Retargeting Gram-Positive-Only Adarotene-Derived Antibacterials to Broad-Spectrum Antibiotics

Abstract

Background: Bacterial resistance to antibiotics continues to rise globally, posing a significant public health challenge and incurring substantial social and economic burdens. In response, the World Health Organization (WHO) has published a list of priority pathogens for which effective treatment options are critically limited. Several antibiotics are categorized as Gram-positive-only (GPO) agents due to their lack of activity against Gram-negative species. Although these compounds often target conserved bacterial processes, their limited spectrum is largely attributed to poor penetration of the Gram-negative outer membrane (OM). Results: In this study, we designed and synthesized a series of adarotene-derived compounds to evaluate the impact of introducing positively charged groups on their interaction with the Gram-negative OM. One of the newly synthesized derivatives, SPL 207, displayed minimum inhibitory concentration (MIC) values ranging from 8 to 64 µM against a panel of Gram-positive and Gram-negative bacteria. The ability of SPL207 to disrupt outer and inner membrane permeability was evaluated using fluorescence assays and confocal microscopy, revealing that the compound compromises membrane integrity across all tested Gram-negative bacteria. Strong synergistic activity was observed in combination with colistin against three P. aeruginosa colistin-resistant strains. Atomistic details of membrane interference were elucidated by molecular dynamics (MD) simulations, with SPL207 clearly acting as a membrane destabilizer by enhancing Ca²⁺ ions diffusion and lipids destabilization. Conclusions: Although the observed MIC values remain above clinically acceptable thresholds, these findings provide a promising proof of concept. The further structural optimization of adarotene derivatives may yield novel broad-spectrum agents with improved antimicrobial potency against MDR pathogens.

1. Introduction

The rising prevalence of multidrug-resistant (MDR) pathogens—and those resistant to nearly all available antibiotics—is contributing to a growing number of deaths from bacterial infections worldwide. In addition, this trend is inversely paralleled by a decline in the discovery and development of new antibiotics. It has been predicted that 10 million people worldwide will die per year from antibiotic-resistant infections by 2050 [1,2,3].

From an economic perspective, infections due to MDR pathogens lead to prolonged hospitalization and repeated antibiotic courses, causing estimated extra healthcare costs of about EUR 1.1 billion to the healthcare systems in EU/EEA countries [4]. These costs are set to rise dramatically if novel antimicrobials will not be identified. In the last two decades Gram-negative bacteria have emerged as particularly problematic pathogens. With the launch of the EU4Health 2021–2027 program, and in line with HORIZON 2020, the European Commission has confirmed the fight against communicable diseases in general and, more specifically, against antibiotic resistance as a critical priority. The World Health Organization issued a list of ‘Priority Pathogens’ for which new antibiotics are urgently needed. The vast majority of these pathogens are Gram-negative, including third-generation cephalosporin- and carbapenem-resistant Escherichia coli and Klebsiella pneumoniae, carbapenem-resistant Acinetobacter baumannii and Pseudomonas aeruginosa, and fluoroquinolone-resistant Shigella species and Salmonella enterica ser. Typhi [1,5].

The lack of Gram-negative antibacterial discovery could be traced to the unique structure of their outer membrane (OM), usually composed of lipopolysaccharides (LPS). LPS are made up of lipid A and a long, negatively charged, oligosaccharide. Lipid A molecules contain 4–7 acyl chains (predominantly saturated), attached to a phosphorylated disaccharide. The adjacent negative charges of the LPS are stabilized by divalent cations. These features allow LPS molecules to stack together very tightly, making passive diffusion extremely challenging for most small molecules [6,7]. As many compounds cannot readily passively diffuse across the OM, small molecules permeate the OM through porins and/or the self-promoted uptake (SPU) pathway. Porins are water-filled β-barrels lined interiorly with charged amino acids. They form a narrow channel allowing only certain molecules to rapidly diffuse (e.g., certain β-lactams and fluoroquinolones). Other small molecules can pass through the OM via the SPU pathway by displacing the divalent cations, temporarily destabilizing the LPS layer. Compounds that utilize SPU are typically polycationic, as for instance colistin [8]; however, some compounds likely enter through a combination of both porins and SPU pathways [6].

For the above reasons, many antibiotics are inactive against Gram-negative pathogens, hence they have been termed “Gram-positive-only” (GPO) antibiotics [6]. The spectrum of activity of GPO antibiotics can be extended to Gram-negative bacteria through the addition of chemical groups that facilitate the passage of the pharmacophore across the OM. So far, the unique marketed example of broad-spectrum antibiotic derived from a GPO antibiotic is cefiderocol. This is made of a cephalosporin pharmacophore conjugated to a catechol-type siderophore. During the infection, bacteria experience iron starvation and respond expressing siderophores. The iron-bound catechol moiety is recognized by specific bacterial receptors located on the OM, promoting cefiderocol uptake. Once in the periplasm, the cephalosporin moiety can bind to the penicillin binding protein and inhibit cell wall synthesis [9]. A similar strategy has been recently used for expanding the range of activity of aztreonam, a GPO β-lactam antibiotic [10].

In the past, extensive screening campaigns for new classes of antibiotics gave no relevant results for Gram-negative bacteria. The lack of cell permeability was recognized as a major problem, since no rationale processes existed to improve hit compounds. However, recent structure-activity relationship (SAR) studies have provided some cues useful to transform GPO into broad-spectrum antibiotics. According to these criteria, named the eNTRy rules, compounds are most likely to be accumulated in Gram-negative bacteria if they: (i) contain a non-sterically encumbered ionizable Nitrogen; (ii) have low Three dimensionality; (iii) are relatively Rigid; (iv) have some non-polar functionality [6]. In support of the validity of the eNTRy rules, promising results have been recently published about modifications of bakuchiol, a prenylated phenolic monoterpene with antibacterial activity only against Gram-positive bacteria. Bakuchiol derivatives with antibacterial activity against Gram-negative pathogens and low toxicity against eukaryotic cells were obtained by introducing an aliphatic tail containing up to 3 positively charged amine groups. This tail favors the crossing through the OM, allowing the bakuchiol derivatives to disrupt the OM and cause cell death [11].

In 2018, the ability of atypical retinoids to penetrate and embed lipid bilayers was correlated with their bactericidal activity [12]. Starting from this evidence, we have recently synthesized novel adarotene-like molecules, belonging to the class of the atypical retinoids, endowed with significant antimicrobial activity against MDR strains of Enterococcus faecalis and Staphylococcus aureus. The most promising compound (2, AB473) showed a Minimal Inhibitory Concentration (MIC) of 4 µM (2 μg/mL) for all tested strains and very low cytotoxicity on human cells (Figure 1) [13,14] suggesting a potential high therapeutic index.

However, the great limitation of the adarotene derivatives obtained so far, including AB473, is their poor activity against Gram-negative bacteria (MIC > 128 μM against E. coli, P. aeruginosa, A. baumannii) [12,13]. To overcome this limitation, we designed new adarotene derivatives by following the above-mentioned eNTRy rules [6].

In particular, we aimed to investigate whether various positively charged nitrogen species could enhance the translocation of compounds across the Gram-negative OM. We hypothesized that the incorporated basic cationic moieties could favor the interaction between the derivatives and the negatively charged bacterial cell membrane. The subsequent insertion of the hydrophobic moiety into the phospholipid bilayer could lead to the destruction of the integrity of the bacterial membrane, leakage of intracellular content and bacterial cell death. Beyond enhanced OM crossing, amino moieties could provide other advantages, including higher solubility in aqueous environments (e.g., biological fluids) [15,16]. Thus, a diverse set of amine-containing adarotene-derived compounds were designed and synthesized (Figure 2). Additionally, a shift from simple amino groups to polycationic groups was implemented, with the aim of obtaining molecules capable of penetrating Gram-negative bacteria via the SPU pathway [6].

Bearing in mind that the amino groups need to be embedded within a compound possessing appropriate flexibility, the structural features of the linker (e.g., geometry, orientation, size, and chemical composition) connecting the charged group(s) to the core skeleton, as well as the position of attachment, were carefully modulated. The installation of the selected moieties was designed not only on the biphenyl skeleton, but also on the key functional groups of the parent compound (e.g., the carboxylic group and the phenolic OH), which will serve as sites of attachment of the chains by amide and ether bonds, respectively (Figure 2). To assess membrane integrity disruption, fluorescence assays and microscopy imaging evaluation were planned for the most promising compounds identified by their MIC profiles. A validated P. aeruginosa OM model was developed to enable in silico investigations of ligand penetration. Finally, considering that colistin is one of the last-resort treatments for P. aeruginosa, we aimed to evaluate whether the most promising compounds exhibit synergistic effects against colistin-resistant strains.

2. Results and Discussion

2.1. Chemistry

Adarotene analogs and derivatives were synthesized to explore their structural and biological properties. Given the critical role of protonatable nitrogen-containing side chains in enhancing anti-Gram-negative activity, all synthesized compounds incorporated (poly)aminic residues.

Aldehydes 29a,b [14,17] were used as starting material for the preparation of benzylamines 9 and 14–17, as well as oximes 10 and 12 (Scheme 1). Reductive amination of 29a with N-methyl piperazine in presence of formic acid and BF₃Et₂O afforded derivative 9. Treatment with ethanolamine and N,N-dimethylethylenediamine gave 14 and 15, respectively; longer and more flexible chains were installed by reductive amination using diethylenetriamine and its corresponding alcohol (compound 17 and 16, respectively). Oximes 10 and 12 were prepared by treatment of the suitable aldehyde with the corresponding O-alkylated hydroxylamine in EtOH at reflux temperature. Similarly, aldehyde 30 furnished oxime 13 upon reaction with 2-(aminooxy)-ethan-1-amine hydrochloride.

Hydroxamate 11 and amide 18 were obtained from adarotene, by condensation with O-[2-(morpholin-4-yl)ethyl]hydroxylamine and N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine, respectively. Common brominated intermediate 31 was employed to introduce a polyamine chain on different portions of the biphenyl adarotene-like skeleton (Scheme 2). Heck reaction with t-butyl acrylate promptly afforded t-butyl ester 32, ready for further functionalization. Alkylation of the phenolic residue with 1,4-dibromobutane resulted in compound 33, which was alkylated and final hydrolyzed to give acid 19. Derivative 34 was obtained by Heck reaction with acrylonitrile, whose cyano group was reduced to allyl amine 7 and its saturated analog 8. Suzuki coupling between bromine 31 and phenylboronic acid afforded compound 36, which was O-alkylated with 1,4-dibromobutane and functionalized with N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine to give derivative 25. Application of the same synthetic route on 31 gave compound 23.

A similar procedure was employed for the O-alkylation of ester 38, obtained by Friedel-Crafts reaction of adamantanol on the biphenyl moiety (Scheme 3). In order to change the substituent at C(4′), treatment of boronate 39 [18] with dibromobutane (compound 40) was followed by Suzuki-Myiaura coupling with chlorobenzene or toluene, to afford 41b and 41c, respectively. Substitution of the terminal bromide with the triamine resulted in the obtainment of compounds 21, 22 and 24. Installation of the basic side chain at the phenolic oxygen was carried out also on biphenyls 42a and 42b, as well as on adamantanyl bromophenol 42c.

According to the eNTRy rules, incorporating protonatable nitrogen-containing groups into the retinoid biphenyl scaffold may facilitate the compound’s passage through the Gram-negative OM [6].

