissimilar poses with energies that are not significantly different. This modelling suggests that SF235 and SF354 could be effective as inhibitors of Mips from all three of these pathogens. However, the modelling gives higher confidence for SF235 acting on LmMip and KpMip, and SF354 acting on LmMip.

Figure 4.

Mips from diverse species have similar active-site pockets to BpMip. The Mips from N. meningitidis, K. pneumoniae and L. major were homology modelled using YASARA. The active-site pockets were compared, focusing on amino acids within 4 Å of the position of SF235 in the BpMip structure. (a) Comparison of NmMip (magenta; oxygen atoms, red; nitrogen atoms, blue) and BpMip (yellow). NmMip has a very similar active site to BpMip, with all key amino acids binding to the compound core maintained. F53 (black arrow), on the lining of the pocket, interacts with the pyridine ring that distinguishes SF235 from other compounds in the series; this is mutated to valine (black dashed arrow), which may affect compound binding. (b) Comparison of KpMip and BpMip. The active site is again similar to BpMip. Here, the loop on the left-hand side of the pocket is drawn towards the centre of the pocket by a proline residue not found in BpMip (grey arrow); this loop interacts with the phenyl ring in the compound core, which may affect binding of this series. (c) Comparison of LmMip and BpMip. LmMip is very similar to BpMip; although F53 (orange arrow) is not conserved, a phenylalanine in a neighbouring position occupies the same space (orange dashed arrow). Proteins are shown in cartoon format with amino acids within 4 Å of SF235 shown as sticks. Image generated using the PyMOL Molecular Graphics system v.2.4.1.

Table 2.

Docking of SF235 and SF354 into Mips

Species	Ligand	Pose	Binding energy (kcal/mol)	Similar to experimental?	Similar side chain?
N. meningitidis	SF235	1	6.6 ± 0.2	Yes	No
		2	6.4 ± 0.5	No
		3	6.2 ± 0.3	Yes	Yes
		4	6.05 ± 0.08	No
		5	6.0	No
		6	5.9	No
	SF354	1	6.8 ± 0.3	Yes	No
		2	6.8 ± 0.1	Yes	No
		3	6.8 ± 0.2	Yes	Yes
		4	6.8 ± 0.3	No
		5	6.5	No
		6	6.45 ± 0.06	No
L. major	SF235	1	7.7 ± 0.9	Yes	Yes
		2	7.5 ± 0.9	Yes	No
		3	6.8 ± 0.2	No
		4	6.8	No
		5	6.7	No
		6	6.5 ± 0.2	No
		7	6.2	No
	SF354	1	6.4 ± 0.3	Yes	No
		2	6.3 ± 0.5	Yes	Yes
		3	6.1 ± 0.3	Yes	No
		4	5.8	Yes	No
		5	5.6	No
		6	5.3	No
K. pneumoniae	SF235	1	6.6	Yes	No
		2	6.4 ± 0.2	Yes	No
		3	6.2 ± 0.3	No
		4	6.1 ± 0.2	No
		5	6.0 ± 0.2	No
		6	5.9 ± 0.2	No
	SF354	1	6.3 ± 0.1	Yes	No
		2	6.3	No
		3	6.3 ± 0.2	No
		4	6.1 ± 0.2	No
		5	6.1	No

Homology models were built of the Mips from N. meningitidis, L. major and K. pneumoniae, and these and the structure of Mip from B. pseudomallei were refined by molecular dynamics. SF235 and SF354 were docked into the active site (25 poses per compound) and the preferred poses compared with the structures of B. pseudomallei bound to each compound. For each preferred pose, the calculated binding energy given is the mean of similar poses sampled, with standard deviation where multiple poses were available. Docking poses were considered to reflect the B. pseudomallei structures if the pipecolic acid and sulphate groups occupied similar spaces; and to have a similar ‘side group’ if the amide-linked side group occupies a similar space to that in the B. pseudomallei structures. Examples of each case, and of docking poses that did not reflect the experimental structures, are provided in Figure S3. All modelling was performed using YASARA.

Broad-spectrum inhibition of Mip PPIase activity by SF235 and SF354

Recombinant Mip proteins from N. meningitidis, K. pneumoniae and L. major (all including the dimerization domain) all demonstrate PPIase activity (Figure S4). SF235 (Table 3) and SF354 (Table 4) reduce the activity of Mips from all three pathogens. Consistent with the docking results, SF235 and SF354 were effective inhibitors of LmMip (K= 0.9 ± 0.2 and 0.7 ± 0.1 μM for SF235 and SF354, respectively). For NmMip, both compounds were effective, but higher concentrations of SF354 were required to achieve inhibition (K= 1.0 ± 0.2 and 2.5 ± 0.4 μM for SF235 and SF354, respectively). Both compounds could inhibit KpMip, but substantially higher concentrations were required (K= 16 ± 3 and 23 ± 4 μM for SF235 and SF354, respectively). The inhibition of each of the three proteins by SF235 and SF354 is demonstrated in Figure S5. These results confirm that SF235 particularly, and SF354 to a lesser extent, have potential as broad-spectrum inhibitors. The results are largely consistent with the docking approach. Two of the three interactions with the strongest support from docking proved to give the strongest inhibition (SF235 and SF354 binding to LmMip), whilst two of the three interactions with equivocal docking support showed weaker inhibition (SF354 binding to KpMip and NmMip). This suggests that the approach of modelling and local docking is sufficiently robust to predict Mip orthologues that are likely to be inhibited by this class of compounds.

