Scope and Limitations of Exploiting the Ability of the Chemosensitizer NV716 to Enhance the Activity of Tetracycline Derivatives against Pseudomonas aeruginosa

The spread of antibiotic resistance is an urgent threat to global health that requires new therapeutic approaches. Treatments for pathogenic Gram-negative bacteria are particularly challenging to identify due to the robust OM permeability barrier in these organisms. One strategy is to use antibiotic adjuvants, a class of drugs that have no significant antibacterial activity on their own but can act synergistically with certain antibiotics. Previous studies described the discovery and development of polyaminoisoprenyl molecules as antibiotic adjuvants with an OM effect. In particular, the compound NV716 has been shown to sensitize Pseudomonas aeruginosa to tetracycline antibiotics such as doxycycline. Here, we sought to explore the disruption of OM to sensitize P. aeruginosa to otherwise inactive antimicrobials using a series of tetracycline derivatives in the presence of NV716. We found that OM disruption expands the hydrophobicity threshold consistent with antibacterial activity to include hydrophobic molecules, thereby altering permeation rules in Gram-negative bacteria.


Introduction
There is a widely acknowledged need for new antibacterial agents to address the global increase in resistance, and this need is especially urgent to combat antibiotic-resistant Gram-negative bacteria [1,2]. In 2017, the World Health Organization (WHO) assigned the highest priority to antibacterial drug research and development against the Gramnegative bacteria Acinetobacter, Pseudomonas, and species of Enterobacteriaceae that are resistant to carbapenems and usually extensively drug-resistant (XDR) [3]. The same year, the WHO released a clinical pipeline report, which has been updated every year since then. Until 2021, 12 new antibacterial drugs have been approved. However, recently approved antibacterial agents show a limited degree of innovation as most of them are derivatives of existing classes and/or poorly address global public health priorities. The 2021 clinical antibacterial pipeline contains 77 antibiotics and/or combinations that include at least one new therapeutic entity. Of these, 45 are traditional antibacterial agents, including 27 against the WHO priority pathogens, and 32 are non-traditional [4].
The failure of the development of antibiotics active against Gram-negative bacteria is mainly due to the inability of small molecules to accumulate within these bacteria and reach a threshold concentration to inhibit their target [5,6]. Indeed, antibiotics are constrained to two opposite-direction fluxes across the Gram-negative bacterial envelope. First, Gramnegative bacteria are protected by an outer membrane (OM), which reduces the uptake of Parkinson's disease [33]. New tetracycline derivatives were recently synthesized with improved neuroprotective properties and reduced antibiotic activity [34,35]. Herein, we sought to determine whether the potential of OM disruption mediated by NV716 could be used as a therapeutic approach in combination with these new tetracycline derivatives. For this, we studied how the disturbance of the OM affects the entry rules of these compounds in P. aeruginosa and identified a significant expansion of the hydrophobicity threshold compatible activity.

OM Perturbation Induced by NV716 Increases the Range of Hydrophobicity of Tetracycline Derivatives Compatible with Antibacterial Activity
NV716 was prepared as previously described [36] in 64% yield that includes a direct nucleophilic substitution of spermine on farnesyl chloride (Figure 1). F.; & Masi, M.; in preparation). One can assume this mode of action has much more impact on P. aeruginosa than on E. coli, whose OM contains the most effective porins.
In recent years, tetracyclines have been considered beyond their antibiotic activity [32]. In particular, the use of tetracyclines is associated with a reduced risk of developing Parkinson's disease [33]. New tetracycline derivatives were recently synthesized with improved neuroprotective properties and reduced antibiotic activity [34,35]. Herein, we sought to determine whether the potential of OM disruption mediated by NV716 could be used as a therapeutic approach in combination with these new tetracycline derivatives. For this, we studied how the disturbance of the OM affects the entry rules of these compounds in P. aeruginosa and identified a significant expansion of the hydrophobicity threshold compatible activity.