With this rationale, we initially evaluated a set of compounds (3–6) from our in-house collection, each bearing an amino group on the same carbon that carries the substituent in AB473 (Figure 2,

Table 1. In vitro antimicrobial activity of adarotene derivatives against representative Gram-positive (S. aureus) and Gram-negative (E. coli) bacteria.

Compound	MIC (µM) *
	S. aureus ATCC 25923	E. coli ATCC 25922
2	4 **	>128 **
3	>128	>128
4	64	>128
5	32	>128
6	32	>128
7	4	>128
8	8	>128
9	>128	>128
10	>128	>128
11	32	>128
12	4	>128
13	64	>128
14	8	>128
15	64	>128
16	64	>128
17	16	32
18	8	32
19	8	16
20	8	64
21	4	8
22	4	4
23	4	4
24	4	8
25	4	128
26	32	128
27	64	64
28	8	8

* Values represent the mode of at least three independent experiments; ** values taken from [14].

).

These compounds were tested against S. aureus (Gram-positive) and E. coli (Gram-negative) model strains by determining the MIC (Table 1). Although the compounds 4, 5, and 6 showed mild activity against the Gram-positive strain, they were inactive against the Gram-negative one. These results prompted us to explore the introduction of protonatable chains in other positions of the biphenyl scaffold.

We therefore evaluated adarotene analogs featuring chains that differ in length, polarity, and flexibility in place of the acrylic moiety, aiming to balance the lipophilic portions of the molecule with the protonatable groups (Figure 2).

Compound 7, having a -CH₂NH₂ group in place of the carboxylic acid moiety of adarotene (Figure 2), retained a good activity against Gram-positive (MIC = 4 µM), not showing however any activity against Gram-negative. The increase in flexibility of the chain (compound 8) did not give any improvement (MIC = 8 µM), whereas the introduction of a methylpyrazine group (compound 9) completely abolished the activity. The derivative obtained by installing a morpholino moiety through an oxime spacer (compound 10) was inactive against both S. aureus and E. coli, while a weak effect against S. aureus (MIC = 32 µM) was obtained using the hydroxamate of adarotene as a spacer for the morpholino group (compound 11). Thus, the morpholino group was replaced with a primary amine (compound 12); however, the compound remained inactive against E. coli while showing potent activity against S. aureus (MIC = 4 µM).

The removal of one of the aromatic rings (compounds 13 vs. 12) produced a further decrease in activity against the Gram-positive strain. The introduction of other secondary and tertiary amino groups (compounds 14, 15, and 16) confirmed a good activity against the Gram-positive bacterium (MIC = 8–64 µM) but still no effect against the Gram-negative one.

Interestingly, compound 17, carrying a polyamine chain in place of the acrylic moiety, showed moderate activity against E. coli, with MIC = 32 µM, retaining also activity against S. aureus (MIC = 16 µM). Moreover, the insertion of a polyamine chain through an amide bond with the carboxylic group of adarotene gave compound 18 with good activity against both E. coli and S. aureus (MIC = 32 µM and 8 µM, respectively).

The antimicrobial activity of compounds 7, 12, and 17 was also tested against other representative Gram-negative pathogens (Table S1). Compounds 7 and 12 were not active (MIC values > 128 μM), while compound 17 showed mild activity against A. baumannii, E. coli, K. pneumoniae and P. aeruginosa strains, with MIC values ranging from 64 µM to 128 µM (Table S1).

Successively, we focused our attention on the position of the polyamine moieties. We thus installed a branched protonatable chain on the phenolic OH. The acid (compound 19) showed a comparable MICs (MIC = 8 and 16 µM) against both S. aureus and E. coli, respectively, while the corresponding t-butyl ester (compound 20) showed reduced activity against E. coli (MIC = 64 µM). Antibacterial activity was improved by directly linking an ester group to the aromatic system. Notably, compound 21 exhibited good activity against S. aureus and E. coli, with MIC values of 4 µM and 8 µM, respectively. Accordingly, we evaluated the effect of introducing various substituents on the aromatic ring while keeping the rest of the molecular structure unchanged (compounds 22–25). The same range of activities of compound 21 was retained in compounds 22, 23, and 24, having a chloro, bromo, and methyl group, respectively, in place of the ester moiety.

A similar activity was maintained after the removal of one of the phenyl groups (compound 28), whereas the introduction of a third phenyl group (compound 25) abolished the activity against the Gram-negative strain (MIC = 128 µM). A further simplification of the scaffold with the removal of the adamantyl group (compounds 26, 27) caused an overall reduction in the activity, even against the Gram-positive strain.

This initial screening allowed the identification of compounds capable of inhibiting the growth of a Gram-negative bacterium such as E. coli. For this reason, the active compounds were subsequently tested against another opportunistic Gram-negative pathogen, the more clinically challenging P. aeruginosa. As reported in

Table 2. MICs of active adarotene derivatives against P. aeruginosa ATCC 27853.

MIC (µM) *
Compound	P. aeruginosa ATCC 27853
18	>128
19	128
20	128
21	64
22	128
23	32
24	128
25	>128
26	>128
27	>128
28	64

* Values represent the mode of at least three independent experiments.

, compounds 18, 19, 20, and 22 were inactive (MIC ≥ 128 µM). Compounds 21 and 28 showed a mild activity with MIC = 64 µM, while the most active compound resulted to be compound 23 (MIC = 32 µM), which was also tested against other Gram-negatives (A. baumannii, E. cloacae, E. coli MG1655, and K. pneumoniae;

Table 3. In vitro antimicrobial activity of the adarotene-derivative compound SPL207.

Strain	MIC SPL207 (µM) *
A. baumannii ATCC 19606	8
A. baumannii ACICU	8
E. cloacae ATCC 13047	16
E. coli MG1655	8
K. pneumoniae ATCC 27736	8
P. aeruginosa PAO1	64

* Values represent the mode of at least three independent experiments.

). The results showed MIC values ranging from 8 µM to 64 µM, confirming compound 23 (hereafter referred to as SPL207) as the most promising candidate for further investigation. As a preliminary assessment of potential cytotoxicity, the effect of the compound SPL207 on eukaryotic cells was tested against the human alveolar lung epithelial A549 cells and the human keratinocyte HaCaT cells (Figures S3 and S4). Compounds 7, 12 and 18 were also included. As reported in Figure S3, for the A549 cell line, the percentage of cell viability remained above 80% upon treatment with all four tested compounds up to a concentration of 16 μM. In the case of the HaCaT cell line (Figure S4), viability remained at 100% up to 16 μM for all four compounds. Treatment with compound 7 did not affect cell viability even at concentrations up to 32 μM. In any case, the concentrations at which an effect on cell viability was observed are higher than the MIC values of the selected molecules against S. aureus and most of the tested Gram-negative strains, except for P. aeruginosa.

Time-kill assays performed on A. baumannii, E. cloacae, K. pneumoniae, and P. aeruginosa cultures showed a drastic drop of CFU counts after treatment with SPL207 at 2 × MIC and 4 × MIC (Figure 3). Hence, SPL207 displays bactericidal activity.

2.2. Biology and Computational Analysis

2.2.1. Investigation About P. aeruginosa Resistance to SPL207

We reasoned that the antimicrobial activity of SPL207 against P. aeruginosa could be mainly weakened by efflux pumps capable of extruding this compound outside of the cell [19] and/or by low permeability of the OM [20,21]. To test these hypotheses, the activity of SPL207 was assessed in P. aeruginosa mutant strains deleted in efflux pump genes or defective in OM biogenesis.

Since the efflux pumps MexAB-OprM, MexXY, MexCD-OprJ, and MexEF-OprN play a key role in P. aeruginosa resistance to many antimicrobial agents [19,22], MIC assays were performed on PAO1-KP ∆efflux, a P. aeruginosa PAO1 mutant with deletion of the mexAB-oprM, mexXY, mexCD-oprJ, and mexEF-oprN operons [23,24]. As reported in

Table 4. In vitro antimicrobial activity of SPL207 against P. aeruginosa PAO1 strains grown in MHB-II without or with EDTA (mM) or arabinose (%).

Strain	Growth Medium Supplemented with	MIC SPL207 (µM) *
PAO1-KP	-	64
PAO1-KP ∆efflux	-	64
PAO1 lptE	0.002% arabinose	16
	0.5% arabinose	64
PAO1 lptH	0.125% arabinose	16
	0.5% arabinose	64
PAO1	-	64
	0.4 mM EDTA	32
	0.8 mM EDTA	16

* Values represent the mode of at least three independent experiments.

, SPL207 showed the same MIC in PAO1-KP ∆efflux and its isogenic wild-type strain PAO1-KP, indicating that efflux pumps are not responsible for poor activity of SPL207 in P. aeruginosa.

MIC assays were also performed in a P. aeruginosa PAO1 mutant strains with increased OM permeability. These strains are conditional mutants in which lptE or lptH expression can be induced by arabinose supplementation. LptE and LptH proteins are involved in the transport of lipopolysaccharides to the OM, hence low expression of lptE or lptH affects OM biogenesis and increases OM permeability [25,26].

Preliminary experiments were performed to identify the concentrations of arabinose that support growth of the lptE and lptH conditional mutants. In agreement with literature data [25,26], in the absence of arabinose P. aeruginosa growth was slightly reduced or abolished in the lptE and lptH conditional mutants, respectively, while arabinose supplementation did not affect PAO1 wild-type growth rate (Figure S1A,B). As shown in Table 4, the antimicrobial activity of SPL207 increased when the lptE and lptH conditional mutants were grown in the presence of the lowest arabinose concentrations required to minimize their growth defect (i.e., 0.002% [w/v] and 0.125% [w/v] arabinose for the lptE and lptH conditional mutants, respectively). Conversely, MIC was restored to wild type levels when the conditional mutants were grown with high arabinose concentration (i.e., 0.5% [w/v] for both conditional mutants). Interestingly, increased activity of SPL207 in P. aeruginosa was observed also in the presence of ethylenediaminetetraacetic acid (EDTA), a divalent cation chelating agent that increases OM permeability [27] (Table 4 and Figure S1C).

These results suggest that low OM permeability of P. aeruginosa could account for the limited activity of SPL207 against this pathogen.

2.2.2. Investigation of SPL207 Mechanism of Action and Synergy with Colistin

Previous studies showed that adarotene derivatives can penetrate lipid bilayers and permeabilize the membranes of Gram-positives, ultimately causing cell lysis [12,14]. To investigate whether SPL207 shares a similar mechanism of action in Gram-negative bacteria, its effect on membrane integrity was evaluated in A. baumannii, E. cloacae, K. pneumoniae, and P. aeruginosa. Briefly, a potential impact of SPL207 on OM and inner membrane (IM) permeability was evaluated by fluorescent assays based on the probes N-phenyl-1-naphthylamine (NPN) and propidium iodide (PI) [28,29]. In all four strains tested, an increase in NPN and PI fluorescence was observed as the concentration of SPL207 increased (Figure 4A–D). These results indicate that SPL207 impairs the OM and IM integrity in all the Gram-negative bacteria tested.

Additionally, to evaluate a potential loss of P. aeruginosa envelope integrity in response to SPL207, a Live/Dead staining assay [30] was performed. As shown by confocal microscopy imaging, SPL207 treatment strongly increased the fraction of red fluorescent PAO1 cells (Figure 4E), indicative of membrane damage and cell death (only these cells can uptake the red fluorescent dye PI). Impairment of membrane integrity in SPL207 treated cells is in line with the fast-killing activity exerted by this compound (Figure 3D).