Table 3.

IC50 and K values of Mip inhibitor SF235 against recombinant Mip proteins of N. meningitidis, K. pneumoniae and L. major

Recombinant Mip protein of:	IC50 (μM)		Ki (μM)
	Mean	SE	Mean	SE
N. meningitidis	0.9	0.2	1.0	0.2
K. pneumoniae	10	5	16	3
L. major	0.7	0.3	0.9	0.2

SE, standard error.

Table 4.

IC50 and K values of Mip inhibitor SF354 against recombinant Mip protein of N. meningitidis, K. pneumoniae and L. major

Recombinant Mip protein of:	IC50 (μM)		Ki (μM)
	Mean	SE	Mean	SE
N. meningitidis	2.7	0.6	2.5	0.4
K. pneumoniae	22	4	23	4
L. major	0.7	0.2	0.7	0.1

SE, standard error.

Mip inhibitors SF235 and SF354 decrease intracellular burden of N. meningitidis, K. pneumoniae and L. major in vitro

Previously, the inhibition of adherence, invasion and/or survival of N. meningitidis in epithelial cells in the presence of Mip inhibitors was observed.[7] However, these compounds were not suitable for therapeutic use due to low solubility. We therefore evaluated the efficacy of improved, more-soluble Mip inhibitors SF235 and SF354 against N. meningitidis using the same bacterial survival assay. This involved infecting epithelial cells with N. meningitidis pre-treated with a solvent control (DMSO), SF235 or SF354 at 50 μM and determining viable bacteria and survival levels 6 h post-infection. Consistent with Reimer et al.,[7] the results from our present study showed that both SF235 and SF354 significantly reduced the ability of N. meningitidis to invade epithelial cells to approximately 30% (Figure 5a; P < 0.01). These findings further highlight the important role of Mip in the survival of N. meningitidis in epithelial cells and efficacy of SF235 and SF354 against N. meningitidis.

Figure 5.

In vitro efficacy of Mip inhibitors SF235 and SF354 against (a) N. meningitidis, (b) K. pneumoniae and (c) L. major. N. meningitidis intracellular survival assays were performed on Detroit 562 epithelial cells and intracellular bacterial counts determined 6 h post-infection (n = 6). K. pneumoniae internalization and survival assays were performed on murine macrophage RAW 264.7 cells (n = 7). L. major intracellular survival assays were performed 3 h post-infection on RAW 264.7 cells (n = 9). *P < 0.05; **P < 0.01; ****P < 0.0001, Mann–Whitney U-test. Mean and individual biological replicates are shown.

K. pneumoniae is a Gram-negative bacterium, the MDR strains of which are in urgent need of novel antibacterial therapies for treatment of infection.[27] In particular, there has been a call for new antibiotic scaffolds, such as new small molecules with meaningful cellular activity.[28] Both SF235 and SF354 significantly reduced the intracellular growth of K. pneumoniae (Figure 5b). Following infection of RAW 264.7 macrophage cells with a clinical isolate of K. pneumoniae, a 50% reduction in the number of bacteria that were internalized and survived was observed in the presence of SF235 (Figure 5b; P < 0.001). Treatment with SF354 resulted in a similar effect (P < 0.05). These results demonstrate that Mip inhibitors have activity against K. pneumoniae, an important cause of healthcare-related infections.

In addition to bacterial species, Mips have also been identified in intracellular parasites, reinforcing their ubiquity and their relevance for invasion and virulence. Following identification of a T. cruzi Mip protein, a Mip was found to be actively secreted by the parasite and is functionally involved in host cell invasion.[8] Although Mip proteins have not been fully characterized in other intracellular parasites such as Leishmania spp., given the close evolutionary relationship between Trypanosoma and Leishmania,[29] we hypothesized that Mips may be present in Leishmania and may play a role in cell invasion and intracellular parasitism. Indeed, our studies with the pathogenic L. major species demonstrated that Mip inhibitors SF235 and SF354 significantly reduced the percentage of infected macrophage cells compared with DMSO controls (Figure 5c; P < 0.0001 and P < 0.0001, respectively). These results firstly corroborate the presence of Mip in L. major, and secondly suggest that Mip has a role in cell invasion. The stronger effect of SF235 and SF354 against L. major (compared with N. meningitidis or K. pneumoniae) may reflect either the predicted extracellular location of LmMip, or the compounds’ stronger inhibitory activity against this orthologue (Tables 3 and 4). Preliminary in vivo pharmacokinetic studies demonstrated that both SF235 and SF354 were absorbed and then metabolized within 30 min. There were no signs of toxicity (Figure S7). Further optimization of SF235 and SF354 is required for future in vivo evaluation of these compounds, including confirmation that they do not cause immunosuppressive effects.

In conclusion, this investigation demonstrates that antivirulence compounds against the Mip protein have broad-spectrum activity in Gram-negative bacteria and against the eukaryotic parasite L. major. Binding activity of three compounds with BpMip was verified using SPR spectroscopy. Based on our compound-bound crystal structures, computational modelling and biological activity, we have a strong basis for predicting a wider range of species that might be targeted by these compounds. Due to their broad-spectrum activity and low cell cytotoxicity, these compounds offer the potential as a new class of drugs against important pathogens of public health interest.

Click here for additional data file.