OM Perturbation Induced by NV716 Increases the Range of Hydrophobicity of Tetracycline Derivatives Compatible with Antibacterial Activity
NV716 was prepared as previously described [36] in 64% yield that includes a direct nucleophilic substitution of spermine on farnesyl chloride (Figure 1). Tetracyclines inhibit protein synthesis in Gram-positive and Gram-negative bacteria by preventing the attachment of aminoacyl-tRNA to the ribosomal acceptor (A) site [37]. This mechanism has been confirmed by X-ray crystallography [38]. Original tetracycline derivatives are shown in Figure 2. The rigid skeleton of tetracyclines contains four rings, a lower non-modifiable side that makes contact with the 30S sRNA, and an upper modifiable side (Supplementary Figure S2). All tetracyclines that act as inhibitors of protein synthesis in bacteria require the amino group in position C4 (C4-dimethylamino group for optimal antibacterial activity) and keto-enolic tautomers in positions C1 and C3 of the A ring (R1 in Supplementary Figure S2). COL-3, 4-dedimethylaminosancycline, or incyclinide (3), was the first chemically modified tetracycline that has been structurally rearranged to suppress antibacterial properties by eliminating the C4-dimethylamino group in the A ring while preserving other activities of interest [39]. The D ring is the most permissive for changes. Thus, it can be hypothesized that all subsequent modifications could affect the bacterial activity due to the modification of clogD (R2-R5 in Supplementary Figure S2). Tetracyclines inhibit protein synthesis in Gram-positive and Gram-negative bacteria by preventing the attachment of aminoacyl-tRNA to the ribosomal acceptor (A) site [37]. This mechanism has been confirmed by X-ray crystallography [38]. Original tetracycline derivatives are shown in Figure 2. The rigid skeleton of tetracyclines contains four rings, a lower non-modifiable side that makes contact with the 30S sRNA, and an upper modifiable side (Supplementary Figure S2). All tetracyclines that act as inhibitors of protein synthesis in bacteria require the amino group in position C4 (C4-dimethylamino group for optimal antibacterial activity) and keto-enolic tautomers in positions C1 and C3 of the A ring (R1 in Supplementary Figure S2). COL-3, 4-dedimethylaminosancycline, or incyclinide (3), was the first chemically modified tetracycline that has been structurally rearranged to suppress antibacterial properties by eliminating the C4-dimethylamino group in the A ring while preserving other activities of interest [39]. The D ring is the most permissive for changes. Thus, it can be hypothesized that all subsequent modifications could affect the bacterial activity due to the modification of clogD (R2-R5 in Supplementary Figure S2).
We screened 28 tetracycline derivatives, including three known tetracycline antibiotics and 25 chemically modified derivatives, for their antibacterial activity in P. aeruginosa PAO1 and their degree of potentiation by the OM disruptor NV716. In the absence of NV716, the three known tetracycline antibiotics (3a-3c) showed modest activity with MICs ranging from 6.25 to 12.5 µg/mL, while all the chemically modified derivatives (3d-3ab) were completely inactive with MICs up to 200-400 µg/mL or higher ( Figure 3A and Supplementary Table S1). Of the 28 compounds tested, 18 were potentiated by NV716 with a reduction in MIC of more than 5-fold in bacteria treated with NV716 compared to untreated bacteria ( Figure 3A and Supplementary