These findings indicate that SPL207 may exert its antibacterial effect via a mechanism akin to that of colistin, involving interaction with the OM and consequent membrane disruption [31]. Interestingly, similar to colistin and in contrast to antibiotics with different mechanisms of action (e.g., ciprofloxacin, chloramphenicol, gentamycin, and tetracycline), SPL207 has no effect on the bacterial growth kinetics when at sub-MICs (Figure 5). The results of the permeability assays and the all-or-nothing effect shared by SPL207 and colistin on P. aeruginosa growth support the mechanism of action of SPL207 as a membrane perturbating agent.

Based on the similarities in the mechanisms of action of SPL207 and colistin, the potential interaction between these two antimicrobials was evaluated. Checkerboard assays revealed that colistin and SPL207 have a synergistic antimicrobial activity against P. aeruginosa, with a Fractional Inhibitory Concentration Index (FICI) = 0.5 (

Table 5. Effect of the SPL207-colistin combination on the MIC of the indicated P. aeruginosa strains.

Strain	Colistin MIC (µg/mL) at SPL207 conc. (µM) of:							Maximum Fold Change a	ColSPL207 b MIC	SPL207Col c MIC	FICI d
	0	2	4	8	16	32	64
PAO1	1	1	0.5	0.5	0.25	0.25	0.25	4	0.25	16	0.5
PAO1 colR1	64	8	8	4	2	2	2	32	2	16	0.094
PAO1 colR3	64	4	2	2	2	2	2	32	2	4	0.047
PAO1 colR5	8	8	1	0.5	0.5	0.5	0.5	16	0.5	8	0.094

^a Ratio between the MIC of colistin and the MIC of Col_SPL207. ^b The MIC value of colistin in combination with SPL207 used to calculate the FICI (µg/mL). ^c The MIC value of SPL207 in combination with colistin used to calculate the FICI (µM). ^d Fractional inhibitory concentration index. For the PAO1 strain, the MIC of SPL207 was 64 µM; for PAO1 col^R1, PAO1 col^R3 and PAO1 col^R5, the MIC of SPL207 was >128 µM, hence it was considered 256 µM for calculation of the FICI [FICI = (MIC Col_SPL207/MIC Col) + (MIC SPL207_Col/MIC SPL207)] [32].

and Figure S2; FICI ≤ 0.5 indicates synergy between antimicrobials) [32]. Conversely, checkerboard assays showed no interaction, either positive or negative, between SPL207 and other two anti-P. aeruginosa antibiotics of clinical relevance such as ciprofloxacin or tobramycin. The effect of the colistin-SPL207 combination was investigated also in P. aeruginosa colistin-resistant strains obtained by in vitro evolution experiments [33]. As reported in Table 5 and Figure S2, the MIC of colistin in these strains was higher than 4 μg/mL (corresponding to the colistin breakpoint in P. aeruginosa according to the EUCAST guidelines; i.e., colistin MIC ≤ 4 μg/mL, susceptible strain; colistin MIC > 4 μg/mL, resistant strain) [34]. The results demonstrated strong synergistic activity between colistin and SPL207 against the three colistin-resistant strains tested (FICI ranging from 0.047 to 0.094; Table 5 and Figure S2). It is important to note that, in the presence of SPL207 at concentrations of 8 μM (for PAO1 col^R1), 2 μM (for PAO1 col^R3), and 4 μM (for PAO1 col^R5), the MIC of colistin against the tested strains decreased below the breakpoint [34].

Overall, SPL207 possesses intrinsic antimicrobial activity and acts as a resistance breaker by resensitizing colistin-resistant P. aeruginosa strains to this last-resort antibiotic.

2.2.3. Small Molecule/Membrane Interaction Analysis by MD Simulations

To provide atomistic details of the interaction between the most promising representative SPL207 and the P. aeruginosa OM, extended molecular dynamics (MD) were carried out by adapting a computational protocol validated previously [35].

The P. aeruginosa OM computational model was validated by 500 ns of all-atom MD simulations, analyzed through frames clustering, mass density distribution along the z-axis, and root mean square deviation (RMSD) (Figure 6). While the representative MD frame (i.e., the centroid frame of the most populated cluster) shows a homogeneous distribution of POPE, POPG, POCL1 phospholipids in the inner leaflet (Figure 6A,B), the RMSD plot indicates that the membrane model is conformationally stable in the MD simulation time (Figure 6C). Visual inspection of the mass density plot (Figure 6D) highlights the peculiar density of Ca²⁺ ions in the PA-LPS region, supporting their crucial role in stabilizing the OM of Gram-negative bacteria through electrostatic interactions with anionic components of the LPS [36,37,38]. Overall, this analysis suggests that the P. aeruginosa OM model is coherent with biological data, and it is suitable for ligand penetration studies in silico.

To evaluate the interaction of SPL207 and its parent compound adarotene (1) into the P. aeruginosa OM model, 500 ns of all-atom MD simulations were performed. The pK_a analysis of amino groups in SPL207 carried out with the MoKa software version 4.0.12 [39] at pH = 7.4 (Figure 7) led to the selection of the protonated form of the molecule with total formal charge = +2 for this in silico study.

Cluster analysis of MD frames shows that, in the most populated cluster, SPL207 penetrates deeply into the OM (Figure 8A) compared to 1, this latter remaining locked within the PA-LPS region (Figure 8B). This behavior is also confirmed by the mass density plot, showing that 1 is unable to overcome the first phosphate layer of core 1a (orange dashed lines in Figure 8C), while SPL207 successfully reaches the central part of the OM (Figure 8D). In evaluating atomistic details of the binding mode of 1 and SPL207, significant differences were observed. The carboxylic acid group of 1 (Figure 8E) establishes water-bridged H-bond interactions with the phosphate groups within the PA-LPS. The adamantyl moiety and the phenolic –OH group are projected towards the outer space and do not establish significant interactions with the P. aeruginosa OM model. In contrast, the positively charged amino groups of SPL207 (Figure 8F) establish direct and water-bridged H-bonds interactions with negatively charged phosphate groups of the lipid A. This suggests that positively charged groups in SPL207 may play a crucial role in OM embedding. Finally, the adamantyl moiety of SPL207 is positioned towards the hydrophobic tails of POPE, POPG and POCL1 of the P. aeruginosa OM model.

2.2.4. Displacement of Ca²⁺ Ions and Membrane Destabilization by SPL207

Based on the stabilizing role of Ca²⁺ ions in bacterial OMs, [34,35,36] it is expected that small molecules acting as membrane destabilizers can displace Ca²⁺ ions from the LPS. To assess this ability by 1 and SPL207, the mass density of Ca²⁺ ions within 20 Å of each molecule was monitored along MD trajectories (Figure 9). The mass density peak (MDP) of 1 fails to affect the mass density profile of Ca²⁺ ions, such as underlined by the comparison with the Ca²⁺ ions density in the free membrane (Figure 9A). In contrast, SPL207 modifies the mass density profile of Ca²⁺ ions with respect to the free membrane (Figure 9B) suggesting that SPL207 might displace Ca²⁺ ions more effectively than 1. To further confirm these results, the mean squared displacement analysis of Ca²⁺ ions was carried out [40,41,42]. Results in

Table 6. Diffusion coefficient for Ca²⁺ ions in PA OM model in the absence and in the presence of SPL207 and 1.

	Diffusion Coefficient [10−7 cm2/s]
System	xy Plane	z Axis	Total
PA free OM	1.36 ± 0.032	0.38 ± 0.025	1.03 ± 0.051
PA–OM + 1	2.90 ± 0.005	0.65 ± 0.049	2.15 ± 0.091
PA-OM + SPL207	4.74 ± 1.132	0.84 ± 0.057	4.73 ± 0.474

clearly show that SPL207 imprints a significant Ca²⁺ ions displacement in the three axes compared to the baseline displacement in the free PA OM model, which is greater than that observed for 1.

Finally, the membrane z thickness and lipid diffusion analysis were calculated along MD trajectories as descriptors of the physical change that occur in the structure of the PA-OM. Membrane z thickness represents the distances in the z axis between the outer and the inner leaflets of the PA OM model, and its comparison in Figure 10A clearly shows that the free OM and the OM/1 system are highly superimposable, whereas the presence of SPL207 induces a greater movement of phospholipids along the z axis, further confirming SPL207’s ability to destabilize the PA OM in silico.

The analysis of lateral lipid diffusion further allows us to describe the movement of lipids along the xy axis over the MD trajectory. The results show that SPL207 triggers a larger movement of lipids than 1. Moreover, unlike compound 1, SPL207 induces a marked increase in lipid diffusion from the very beginning of the MD production trajectory.

Overall, MD results clearly point to the potential mechanism of action of SPL207 as a membrane destabilizer through improved diffusion of Ca²⁺ ions and lipids destabilization.

3. Materials and Methods

3.1. Chemistry

All reagents and solvents were reagent grade or were purified by standard methods before use. ¹H and ¹³C NMR spectra were recorded on Bruker AV600 (Billerica, MA, USA) spectrophotometer at 600 and 150 MHz, respectively, or Bruker Neo 400 spectrophotometer at 400 and 100 MHz, respectively. Chemical shifts are reported in ppm (δ) relative to TMS. The coupling constants, J are reported in Hertz (Hz). All compounds were routinely checked by thin layer chromatography (TLC) using precoated silica gel 60 F254, aluminum foil and the spots were detected under UV light at 254 nm and 365 nm or were revealed spraying with 10% phosphomolybdic acid (PMA) in ethanol.

Compounds 3, 5, 6, 39 [18], 4 [13], 1, 29a-b, 34 [17], 30 [43], 31 [44], 32 [45], 42c [46] were prepared according to reported procedures.

3-((3r,5r,7r)-adamantan-1-yl)-4′-((E)-3-aminoprop-1-en-1-yl)-[1,1′-biphenyl]-4-ol (7). To a solution of AlCl₃ (49 mg, 0.37 mmol) in dry THF (1.25 mL) 1 M LiAlH₄ in THF (0.56 mL) was added at 0 °C and the solution was stirred 10 min, then compound 34 [17] (100 mg, 0.28 mmol) dissolved in THF (1.25 mL) was added at 0 °C. The mixture was stirred for 1 h at 0 °C, then quenched with water and 1M NaOH and extracted with AcOEt. The organic phase was dried with Na₂SO₄ and concentrated under reduced pressure. The crude was purified by preparative TLC (CH₂Cl₂: MeOH = 9:1) to afford product in 28% yield (28 mg) as a white solid. R_f = 0.39 (CH₂Cl₂: MeOH = 9:1). ¹H-NMR (400 MHz, DMSO-d₆) δ: 7.56–7.49 (2H, m); 7.47–7.39 (2H, m); 7.37–7.27 (2H, m); 6.86 (1H, d, J = 8.3 Hz); 6.54 (1H, d, J = 16.0 Hz); 6.35 (1H, dt, J = 16.0 Hz, J = 5.7 Hz); 3.37 (2H, d, J = 5.7 Hz); 2.22–2.10 (6H, m); 2.09–2.01 (3H, m); 1.82–1.67 (6H, m). ¹³C-NMR (100 MHz, DMSO-d₆) δ: 156.4; 140.1; 136.3; 135.4; 131.2; 130.7; 129.0; 127.0 (×2C); 126.5 (×2C); 125.0; 124.9; 117.3; 43.8; 40.3 (×3C); 37.1 (×3C); 36.7; 28.9 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-(3-aminopropyl)-[1,1′-biphenyl]-4-ol (8). To a solution of compound 34 (128 mg, 0.359 mmol) in dry THF (2.5 mL) 1 M LiAlH₄ in THF (0.05 mL) was added at 0 °C. The mixture was stirred for 2 h at 0 °C, then quenched with water and 1 M NaOH and extracted with AcOEt. The organic phase was dried with Na₂SO₄ and concentrated under reduced pressure. The crude was purified by column chromatography (CH₂Cl₂: MeOH = 95:5 with 1% of conc. NH₃ aqueous solution) to afford product 8 in 15% yield (20 mg) as a white solid. R_f = 0.39 (CH₂Cl₂: MeOH = 9:1). ¹H-NMR (400 MHz, CH₃OH-d₄) δ: 7.48–7.43 (2H, m); 7.36 (1H, d, J = 2.3 Hz); 7.27–7.21 (3H, m); 6.78 (1H, d, J = 8.2 Hz); 2.83–2.78 (2H, m); 2.72 (2H, t, J = 7.2 Hz); 2.27–2.20 (6H, m); 2.12–2.05 (3H, m); 1.95–1.79 (8H, m). ¹³C-NMR (100 MHz, CH₃OH-d₄) δ: 157.1; 140.5; 137.7; 132.2; 129.7 (×2C); 127.5 (×2C); 126.1; 125.8; 117.6; 41.6 (×3C); 41.4; 38.3 (×3C); 38.0; 33.5; 30.66 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-((4-methylpiperazin-1-yl)methyl)-[1,1′-biphenyl]-4-ol (9). To a suspension of aldehyde 29b [17] (200 mg, 0.448 mmol) in CH₃CN (1 mL) 1-methylpiperazine (49 mg, 0.49 mmol), formic acid (41 mg, 0.90 mmol) and BF₃Et₂O (0.002 mmol) were added and the mixture was heated at 85 °C for 6 h. The reaction was concentrated and treated with Et₂O to obtain a precipitate. Compound 9 was obtained in 38% yield (70 mg) as a white solid upon chromatographic purification (CH₂Cl₂: MeOH = 20: 1 with 1% of conc. NH₃). ¹H-NMR (400 MHz, DMSO-d₆) δ: 9.42 (1H, s); 7.52–7.45 (2H, m); 7.35–7.23 (4H, m); 6.84 (1H, d, J = 8.1 Hz); 3.46 (2H, s); 2.48–2.20 (8H, m); 2.18–2.10 (9H, m); 2.07–2.00 (3H, m); 1.82–1.70 (6H, m). ¹³C-NMR (100 MHz, DMSO-d₆) δ: 155.8; 139.6; 136.2; 135.8; 130.6; 129.3 (×2C); 125.8 (×2C); 124.7 (×2C); 116.8; 61.8; 54.7 (×2C); 52.6 (×2C); 45.8; 39.9 (×3C); 36.6 (×3C); 36.3; 28.4 (×3C).

(E)-3′-((3r,5r,7r)-adamantan-1-yl)-4′-hydroxy-[1,1′-biphenyl]-4-carbaldehyde O-(2-morpholinoethyl) oxime (10). To a solution of compound 29a [17] (20 mg, 0.06 mmol) in EtOH (0.5 mL) O-(2-morpholinoethyl)hydroxylamine hydrochloride (22 mg, 0.12 mmol) and pyridine (95 mg, 1.2 mmol) were added and the resulting mixture was refluxed for 5 h. The reaction was concentrated and treated with H₂O to obtain product 10 as a white precipitate in 67% yield (18 mg) without further purification. R_f = 0.14 (TLC hexane: acetone = 7:3). ¹H-NMR (400 MHz, DMSO-d₆) δ: 9.54 (1H, s); 8.26 (1H, s); 7.68–7.57 (4H, m); 7.39–7.30 (2H, m); 6.87 (1H, d, J = 8.0 Hz); 4.24 (2H, t, J = 5.8 Hz); 3.63–3.51 (4H, m); 2.63 (2H, t, J = 5.8 Hz); 2.48–2.40 (4H, m); 2.20–2.10 (6H, m); 2.09–2.00 (3H, m); 1.83–1.67 (6H, m). ¹³C-NMR (100 MHz, DMSO-d₆) δ: 156.3; 148.5; 142.3; 135.9;129.8; 129.8; 127.3 (×2C); 126.2 (×2C); 124.9; 124.7; 116.9; 71.3; 66.2 (×2C); 57.0; 53.6 (×2C); 39.8 (×3C); 36.6 (×3C); 36.3; 28.4 (×3C).

(E)-3-(3′-((3r,5r,7r)-adamantan-1-yl)-4′-hydroxy-[1,1′-biphenyl]-4-yl)-N-(2-morpholinoethoxy)acrylamide (11). To a solution of compound 1 (adarotene) [17] (50 mg, 0.13 mmol) in DMF (2 mL) cooled to 0 °C DIPEA (52 mg, 0.40 mmol) and HBTU (51 mg, 0.13 mmol) were added and the resulting mixture was stirred for 10 min. After addition of O-(2-morpholinoethyl)hydroxylamine hydrochloride (24 mg, 0.13 mmol) the reaction was stirred for 48 h at room temperature and then concentrated and treated with H₂O to obtain a precipitate. Product 11 was obtained in 24% yield (16 mg) as a sticky solid upon chromatographic purification (DCM: MeOH = 9:1). R_f = 0.26 (DCM: MeOH = 9:1). ¹H-NMR (400 MHz, CH₃OH-d₄) δ: 7.65 (1H, d, J = 15.8 Hz); 7.61–7.56 (4H, m); 7.42 (1H, d, J = 2.3 Hz); 7.31 (1H, dd, J = 2.3 Hz, 8.3 Hz); 6.82 (1H, d, J = 8.3 Hz); 6.46 (1H, d, J = 15.8 Hz); 4.13 (2H, t, J = 5.2 Hz); 3.78–3.71 (4H, m); 2.76 (2H, t, J = 5.2 Hz); 2.68–2.57 (4H, m); 2.27–2.18 (6H, m); 2.12–2.03 (3H, m); 1.92–1.80 (6H, m). ¹³C-NMR (100 MHz, CH₃OH-d₄) δ: 166.4; 157.8; 145.0; 142.5; 137.9; 133.7; 132.2; 129.4 (×2C); 127.8 (×2C); 126.2; 126.1; 117.8; 117.1; 74.4; 67.6 (×2C); 57.5; 54.8 (×2C); 41.5 (×3C); 38.3 (×3C); 38.0; 30.6 (×3C).

(E)-3′-((3r,5r,7r)-adamantan-1-yl)-4′-hydroxy-[1,1′-biphenyl]-4-carbaldehyde O-(2-aminoethyl) oxime (12). To a solution of compound 29a [17] (30 mg, 0.09 mmol) in EtOH, (0.5 mL) 2-(aminooxy)ethan-1-amine hydrochloride (27 mg, 0.18 mmol) and pyridine (142 mg, 1.80 mmol) were added and the resulting mixture was refluxed for 5 h. The reaction was concentrated under reduced pressure and the crude was finely shredded and washed with H₂O to afford product 12 in 51% yield (18 mg) as a white solid without further purification. R_f = 0.16 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 8.27 (1H, s); 7.22–7.66 (2H, m); 7.64–7.58 (2H, m); 7.43 (1H, d, J = 1.8 Hz); 7.32 (1H, dd, J = 1.8Hz, 8.4 Hz); 6.82 (1H, d, J = 8.4 Hz); 4.39 (2H, t, J = 4.9 Hz); signal overlapped with the solvent, 2.28–2.22 (6H, m); 2.12–2.07 (3H, m); 1.90–1.82 (6H, m). ¹³C-NMR (150 MHz, CH₃OH-d₄) δ: 157.9; 151.9; 145.3; 137.9; 132.2; 130.8; 128.7 (×2C); 127.6 (×2C); 126.2; 126.1; 117.8; 70.9; 41.5 (×3C); 40.3; 38.3 (×3C); 38.0; 30.6 (×3C).

(E)-3-((3r,5r,7r)-adamantan-1-yl)-4-hydroxybenzaldehyde O-(2-aminoethyl) oxime (13). To a solution of compound 30 [43] (20 mg, 0.078 mmol) in EtOH, (0.5 mL) 2-(aminooxy)ethan-1-amine hydrochloride (23 mg, 0.156 mmol) and pyridine (123 mg, 1.56 mmol) were added and the resulting mixture was refluxed for 3 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and again concentrated under reduced pressure. Product 13 was obtained in 37% yield (9 mg) as a sticky solid without further purification. R_f = 0.34 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (400 MHz, CH₃OH-d₄) δ: 8.09 (1H, s); 7.43 (1H, d, J = 2.0 Hz); 7.25 (1H, dd, J = 2.0 Hz, 8.3 Hz); 6.74 (1H, d, J = 8.3 Hz); 4.19 (2H, t, J = 5.2 Hz); 3.03 (2H, t, J = 5.2 Hz); 2.20–2.13 (6H, m); 2.09–2.03 (3H, m); 1.86–1.79 (6H, m). ¹³C-NMR (100 MHz, CH₃OH-d₄) δ: 159.8; 151.5: 137.9; 127.1; 126.8; 124.3; 117.5; 74.5; 41.4; 41.4 (×3C); 38.2 (×3C); 37.9; 30.5 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-(((2-hydroxyethyl)amino)methyl)-[1,1′-biphenyl]-4-ol (14). Ethanolamine (5 mg, 0.09 mmol) was added dropwise to a solution of compound 29a [17] (30 mg, 0.09 mmol) in MeOH (1.5 mL) at 0 °C and the resulting mixture was stirred for 18 h at room temperature. NaBH₄ (7 mg, 0.18 mmol) was then gradually added at 0 °C and after stirring for 30 min the reaction was diluted with 1 M HCl. The aqueous phase was washed with CH₂Cl₂ in order to remove organic impurities, then added with 1 M NaOH and extracted with ethyl acetate. The organic phase was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure to afford product 14 in 97% yield (32 mg) as a sticky yellow solid without further purification. R_f = 0.32 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 7.54–7.49 (2H, m); 7.41–7.35 (3H, m); 7.26 (1H, dd, J = 2.1 Hz, 8.3 Hz); 6.80 (1H, d, J = 8.3 Hz); 3.83 (2H, s); 3.71 (2H, J = 5.5 Hz); 2.77 (2H, t, J = 8.3 Hz); 2.31–2.17 (6H, m); 2.13–2.05 (3H, m); 1.90–1.79 (6H, m). ¹³C-NMR (150 MHz, CH₃OH-d₄) δ: 157.3; 142.3; 138.1; 137.7; 133.0; 129.9 (×2C); 127.5 (×2C); 126.2; 125.9; 117.7; 61.4; 54.0; 51.6; 41.6 (×2C); 38.3 (×3C); 38.0; 30.6 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-(((2-(dimethylamino)ethyl)amino)methyl)-[1,1′-biphenyl]-4-ol (15). N¹,N¹-dimethylethane-1,2-diamine (11 mg, 0.12 mmol) was added dropwise to a solution of compound 29a [17] (40 mg, 0.12 mmol) in MeOH (1.3 mL) at 0 °C and the resulting mixture was stirred for 18 h at room temperature. NaBH₃CN (15 mg, 0.24 mmol) was then gradually added at 0 °C and after stirring for 40 min the reaction was concentrated under reduced pressure. The residue was treated with H₂O and the resulting precipitate was filtered, washed with H₂O and dried to afford product 15 in 84% yield (41 mg) as a white solid without further purification. R_f = 0.14 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 8.42 (1H, s); 7.84–7.75 (2H, m); 7.69–7.61 (2H, m); 7.45 (1H, d, J = 2 Hz); 7.34 (1H, dd, J = 2 Hz, 8.2 Hz); 6.82 (1H, d, J = 8.2 Hz); 3.79 (2H, t, J = 6.7 Hz); 2.72 (2H, t, J = 6.7 Hz); 2.41–2.32 (3H, s); 2.31–2.20 (6H, m); 2.13–2.06 (3H, m); 1.92–1.80 (6H, m). ¹³C-NMR (150 MHz, CDCl₃) δ: 157.9; 146.0; 137.8; 134.9; 132.2; 129.9 (×2C); 127.5 (×2C); 126.3; 126.2; 117.8; 60.7; 59.7; signal overlapped with the solvent; 45.8 (×2C); 41.5 (×3C); 38.3 (×3C); 38.0; 30.6 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-(((2-(2-hydroxyethoxy)ethyl)amino)methyl)-[1,1′-biphenyl]-4-ol (16). 2-(2-aminoethoxy)ethan-1-ol (9 mg, 0.09 mmol) was added dropwise to a solution of compound 29a [17] (30 mg, 0.09 mmol) in MeOH (1.3 mL) at 0 °C and the resulting mixture was stirred for 18 h at room temperature. NaBH₃CN (11 mg, 0.18 mmol) was then gradually added at 0 °C and after stirring for 30 min the reaction was concentrated under reduced pressure. The residue was treated with H₂O and resulting precipitate was filtered, washed with H₂O and dried to afford product 16 in 54% yield (24 mg) as a white solid without further purification. R_f = 0.40 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (400 MHz, CH₃OH-d₄) δ: 7.65–7.55 (2H, m); 7.51–7.43 (2H, m); 7.39 (1H, d, J = 2.1 Hz); 7.28 (1H, dd, J = 2.1 Hz, J = 8.2 Hz); 6.81 (1H, d, J = 8.2 Hz); 4.09 (2H, s); 3.77–3.67 (4H, m); 3.63–3.55 (2H, m); 3.15–3.05 (2H, m). ¹³C-NMR (100 MHz, CH₃OH-d₄) δ: 157.6; 143.7; 137.8; 133.0; 132.4; 130.9 (×2C); 127.9 (×2C); 126.2; 126.1; 117.7; 73.4; 68.1; 62.1; 52.7.0; 48.5; 41.5 (×3C); 38.3 (×3C); 38.0; 30.6 (×3C).