Table S1). As shown previously, NV716 at a concentration of 10 µM consistently decreased the MIC of doxycycline (3c) by 62-fold in P. aeruginosa PAO1 [27,30]. We next determined the lipophilicity (cLogD, calculated partition coefficient cLogD at pH 7.4) for all the tested compounds and ranked them according to their ability to be potentiated by the addition of NV716 ( Figure 3B). As mentioned above, the loss of antibacterial activity of COL-3 and the other derivatives is likely attributable to the removal of the C4-dimethylamino group in the A ring. This modification also increases the hydrophobicity of most of the compounds. Nevertheless, the addition of NV716 expanded the range of cLogDs compatible with antibacterial activity towards more hydrophobic compounds ( Figure 3B). Indeed, the MIC values of tetracycline derivatives with a cLogD value between −1.8 and −4.3 are either reduced to a level below the sensitivity threshold of P. aeruginosa (MIC < 2 µg/mL for compounds 3a-3c; Figure 3B, blue dots) or potentiated from 16 to 128-fold (compounds 3d-3n, Figure 3B, green dots). On the other hand, tetracycline derivatives with a cLogD value above −1.8 cannot be potentiated (compounds 3s-3ab, Figure 3B, red dots). Some compounds with a cLogD very close to −1.8 are less well potentiated, from 4 to 32-fold, in the presence of NV716 (compounds 3o-3r, Figure 3B, yellow dots). We screened 28 tetracycline derivatives, including three known tetracycline antibiotics and 25 chemically modified derivatives, for their antibacterial activity in P. aeruginosa PAO1 and their degree of potentiation by the OM disruptor NV716. In the absence of NV716, the three known tetracycline antibiotics (3a-3c) showed modest activity with MICs ranging from 6.25 to 12.5 µg/mL, while all the chemically modified derivatives (3d-3ab) were completely inactive with MICs up to 200-400 µg/mL or higher ( Figure 3A and Supplementary Table S1). Of the 28 compounds tested, 18 were potentiated by NV716 with a reduction in MIC of more than 5-fold in bacteria treated with NV716 compared to untreated bacteria ( Figure 3A and Supplementary Table S1). As shown previously, NV716 at a concentration of 10 µM consistently decreased the MIC of doxycycline (3c) by 62-fold in P. aeruginosa PAO1 [27,30]. We next determined the lipophilicity (cLogD, calculated partition coefficient cLogD at pH 7.4) for all the tested compounds and ranked them according to their ability to be potentiated by the addition of NV716 ( Figure 3B). As mentioned above, the loss of antibacterial activity of COL-3 and the other derivatives is likely attributable to the removal of the C4-dimethylamino group in the A ring. This modification also increases the hydrophobicity of most of the compounds. Nevertheless, the addition of NV716 expanded the range of cLogDs compatible with antibacterial activity towards more hydrophobic compounds ( Figure 3B). Indeed, the MIC values of tetracycline derivatives with a cLogD value between −1.8 and −4.3 are either reduced to a level below the sensitivity threshold of P. aeruginosa (MIC < 2 µg/mL for compounds 3a-3c; Figure 3B, blue dots) or potentiated from 16 to 128-fold (compounds 3d-3n, Figure 3B, green dots). On the other hand, tetracycline derivatives with a cLogD value above −1.8 cannot be potentiated (compounds 3s-3ab, Figure 3B, red dots). Some compounds with a cLogD very close to −1.8 are less well potentiated, from 4 to 32-fold, in the presence of NV716 (compounds 3o-3r, Figure 3B, yellow dots).

Antibacterial Activity of the Tetracycline Derivatives correlates with Increased Uptake in t Presence of NV716
The OM is a main barrier to the diffusion of antibiotics. Therefore, we next examin how OM perturbation by NV716 affects the accumulation of tetracycline derivatives side bacteria. For this, we chose doxycycline (3c) as a positive control, and 3j and 3u representative compounds respectively potentiated or not by NV716. It is worth noti that the chemical structure 3u is not fully represented (

Antibacterial Activity of the Tetracycline Derivatives correlates with Increased Uptake in the Presence of NV716
The OM is a main barrier to the diffusion of antibiotics. Therefore, we next examined how OM perturbation by NV716 affects the accumulation of tetracycline derivatives inside bacteria. For this, we chose doxycycline (3c) as a positive control, and 3j and 3u as representative compounds respectively potentiated or not by NV716. It is worth noting that the chemical structure 3u is not fully represented ( Figure 2) since it is patent pending in another context. This compound was chosen for this study because of its fluorescence properties and lack of antibacterial activity (i.e., cLogD is incompatible with potentiation by NV716). The fluorescence of doxycycline increases when it binds to its target (i.e., the ribosomes) inside the bacteria. We recently set up a doxycycline uptake assay with intact P. aeruginosa cells and demonstrated that NV716 causes an increase in doxycycline uptake (as shown by an increase in fluorescence, AUC × 3.6), indicating that NV716 helps doxycycline get inside the bacteria [Draveny, M.; et al., in prep.] ( Figure 4A). Here, NV716 also increased the uptake of 3j (AUC × 4.2) but not 3u, consistent with the ability of NV716 to potentiate or not their antibacterial activity ( Figure 4B). Together, these results show that disruption of OM by NV716 overcomes the intrinsic resistance of P. aeruginosa to hydrophobic compounds as it helps to increase their accumulation and activity. However, these observations could also result from a difference in the affinity of the molecules for their target. Therefore, we elucidated the inhibition of translation in an E. coli cell-free transcription-translation system using GFP as a reporter protein. Here, we found that compounds 3j and 3u similarly inhibit the production of GFP (inhibitory concentration 50%, IC 50 = 575.2 ± 1.07 µM and IC 50 = 603.3 ± 1.39 µM, respectively) ( Figure 5). Noteworthy, doxycycline was approximately 30-fold more active in inhibiting bacterial translation compared to the tetracycline derivatives (IC 50 = 19.71 ± 1.38 µM) ( Figure 5). observations could also result from a difference in the affinity of the molecules for their target. Therefore, we elucidated the inhibition of translation in an E. coli cell-free transcription-translation system using GFP as a reporter protein. Here, we found that compounds 3j and 3u similarly inhibit the production of GFP (inhibitory concentration 50%, IC50 = 575.2 ± 1.07 µM and IC50 = 603.3 ± 1.39 µM, respectively) ( Figure 5). Noteworthy, doxycycline was approximately 30-fold more active in inhibiting bacterial translation compared to the tetracycline derivatives (IC50 = 19.71 ± 1.38 µM) ( Figure 5).

Bacterial Strains and Culture Conditions
P. aeruginosa strains PAO1 and PAS263 (PAO1ΔpvdF) [40] were used for the experiments. The bacterial cultures were carried out routinely in LB-Miller medium or Mueller-Hinton 2 (MH2) at 37 °C with shaking (160 rpm).  9.7 mmol, 10 equiv) was added, and the reaction mixture was stirred at room temperature for 2 h. The resulting solution was filtered through a small pad of Celite with AcOH. The organic phase was extracted with CH 2 Cl 2 , washed with HCl (1 M) and brine, dried over MgSO4, filtered off, and concentrated in vacuo. The crude mass was 275 mg. Purification with combiflash chromatography [60 g SiO 2 , elution 0.5% (acetone + 1% formic acid) in CH 2 Cl 2 to 10% over 50 min] afforded DCT (3p) (

Potentiation Assays
MICs were conducted in PAO1 in at least three biological replicates following the CLSI protocol. The fold reduction of MIC was determined by dividing the MIC of the antibiotic alone by its MIC in the presence of 10 µM NV716.

Physicochemical Property Calculations
Structure analysis was conducted and clogD at pH 7.4 was calculated using Marvin Suite, ChemAxon (Budapest, Hungary).

Accumulation Assays
PAS263 is an isogenic mutant of PAO1 that carries a chromosomal deletion of pvdF, which encodes the pyoverdine siderophore. Accumulation assays were performed in PAS263 instead of PAO1, as pyoverdine's spectral properties overlap with those of tetracycline derivatives. Here, bacteria were grown to an exponential phase. The cultures were centrifuged at 6000× g for 20 min at 20 • C and the pellets were resuspended in NaPi-MgCl 2 buffer (50 mM sodium phosphate pH 7.4; 2 mM MgCl 2 ) to obtain 0.6 UDO/mL. Tetracycline derivatives and NV716 were used at final concentrations of 100 µM and 10 µM, respectively. For these assays, the order of the addition of the reagents is important to limit the loss of fluorescence recording at time zero. Thus, 10 µL of individual 10X tetracycline derivative solutions, 10 µL of NaPi-MgCl 2 buffer or 10X NV716, and 180 µL of bacterial suspensions were added sequentially to the wells of a 96-well microplate with black sides and a clear bottom (Thermo Fisher Scientific, ref. 165305, Illkirch-Graffenstaden, France). An additional well contains bacteria alone. The fluorescence (A. U.) of the tetracycline derivatives was immediately measured every minute for 30 min with a TECAN Spark microplate reader using specific wavelengths for excitation and emission as determined in Supplementary Figure S1. The obtained data were then processed to obtain A. U. = f(t). The fluorescence of the cells alone measured at each time was subtracted for all the conditions tested during the same experiment. Finally, the fluorescence at time zero was subtracted for each time point of each condition. The uptake of tetracycline derivatives was quantified as the area under the curve (AUC) by using GraphPad Prism 6 software. Independent experiments were performed at least three times, and the results are shown as the means of the replicates.