3-((3r,5r,7r)-adamantan-1-yl)-4′-(((2-((2-aminoethyl)amino)ethyl)amino)methyl)-[1,1′-biphenyl]-4-ol (17). N¹-(2-aminoethyl)ethane-1,2-diamine (24 mg, 0.23 mmol) was added dropwise to a solution of compound 29a [17] (26 mg, 0.07 mmol) in MeOH (1.3 mL) at 0 °C and the resulting mixture was stirred for 18h at room temperature. NaBH₄ (8 mg, 0.23 mmol) was then gradually added at 0 °C and after stirring for 30 min the reaction was diluted with 1 M HCl. The aqueous phase was washed with CH₂Cl₂ to remove organic impurities and added with 1 M NaOH. The resulting precipitate was filtered, washed with H₂O and dried. Product 17 was obtained in 49% yield (14 mg) as a white solid upon crystallization from Et₂O. R_f = 0.08 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (400 MHz, CH₃OH-d₄) δ: 7.54–7.46 (2H, m); 7.41–7.32 (3H, m); 7.25 (1H, dd, J = 2.2 Hz, 8.2 Hz); 6.79 (1H, d, J = 8.2 Hz); 3.79 (2H, s); 2.81–2.71 (6H, m); 2.71–2.63 (2H, m); 2.27–2.17 (6H, m); 2.11–2.03 (3H, m); 1.89–1.77 (6H, m). ¹³C-NMR (100 MHz, CH₃OH-d₄) δ: 157.3; 142.3; 138.4; 137.7; 133.0; 129.9 (× 2C); 127.5 (× 2C); 126.1; 125.9; 117.7; 54.2; 54.0; 52.3; 49.5; 41.7 (× 3C); 38.3 (× 3C); 38.0; 30.6 (× 3C).

(E)-3-(3′-((3r,5r,7r)-adamantan-1-yl)-4′-hydroxy-[1,1′-biphenyl]-4-yl)-N,N-bis(3-(dimethylamino)propyl)acrylamide (18). To a solution of compound 1 (50 mg, 0.134 mmol) in dry CH₂Cl₂ (2 mL), DIPEA (163 mg, 1.3 mmol), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (32 mg, 0.173 mmol) and BOP (77 mg, 0.173) were added and the resulting mixture was stirred for 48 h at room temperature. The reaction was diluted with CH₂Cl₂, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. The resulting precipitate was filtered, washed with H₂O and dried. The crude was purified by column chromatography (CH₂Cl₂: MeOH = 93:7 with 1% of conc. NH₃ aqueous solution) to afford product 18 in 54% yield (39 mg) as a sticky solid. R_f = 0.20 (CH₂Cl₂: MeOH = 95:5 with 1% of conc. NH₃ aqueous solution). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 7.70–7.65 (2H, m); 7.63–7.58 (3H, m); 7.44 (1H, d, J = 2.0 Hz); 7.32 (1H, dd, J = 2.0 Hz, J = 8.1 Hz) 7.15 (1H, d, J = 15.2 Hz); 6.82 (1H, d, J = 8.1 hz); 3.62 (2H, t, J = 7.7 Hz); 3.52 (2H, t, J = 7.2 Hz); 2.54 (2H, m); 2.47–2.42 82H, m); 2.41 (6H, s); 2.34–2.28 (6H, s); 2.27–2.22 (6H, m); 2.12–2.05 83H, m); 1.93–1.87 (4H, m); 1.86–1.83 (6H, m). ¹³C-NMR (600 MHz, CH₃OH-d₄) δ: 169.1; 157.8; 144.8; 144.1; 137.9; 134.2; 132.2; 129.6 (×2C); 127.7 (×2C); 126.2; 126.0; 117.8; 117.6; 57.7; 57.2; 47.2; 45.8; 45.5 (×2C); 45.1 (×2C); 41.5 (×3C); 38.3 (×3C); 38.0; 30.6 (×3C); 28.2; 26.2.

tert-butyl (E)-3-(3′-((3r,5r,7r)-adamantan-1-yl)-4′-(4-(bis(3-(dimethylamino)propyl)amino)butoxy)-[1,1′-biphenyl]-4-yl)acrylate (20). To a solution of compound 1 (140 mg, 0.32 mmol) in acetone (5 mL), K₂CO₃ (225 mg, 1.62 mmol) and 1,4-dibromobuthane (526 mg, 2.43 mmol) were added and the resulting mixture was heated to reflux for 6 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated again under reduced pressure. Product 33 was obtained in 66% yield (122 mg) as a white solid upon crystallization from hexane. R_f = 0.75 (Hex: AcOEt = 8:2). ¹H-NMR (600 MHz, CDCl₃) δ: 7.64 (1H, d, J = 16.0 Hz); 7.62–7.55 (4H, m); 7.51 (1H, d, J = 2.1 Hz); 7.43 (1H, dd, J = 2.1 Hz, J = 8.6 Hz); 6.96 ( 1H, d, J = 8.6 Hz); 6.41 (1H, d, J = 16.2 Hz); 4.10 (2H, t, J = 6.2 Hz); 3.56 (2H, t, J = 6.6 Hz); 2.24–2.17 (8H, m); 2.16–2.07 (5H, m); 1.86–1.787 (6H, m); 1.58 (9H, s).

To a solution of the above intermediate 33 (50 mg, 0.09 mmol) in dry DMF (2.5 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (165 mg, 0.90 mmol) was added and the mixture was stirred for 24h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Impurities were removed from the crude by dissolving them in Et₂O to obtain product 20 in 47% yield (25 mg) as a sticky solid. R_f = 0.32 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 7.68–7.58 (5H, m); 7.50 (1H, d; J = 2.2 Hz); 7.47 (1H, dd, J = 2.2 Hz, J = 8.2 Hz); 7.06 (1H, d, J = 8.2 Hz); 6.45 (1H, d, J = 16.0 Hz); 4.16 (2H, t, J = 5.9 Hz); 3.52–3.43 (4H, m); 3.13 (6H, s); 2.72–2.60 (4H, m); 2.49–2.40 (2H, m); 2.30 (6H, m); 2.25–2.19 (6H, m); 2.16–2.05 (5H, m); 2.04–1.98 (4H, m); 1.90–1.79 (6H, m); 1.72–1.64 (2H, m); 1.54 (9H, s). ¹³C-NMR (150 MHz, CDCl₃) δ: 168.2; 159.2; 144.8; 142.6; 139.6; 134.1; 133.7; 129.7 (×2C); 127.9 (×2C); 126.4; 126.3; 120.5; 113.9; 81.7; 68.1; 65.2; 63.7; 58.5; 51.3; 47.1; 45.3 (×4C); 42.0 (×3C); 38.3 (×4C); 30.6 (×3C); 28.5 (×3C); 27.9; 27.5; 23.7; 21.0.

(E)-3-(3′-((3r,5r,7r)-adamantan-1-yl)-4′-(4-(bis(3-(dimethylamino)propyl)amino)butoxy)-[1,1′-biphenyl]-4-yl)acrylic acid (19). To a solution of compound 20 (20 mg, 0.03 mmol) in dry CH₂Cl₂ (0.47 mL), trifluoroacetic acid (0.21 mL) was added at 0 °C and the resulting mixture was stirred at 0 °C for 2 h. The reaction was concentrated under reduced pressure. After the azeotropic removal of TFA with toluene, product 19 was obtained in quantitative yield (28 mg) as a sticky solid. R_f = 0.57 (reverse phase TLC MeOH: H₂O = 9:1). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 7.73 (1H, d, J = 16.2 Hz); 7.70–7.61 (4H, m); 7.52 (1H, d, J = 2.1 Hz); 7.49 (1H, dd, J = 2.1 Hz, 8.6 Hz); 7.08 (1H, d, J = 8.6 Hz); 6.52 (1H, d, J = 16.2 Hz); 4.18 (2H, t, J = 6.1 Hz); 3.57–3.49 (4H, m); 3.31–3.26 (2H, m); 3.24–3.15 (10H, m); 2.99–2.92 (6H, m); 2.34–2.28 (2H, m); 2.27–2.18 (8H, m); 2.16–2.08 (5H, m); 2.07–1.99 (2H, m); 1.91–1.84 (6H, m). ¹³C-NMR (150 MHz, CH₃OH-d₄) δ: 170.6; 162.6; 146.1; 144.8; 139.8; 134.2; 133.8; 129.9 (×2C); 128.1 (×2C); 126.6; 126.4; 118.8; 114.0; 68.2; 65.8; 62.3; 55.7; 51.4; 49.5; 46.0; 45.8; 43.7 (×3C); 42.1 (×3C); 38.4 (×4C); 30.5 (×3C); 22.8; 21.1; 21.0.

Methyl 3′-((3r,5r,7r)-adamantan-1-yl)-4′-(4-(bis(3-(dimethylamino)propyl)amino)butoxy)-[1,1′-biphenyl]-4-carboxylate (21). To a 0 °C cooled solution of methyl 4′-hydroxy(1,1′-biphenyl)-4-carboxylate 37 (100 mg, 0.438 mmol) and 1-adamantanol (73 mg, 0.48 mmol) in CH₂Cl₂ (3.5 mL), H₂SO₄ (0.05 mL) was added and the resulting mixture was stirred for 2 h. The reaction was diluted with cold H₂O, extracted with CH₂Cl₂, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The crude was purified by column chromatography (Hex: AcOEt = 8:2) to afford product 38 in 42% yield (67 mg). R_f = 0.30 (Hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 8.14–8.06 (2H, m); 7.68–7.61 (2H, m); 7.52 (1H, d, J = 2.2 Hz); 7.37 (1H, dd, J = 2.2 Hz, 8.2 Hz); 6.77 (1H, d, J = 8.2 Hz); 4.98 (1H, s); 3.97 (3H, s); 2.25–2.20 (6H, m); 2.18–2.11 (3H, m); 1.86–1.81 (6H, m).

To a solution of above intermediate 38 (33 mg, 0.09 mmol) in acetone (1.5 mL), K₂CO₃ (63 mg, 0.46 mmol) and 1,4-dibromobuthane (136 mg, 0.63 mmol) were added and the resulting mixture was refluxed for 4 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated again under reduced pressure. Product 41a was obtained in 53% yield (23 mg) as a white solid upon crystallization from hexane. R_f = 0.36 (Hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 8.14–8.05 (2H, m); 7.70–7.60 (2H, m); 7.53 (1H, d, J = 2.2 Hz); 7.46 (1H, dd, J = 2.2 Hz, 8.4 Hz); 6.97 (1H, d, J = 8.4 Hz); 4.11 (2H, t, J = 6.0 Hz); 3.96 (3H, s); 3.56 (2H, t, J = 6.4 Hz); 2.26–2.17 (8H, m); 2.16–2.05 (5H, m); 1.87–1.78 (6H, m).

To a solution of compound 41a (60 mg, 0.12 mmol) in THF (2 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (0.27 mL, 0.4 mmol) was added and the mixture was refluxed for 6 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Product 21 was obtained in 39% yield (4 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃). R_f = 0.23 (DCM: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃). ¹H-NMR (400 MHz, CDCl₃) δ: 8.13–8.03 (2H, m); 7.68–7.62 (2H, m); 7.52 (1H, d, J = 2.0 Hz); 7.45 (1H, dd, J = 2.0 Hz, 8.3 Hz); 6.95 (1H, d, J = 8.3 Hz); 4.06 (2H, t, J = 6.1 Hz); 3.94 (3H, s); 2.60–2.46 (6H,.m); 2.41–2.33 (2H, m); 2.29 (12H, s); 2.23–2.19 (6H, m); 2.16–2.10 (3H, m); 1.97–1.85 (2H, m); 1.83–1.77 (6H, m); 1.76–1.62 (6H, m). ¹³C-NMR (100 MHz, CDCl₃) δ: 167.3; 158.5; 146.1; 138.7; 131.8; 130.1 (×2C); 128.1; 126.6 (×2C); 125.9; 125.7; 112.4; 67.9; 57.9 (×2C); 53.9; 52.1; 52.0 (×2C); 45.4 (× 4C); 40.7 (×3C); 37.3 (×3C); 29.8; 29.2 (×3C); 27.6; 25.1 (×2C); 24.0.

N¹-(4-((3-(adamantan-1-yl)-4′-chloro-[1,1′-biphenyl]-4-yl)oxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (22). To a solution of compound 39 [18] (463 mg, 1.30 mmol) in dry acetone (12 mL), K₂CO₃ (898 mg, 6.50 mmol) and 1,4-dibromobuthane (1.10 mL, 9.15 mmol) were added and the resulting mixture was refluxed for 4 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄. In vacuo concentration followed by flash column chromatography in Hex: AcOEt 9: 1 furnished the intermediate 2-(3-((3r,5r,7r)-adamantan-1-yl)-4-(4-bromobutoxy)phenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (40) in 76% yield (484 mg). ¹H-NMR (600 MHz, CDCl₃) δ: 7.32 (1H, d, J = 2.4 Hz); 7.27 (1H, dd, J = 2.4 Hz, 8.9 Hz); 6.74 (1H, d, J = 8.9 Hz); 4.01 (2H, t, J = 5.9 Hz); 3.54 (2H, t, J = 6.6 Hz); 2.23–2.15 (2H, m); 2.14–2.02 (11H, m); 1.85–1.72 (6H, m).

To a suspension of intermediate (40, 100 mg, 0.20 mmol) in THF: H₂O 2: 1 (2.4 mL), 1-bromo-4-chlorobenzene (38 mg, 0.20 mmol), K₂CO₃ (69 mg, 0.50 mmol) and Pd(PPh₃)₄ (7 mg, 0.006 mmol) were sequentially added under N₂ atmosphere, and the mixture was heated 4 h at 80 °C. The reaction was diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The intermediate 41b was obtained in 39% yield (37 mg) as a white solid upon crystallization from hexane. ¹H-NMR (600 MHz, CDCl₃) δ: 7.53–7.47 (2H, m); 7.44 (1H, d, J = 2.0 Hz); 7.41–7.34 (3H, m); 6.94 (1H, d, J = 8.3 Hz); 4.09 (2H, t, J = 5.9 Hz); 3.55 (2H, t, J = 6.5 Hz); 2.26–2.15 (8H; m); 2.15–2.08 (5H, m); 1.86–1.75 (6H, m).

To a solution of compound 41b (37 mg, 0.09 mmol) in THF (0.8 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (174 µL, 0.78 mmol) was added and the mixture was refluxed for 4 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Product 22 was obtained in 24% yield (11 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃). ¹H-NMR (400 MHz, CDCl₃) δ: 7.54–7.46 (2H, m); 7.43 (1H, d, J = 2.3 Hz); 7.40–7.33 (3H, m); 6.94 (1H, d, J = 8.7 Hz); 4.05 (2H, t, J = 5.9 Hz); 2.59–2.52 (2H, m); 2.52 −2.44 (4H, m); 2.25 (12H, s); 2.22–2.16 (6H, m); 2.15–2.07 (3H, m); 1.95–1.85 (2H, m); 1.84–1.78 (6H, m); 1.77–1.59 (6H, m). ¹³C-NMR (100 MHz, CDCl₃) δ: 158.0; 140.2; 138.7; 132.5; 131.9; 128.8 (×2C); 128.1 (×2C); 125.6; 125.3; 112.3; 67.9; 58.1 (×2C); 54.0; 52.1 (×2C); 45.6 (×4C); 40.7 (×3C); 37.3 (×3C); 29.8; 29.2 (×3C); 27.7; 25.5 (×2C); 24.1.

N¹-(4-((3-((3r,5r,7r)-adamantan-1-yl)-4′-methyl-[1,1′-biphenyl]-4-yl)oxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (24). To a suspension of intermediate (40) in THF: H₂O 2: 1 (2.4 mL), 1-bromo-4-methylbenzene (34 mg, 0.20 mmol), K₂CO₃ (69 mg, 0.50 mmol) and Pd(PPh₃)₄ (7 mg, 0.006 mmol) were sequentially added under N₂ atmosphere, and the mixture was heated 4 h at 80 °C. The reaction was diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The intermediate (3r,5r,7r)-1-(4-(4-bromobutoxy)-4′-methyl-[1,1′-biphenyl]-3-yl)adamantane 41c was obtained in 47% yield (42 mg) as a white solid upon crystallization from hexane. ¹H-NMR (400 MHz, CDCl₃) δ: 7.51–7.44 (3H, m); 7.38 (1H, d, J = 2.3 Hz, 8.4 Hz); 7.27–7.21 (2H, m); 6.93 (1H, d, J = 8.4 Hz); 4.09 (2H, t, J = 6.0 Hz); 3.55 (2H, t, J = 6.5 Hz); 2.41 (3H, s); 2.23–2.16 (8H, m); 2.15–2.05 (5H, m); 1.88–1.77 (6H, m).

To a solution of compound 41c (42 mg, 0.09 mmol) in THF (0.8 mL) N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (206 µL, 0.92 mmol) was added and the mixture was refluxed for 4 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Product 24 was obtained in 25% yield (13 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃). ¹H-NMR (600 MHz, CDCl₃) δ: 7.48–7.41 (3H, m); 7.38–7.33 (1H, m); 7.24–7.18 (2H, m); 6.91 (1H, d, J = 8.5 Hz); 4.02 (2H, t, J = 6.2 Hz); 2.55–2.49 (2H, m); 2.49–2.42 (4H, m); 2.38 (3H, s); 2.31–2.25 (4H, m); 2.22 (12H, s); 2.20–2.14 (6H, m); 2.12–2.04 (3H, m); 1.90–1.83 (2H, m); 1.82–1.75 (6H, m); 1.74–1.56 (6H, m). ¹³C NMR (101 MHz, CDCl₃) δ 157.51, 138.81, 138.32, 136.03, 133.09, 129.34 (×2C), 126.70 (×2C), 125.53, 125.05, 112.17, 67.78, 57.98 (×2C), 53.93, 52.06 (×2C), 45.51 (×4C), 40.66 (×3C), 37.24 (×3C), 37.16, 29.17 (×3C), 27.65, 25.44 (×2C), 24.06, 21.06.

N¹-(4-((3-((3r,5r,7r)-adamantan-1-yl)-4′-bromo-[1,1′-biphenyl]-4-yl)oxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (23). To a solution of 3-((3r,5r,7r)-adamantan-1-yl)-4′-bromo-[1,1′-biphenyl]-4-ol 31 [17,44] (100 mg, 0.26 mmol) in dry acetone (3 mL), K₂CO₃ (180 mg, 1.3 mmol) and 1,4-dibromobuthane (392 mg, 1.82 mmol) were added and the resulting mixture was refluxed for 6 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated again under reduced pressure. Product 35 was obtained in 65% yield (87 mg) as a white solid upon crystallization from hexane. R_f = 0.77 (Hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 7.57–7.52 (2H, m); 7.48–7.42 (3H, m); 7.37 (1H, dd, J = 2.6 Hz, 8.7 Hz); 6.95 (1H, d, J = 8.7 Hz); 4.10 (2H, t, J = 6.2 Hz); 3.56 (2H, t, J = 6.5 Hz); 2.25–2.16 (8H, m); 2.15–2.04 (5H, m); 1.86–1.77 (6H, m).