Bacterial Cell-Free Expression Assays
The effect of the selected tetracycline derivatives (i.e., 3c, 3j, and 3u) on in vitro bacterial protein synthesis was tested using the Expressway Cell-Free E. coli Expression System (Invitrogen for Thermo Fisher Scientific, Illkirch-Graffenstaden, France). pEXP5-CT TOPO-GFP was used as a template for in vitro transcription-translation. The reactions were performed following the manufacturer's instructions and carried out in 50 µL including the addition of feed buffer at 30 min of incubation to support optimal protein synthesis. Reactions were initiated with 1 µg of plasmid DNA; mixtures were incubated at 30 • C with shaking (300 rpm), and fluorescence was measured after 3 h with a TECAN Spark microplate reader (λex = 475 nm; λem = 520 nm). IC 50 values were calculated using GraphPad Prism 6 software. Independent experiments were performed at least three times, and the results are shown as the means of the replicates ± SD.

Conclusions
New antibacterial discoveries and developments have not kept pace with the spread of resistance. In particular, the membrane impermeability of Gram-negative bacteria represents a major challenge to overcoming the penetration and activity of both existing and new drugs. Perturbation of the OM barrier by chemical or genetic disruption has been shown to increase the sensitivity of Gram-negative bacteria to many antibiotics that are traditionally active against Gram-positive bacteria. Indeed, several groups have published proof-of-concept studies using cationic peptides [23,41], small molecules [26,42], LPS chelators [43], and genetic manipulations [13,44] that support this approach. Despite an upsurge in efforts in this area, as evidenced by reports on the global preclinical antibacterial pipeline, previous studies have mainly focused on studying the effectiveness and characterization of the mode of action of potentiating molecules taken individually [45]. However, this research field lacks studies of the strengths and limits of this approach. In 2020, Eric Brown and colleagues published a comprehensive study to analyze the interaction between OM perturbations and general constraints for effective antibiotic treatment. Interestingly, OM disruption was shown to overcome intrinsic, acquired, and spontaneous antibiotic resistance in E. coli, thus validating this approach [46]. The polyaminoisoprenyl compound NV716 was shown to revive old, disused antibiotics such as doxycycline against P. aeruginosa [27,30]. In this work, we focused on the impact of OM perturbation caused by NV716 to potentiate the activity of a new series of original tetracycline derivatives in this species. We found that OM disruption increased susceptibility towards hydrophobic compounds; this potentiation of activity was restricted to a specific range of cLogD and reflected an increase in uptake without affecting the affinity to the target.
Other mechanisms, including antibiotic inactivation, target modification, and/or overexpression of multidrug efflux pumps, are also responsible for acquired resistance in Gram-negative pathogens. Thus, we will next seek to study the efficacy of NV716 as an OM disruptor against clinical isolates of P. aeruginosa.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules28114262/s1, Figure S1: Excitation and emission spectra of compounds 3c, 3j, and 3u; Figure S2: Structure-activity relationship of tetracyclines; Table S1: List of tetracycline derivatives. Molecules in bold font (3c, 3j, and 3u) were selected for uptake assays shown in Figure 4 and/or in vitro expression assays shown in Figure 5. Funding: M.D. was supported by a Ph.D. grant from Synchrotron SOLEIL. We acknowledge France Parkinson (GAO 2018) and the "Fondation Recherche Alzheimer" for financial support for the chemical part of this study.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.