To a solution of compound 35 (40 mg, 0.077 mmol) in dry DMF (1.5 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (144 mg, 0.77 mmol) was added and the mixture was stirred for 18 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Impurities were removed from the crude by dissolving them in Et₂O: Hex= 1: 1 to obtain product 23 in 42% yield (20 mg) as a sticky solid. R_f = 0.30 (CH₂Cl₂: MeOH = 8:2 with 1% of conc. NH₃ aqueous solution). ¹H-NMR (600 MHz, CH₃OH-d₄) δ: 7.59–7.53 (2H, m); 7.51–7.47 (2H, m); 7.43 (1H, d, J = 2.3 Hz); 7.41 (1H, dd, J = 2.3 Hz, 8.4 Hz); 7.03 (1H, d, J = 8.4 Hz); 4.10 (2H, t, J = 6.1 Hz); 2.64 (2H, t, J = 7.2 Hz); 2.60–2.54 (4H, m); 2.49–2.43 (4H, m); 2.33 (12H, s); 2.28–2.20 (6H, m); 2.15–2.08 (3H, m); 1.97–1.90 (2H, m); 1.90–1.85 (6H, m); 1.85–1.78 (2H, m); 1.77–1.69 (4H, m). ¹³C-NMR (150 MHz, CH₃OH-d₄) δ: 159.3; 142.0; 139.6; 133.1; 132.8 (×2C); 129.3 (×2C); 126.3; 126.1; 121.4; 113.6; 68.8; 58.8 (×2C); 54.7; 53.0 (×2C); 45.3 (×4C); 42.0 (×3C); 38.3 (×3C); 30.8; 30.6 (×3C); 28.7; 25.2 (×2C); 24.8.

N¹-(4-((3-((3r,5r,7r)-adamantan-1-yl)-[1,1′:4′,1″-terphenyl]-4-yl)oxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (25). Compound 31 (100 mg, 0.26 mmol), phenylboronic acid (32 mg, 0.26 mmol), Pd(PPh₃)₄ (9 mg, 0.008 mmol) and K₂CO₃ (90 mg, 0.65 mmol) were placed under N₂ atmosphere and then H₂O (1 mL) and THF (2 mL) were added. The resulting mixture was heated at 80 °C for 4 h. The reaction was diluted with cold water, acidified with 1 M HCl and extracted with ethyl acetate. The organic phase was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The residue was purified by flash column chromatography (petroleum ether: AcOEt = 95: 5) to obtain product 36 in 46% yield (46 mg) as a white solid. R_f = 0.29 (Hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 7.70–7.72 (6H, m); 7.55–7.52 (1H, m); 7.51–7.46 (2H, m); 7.40−7.35 (2H, m); 6.77 (1H, d, J = 8.0 Hz); 2.25–2.21 (6H, m); 2.18–2.11 (3H, m); 1.87–1.80 (6H, m).

To a solution of intermediate 36 (46 mg, 0.12 mmol) in acetone (1 mL), K₂CO₃ (83 mg, 0.6 mmol) and 1,4-dibromobuthane (181 mg, 0.84 mmol) were added and the resulting mixture was refluxed for 6 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated again under reduced pressure. (3r,5r,7r)-1-(4-(4-bromobutoxy)-[1,1′:4′,1″-terphenyl]-3-yl)adamantane was obtained in 40% yield (25 mg) as a white solid upon crystallization from petroleum ether. R_f = 0.53 (Hex: AcOEt = 9:1). ¹H-NMR (400 MHz, CDCl₃) δ: 7.72–7.61 (6H, m); 7.54 (1H, d, J = 2.5 Hz); 7.51–7.43 (3H, m); 7.41–7.34 (1H, m); 6.97 (1H, d, J = 8.4 Hz); 4.10 (2H, t, J = 6.0 Hz); 3.56 (2H, t, J = 6.0 Hz); 2.30–2.17 (8H, m); 2.16–2.03 (5H, m); 1.90–1.77 (6H, m).

To a solution of the above intermediate (17 mg, 0.033 mmol) in THF (1 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (62 mg, 0.33 mmol) was added and the mixture was refluxed for 4 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. Product 25 was obtained in 20% yield (4 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃ aqueous solution). ¹H-NMR (400 MHz, CDCl₃) δ: 7.72–7.60 (6H, m); 7.53 (1H, d, J = 2.4 Hz); 7.51–7.42 (3H, m); 7.41–7.34 (1H, m); 6.97 (1H, d, J = 8.5 Hz); 4.07 (2H, t, J = 6.0 Hz); 2.60–2.53 (2H, m); 2.53–2.46 (4H, m); 2.37–2.30 (4H, m); 2.26 (12H, s); 2.24–2.20 (6H, m); 2.16–2.07 (3H, m); 1.96–1.86 (2H, m); 1.85–1.79 (6H, m); 1.77–1.61 (6H, m). ¹³C-NMR (600 MHz, CDCl₃) δ: 157.8; 141.0; 140.6; 139.2; 132.6; 128.8 (×2C); 127.4 (×2C); 127.2 (×3C); 127.0 (×2C); 125.6; 125.2; 112.2; 67.8, 57.9 (×2C); 53.9; 52.0 (×2C); 45.5 (×4C); 40.6 (×3C); 37.2 (×3C); 29.7; 29.1 (×3C); 27.6; 25.4 (×2C); 24.0.

N¹-(4-([1,1′-biphenyl]-4-yloxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (26). To a solution of 4-hydroxybiphenyl 42a (200 mg, 0.18 mmol) in acetone (6 mL), K₂CO₃ (815 mg, 5.9 mmol) and 1,4-dibromobuthane (1902 mg, 8.8 mmol) were added and the resulting mixture was refluxed for 3 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated again under reduced pressure. Product 43a 4-(4-bromobutoxy)-1,1′-biphenyl was obtained in 60% yield (216 mg) as a white solid upon crystallization from hexane. R_f = 0.71 (Hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 7.60–7.53 (4H, m); 7.47–7.42 (2H, m); 7.35–7.31 (1H, m); 4.08 (1H, t, J = 6.3 Hz); 3.54 (1H, t, J = 6.6 Hz); 2.16–2.09 (2H, m); 2.04–1.97 (2H, m).

To a solution of intermediate 43a (40 mg, 0.131 mmol) in dry DMF (2 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (245 mg, 1.31 mmol) was added and the mixture was stirred for 48 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated under reduced pressure. Product 26 was obtained in 24% yield (13 mg) as a sticky solid without further purification. R_f = 0.16 (CH₂Cl₂: MeOH = 9:1 with 1% of conc. NH₃ aqueous solution).¹H-NMR (600 MHz, CDCl₃) δ: 7.60–7.58 (2H, m); 7.56–7.52 (2H, m); 7.46–7.41 (2H, m); 7.35–7.30 (1H, m); 7.01–6.97 (2H, m); 4.04 (2H, t, J = 6.2 Hz); 2.52 (2H, t, J = 7.4 Hz); 2.50–2.34 (4H, m); 2.32–2.28 (4H, m); 2.33 (12H, s); 1.88–1.77 (6H, m); 1.69–1.60 (2H, m). ¹³C-NMR (150 MHz, CDCl₃) δ: 158.8; 141.0; 133.7; 128.8 (×2C); 128.2 (×2C); 126.8 (×2C); 126.7; 114.9 (×2C); 68.10, 58.2 (×2C); 52.3 (×2C); 45.7 (×4C); 27.5; 25.6 (×2C); 23.8.

N¹-(4-((4′-bromo-[1,1′-biphenyl]-4-yl)oxy)butyl)-N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (27). To a solution of 4′-bromo-(1,1′-biphenyl)-4-ol 42b (100 mg, 0.40 mmol) in acetone (3 mL), K₂CO₃ (276 mg, 2 mmol) and 1,4-dibromobuthane (605 mg, 2.8 mmol) were added and the resulting mixture was refluxed for 3 h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated again under reduced pressure. Product 43b was obtained in 65% yield (90 mg) as a white solid upon crystallization from petroleum ether. ¹H-NMR (600 MHz, CDCl₃) δ: 7.58–7.53 (2H, m); 7.52–7.47 82H, m); 7.46–7.40 (2H, m); 7.01–6.95 (2H, m); 4.06 (2H, t, J = 6.0 Hz); 3.53 (2H, t, J = 6.5 Hz); 2.16–2.07 (2H, m); 2.05–1.97 (2H, m).

To a solution of compound 43b (56 mg, 0.16 mmol) in THF (3 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (300 mg, 1.6 mmol) was added and the mixture was refluxed for 3 h. The reaction was diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. Product 27 was obtained in 26% yield (20 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃ aqueous solution). ¹H-NMR (600 MHz, CDCl₃) δ: 7.57–7.54 (2H, m); 7.53–7.48 (2H, m); 7.46–7.42 (2H, m); 7.01–6.95 (2H, m); 4.03 (2H, t, J = 6.5 Hz); 2.52 (2H, t, J = 7.4 Hz); 2.50–2.46 (4H, m); 2.33–2.27 (4H, m); 2.25 (12H, s); 1.87–1.80 (2H, m); 1.70–1.61 (6H, m). ¹³C-NMR (600 MHz, CDCl₃) δ: 158.9; 139.8; 132.3; 131.8 (×2C); 128.3 (×2C); 127.9 (×2C); 120.7; 114.9 (×2C); 67.9; 57.9 (×2C); 53.8; 52.0 (×2C); 45.5 (×4C); 27.8; 25.3 (×2C); 23.6.

N¹-(4-(2-((3r,5r,7r)-adamantan-1-yl)-4-bromophenoxy)butyl)-N1-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (28). To a solution of 2-((3r,5r,7r)-adamantan-1-yl)-4-bromophenol 42c [46] (200 mg, 0.65 mmol) in acetone (5 mL), K₂CO₃ (450 mg, 3.25 mmol) and 1,4-dibromobuthane (983 mg, 4.5 mmol) were added and the resulting mixture was refluxed for 5h. The reaction was concentrated, diluted with ethyl acetate, washed with H₂O, dried with Na₂SO₄ and concentrated again under reduced pressure. The intermediate (3r,5r,7r)-1-(5-bromo-2-(4-bromobutoxy)phenyl)adamantane 43c was obtained in 69% yield (198 mg) as a white solid upon crystallization from hexane. Rf = 0.78 (hex: AcOEt = 9:1). ¹H-NMR (600 MHz, CDCl₃) δ: 7.32 (1H, d, J = 2.4 Hz); 7.28 (1H, dd, J = 2.4 Hz, J = 8.7 Hz); 6.74 (1H, d, J = 8.7 Hz); 4.01 (2H, t, J = 6.0 Hz); 3.54 (2H, t, J = 6.5 Hz); 2.21–2.14 (2H, m); 2.12–2.09 (9H, m); 1.84–1.74 (6H, m).

To a solution of above compound 43c (126 mg, 0.285 mmol) in THF (5 mL), N¹-(3-(dimethylamino)propyl)-N³,N³-dimethylpropane-1,3-diamine (534 mg, 0.4 mmol) was added and the mixture was refluxed for 3h. The reaction was diluted with ethyl acetate, washed with H₂O, dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. Product 28 was obtained in 24% yield (38 mg) as a sticky solid upon chromatographic purification (CH₂Cl₂: MeOH: toluene = 4.5: 1.5: 4 with 1% of conc. NH₃ aqueous solution). ¹H-NMR (600 MHz, CDCl₃) δ: 7.321(1H, d, J = 2.5 Hz); 7.26 (1H, dd, J = 2.5 Hz, J = 8.6 Hz); 6.74 (1H, d, J = 8.6 Hz); 3.97 (2H, t, J = 6.2 Hz); 2.53 (2H, t, J = 7.2 Hz); 2.48 (4H, t, J = 7.0 Hz); 2.34–2.28 (4H, m); 2.25 (12 H, s); 2.14–2.06 (9H, m); 1.91–1.84 (2H, m); 1.83–1.75 (6H, m); 1.75–1.60 (6H, m). ¹³C-NMR (150 MHz, CDCl₃) δ: 157.3; 140.6; 129.9; 129.4; 113.7; 68.1; 58.1 (×2C); 54.0; 52.2 (×2C); 45.6 (×4C); 40.5 (×3C); 37.4; 37.2 (×3C); 29.1 (×3C); 27.6; 25.6 (×2C); 24.1.

3.2. Bacterial Strains and Growth Conditions

The bacterial strains used in this study are listed in Table S2. All strains were routinely grown at 37 °C in Muller Hinton II broth cation-adjusted (MHB-II) in shaking conditions (200 rpm), or Lysogeny Broth medium (LB) supplemented with 1.5% (w/v) agar. When specified, growth media were supplemented with L-arabinose or EDTA at the indicated concentrations.

Stock solutions of the adarotene derivatives were prepared in dimethyl sulfoxide (DMSO) at concentrations of 12.8 mM; stock solutions of 25% (w/v) L-arabinose, 0.5 M EDTA (pH 8.0), 10 mg/mL colistin, 1 mg/mL tobramycin were prepared in water; the stock solution of 1 mg/mL ciprofloxacin was prepared in 0.1 M HCl.

Growth assays in liquid cultures were performed as follows. Bacterial strains were grown at 37 °C in MHB-II with shaking (for the PAO1 lptE and lptH strains, the medium was supplemented with 0.5% (w/v) L-arabinose). After 16 h, cultures were diluted in fresh medium to an optical density at 600 nm wavelength (OD₆₀₀) of ≈0.001. 100 µL aliquots were dispensed into each well of 96-well microtiter plates, to which 100 µL of SPL207, colistin, ciprofloxacin, L-arabinose or EDTA at increasing concentrations were added. The OD₆₀₀ of the cultures was recorded every 2 h using an automated luminometer-spectrophotometer plate reader Spark10M (Tecan). Results were obtained from at least three independent experiments.

3.3. Antimicrobial Assays

The Minimal Inhibitory Concentration (MIC) of the compounds tested in this study was determined with the standard microdilution method, according to Clinical and Laboratory Standards Institute guidelines [47]. Bacterial strains were grown in MHB-II at 37 °C in shaking conditions. After 8 h of growth, the cultures were diluted in fresh medium at an OD₆₀₀ of ≈0.0005 (ca. 5 × 10⁵ CFU/mL) in 96-well microtiter plates in presence of increasing concentrations of each compound (or the solvent in which the compound was dissolved, as control). When required, L-arabinose or EDTA were added at the concentrations indicated in the text. The MIC values were visually evaluated after 24 h of incubation at 37 °C in static conditions. Results were obtained from at least three independent experiments.

Time–kill assays were performed as reported in [48]. Briefly, bacterial strains were inoculated in MHB-II and incubated at 37 °C with shaking. Overnight cultures were adjusted to an OD₆₀₀ of ≈0.0005 (ca. 5 × 10⁵ CFU/mL) in fresh medium supplemented or not with increasing concentrations of SPL207. Bacterial cultures were incubated at 37 °C in shaking conditions and 100 µL aliquots were harvest at different time points. The number of CFU/mL was determined by serially diluting the cultures in MHB-II and plating them on MHB-II agar plates for CFU count.

3.4. Cell Viability Assay

The human type II alveolar epithelial cell line A549 and the human immortalized keratinocyte cell line HaCaT were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and AddexBio (San Diego, CA, USA), respectively. Both cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with L-glutamine (2mM for A549 or 4mM for HaCaT cells), 10% fetal bovine serum (FBS) and 0.1 mg/mL penicillin-streptomycin.

The effect of compounds 7, 12, SPL207 and 18 on the viability of mammalian cells was determined by the inhibition of 3(4,5-dimethylthiazol-2yl)2,5-diphenyltetrazolium bromide (MTT) reduction to insoluble formazan. Cells suspended in the corresponding culture medium supplemented with glutamine and 2% FBS without antibiotics were plated in triplicate wells of a microtiter plate (4 × 10⁴ cells/well). After overnight incubation at 37 °C in a 5% CO₂ atmosphere, the medium was replaced with 100 μL fresh serum-free medium containing the compounds at different concentrations. The plate was incubated for 24 h at 37 °C in a 5% CO₂ atmosphere and the culture medium was removed and replaced with Hank’s buffer (136 mM NaCl; 4.2 mM Na₂HPO₄; 4.4 mM KH₂PO₄; 5.4 mM KCl; 4.1 mM NaHCO₃, pH 7.2, supplemented with 20 mM d-glucose) containing 0.5 mg/mL MTT. After 4 h incubation, the formazan crystals were dissolved by adding 100 μL of acidified isopropanol and viability was determined by absorbance measurements at 570 nm using a microplate reader (Infinite M200; Tecan, Salzburg, Austria). Cell viability was calculated with respect to untreated cells (in DMEM containing 1% DMSO).

3.5. Fluorescent Probe-Permeability Assays

Permeability assays were performed as previously described [49], with minor modifications. Briefly, overnight cultures were diluted in 5 mL of fresh MHB-II and grown for 6 h at 37 °C in shaking conditions. Then, bacteria were harvested by centrifugation, resuspended in 5 mM HEPES (pH 7.2) at an OD₆₀₀ of ≈1.0 and dispensed in 96-well black microtiter plates in the presence of increasing concentration of SPL207 and NPN or PI at final concentrations of 10 µM or 20 µg/mL, respectively. Fluorescence was measured in an automated luminometer-spectrophotometer plate reader Spark10M (Tecan) after 5 min or 60 min at room temperature (λ_ex 350 nm and λ_em 420 nm, for NPN; λ_ex 580 nm and λ_em 620 nm, for PI).

Membrane permeabilization was also visualized by confocal laser scanning microscopy, as previously described [30], with minor modifications. Overnight cultures were diluted in 5 mL of fresh MHB-II and grown for 6 h at 37 °C in shaking conditions. Then, bacteria were harvested by centrifugation, washed with Phosphate-Buffered Saline (PBS, 1X), resuspended in PBS at an OD₆₀₀ of ≈1.0, and dispensed in 96-well black microtiter plates in the presence of SYTO 9 and PI at final concentrations of 6 µM or 30 µM, respectively, in the presence or in the absence of 64 µM SPL207. After 15 min of incubation at room temperature in the dark, bacteria were washed with PBS and 10 µL aliquots of each sample were spotted on a microscope glass slide covered with 0.5% (w/v) agarose. Imaging was performed with a laser scanning confocal microscope Nikon A1R+, using 40× oil immersion objective. Images were acquired both in bright field and fluorescence channels using the following parameters: λ_ex 488 nm and λ_em ranging from 495 nm to 560 nm for SYTO 9; λ_ex 561 nm and λ_em ranging from 580 nm to 720 nm for PI. The NIS-elements software 6.1 was used to acquire and pre-process the images.

3.6. Checkerboard Assays

Checkerboard assays were performed as reported in [48]. Briefly, the antibiotic (i.e., colistin, ciprofloxacin, or tobramycin) was 2-fold diluted in MHB-II along the abscissa (x-axis), while SPL207 (or DMSO as control) was 2-fold diluted in MHB-II along the ordinate (y-axis), allowing the testing of all possible combinations of the two compounds. The strains were grown in MHB-II at 37 °C in shaking conditions. After 8 h of growth, 100 µL of cultures diluted in fresh medium at an OD₆₀₀ of ≈0.001 (ca. 1 × 10⁶ CFU/mL) was added to each well, previously filled with 100 µL containing the two compounds alone or in combination. After 20 h of incubation in static conditions at 37 °C, the MIC was determined as the lowest concentration of the antibiotic-SPL207 combination that showed no visible bacterial growth. Results were obtained from at least three independent experiments.

3.7. System Preparation for MD Simulations

The asymmetric OM model of P. aeruginosa was built with the CHARMM-GUI Membrane Builder Tool [50,51]. The outer leaflet of the OM consists of 60 molecules of P. aeruginosa type 1 lipid A, core 1a lipopolysaccharide (PA-LPS, 100%) from CHARMM-GUI, while the inner leaflet contains 164 molecules of 1-palmitoyl-2-oleoyl-phosphatidylethanolamine (POPE, 90%), 9 molecules of 1-palmitoyl-2-oleoyl-phosphatidylglycerol (POPG, 5%) and 9 molecules of cardiolipin (POCL1, 5%) for a total of 60 PA-LPS and 182 phospholipid molecules, in agreement with the literature [52,53,54,55,56,57,58,59,60,61,62,63]. The system has a box length of 106.70 Å in the xy dimension and it is surrounded by a water layer with a thickness of 50 Å. 300 Ca²⁺ ions were added to each system, while Na⁺ and Cl^- ions were further added at a concentration of 0.15 M by the CHARMM-GUI tool, up to charge neutrality. In total, there were approximately 180,349 atoms in each simulation system.

Adarotene 1 and SPL207 were sketched in 2D with the Picto software (version 4.5.4.1, OpenEye Cadence Molecular Sciences, Santa Fe, NM, USA) [64] and converted into a 3D structure with OMEGA (version 4.2.0.1, OpenEye Cadence Molecular Sciences, Santa Fe, NM) [65,66]. The protonation state of the molecule was predicted at pH 7.4 using the pK_a prediction software MoKa (Molecular Discovery, version 4.0.12) [39,67,68]; ligand energy minimization was performed with SZYBKI (version 2.5.0.1, OpenEye Cadence Molecular Sciences, Santa Fe, NM, USA) [69]. The two molecules were parameterized by the ligand reader and modeler module of CHARMM-GUI using the standard CHARMM force field (FF) [70]. The Multicomponent Assembler tool was used to randomly distribute the small molecules under investigation within the solvent area [71,72]. Following system construction with CHARMM-GUI, the topology and coordinate files were generated for AMBER [73,74].

3.8. MD Simulations

All-atom MD simulations were run with AMBER22, using a 2 fs time-step and a 10 Å non-bonded cut-off [75,76]. Each system underwent energy minimization for 50.000 steps (1.500 steps using the steepest descent algorithm, followed by 48.500 steps with the conjugate gradient algorithm). Subsequently, heating from 0 to 300 K was achieved over 900 ps at constant volume using the Langevin thermostat with a collision frequency of 2 ps⁻¹, further keeping the temperature at 300 K at constant volume for 100 ps. Box density was equilibrated at constant pressure and constant temperature (300 K) over 1 ns using the Berendsen barostat. Following density equilibration, a preliminary 50 ns MD simulation was conducted at constant pressure. Subsequently, MD trajectories were generated for 500 ns. Analysis of MD trajectories was performed using the CPPTRAJ [77] software (version 6.18.1) and python packages like MDanalysis and LiPyphilic [78,79,80]. Small molecules interactions with the membrane were visually inspected with PyMol [81].

3.9. Statistical Analysis

Statistical analysis was performed with the software GraphPad Prism 5, using one-way analysis of variance (ANOVA) followed by Tukey–Kramer multiple comparison tests. Differences having a p value < 0.05 were considered statistically significant.

4. Conclusions

The search for novel antimicrobials is critical to combating the escalating threat of antibiotic resistance. In this study, we conjugated a retinoid-like scaffold—distinct from any currently used antibiotic class and originally active only against Gram-positive bacteria—with protonatable residues to promote penetration of the Gram-negative OM. We investigated the influence of various functional groups on enhancing permeation through the Gram-negative OM and identified SPL207 as the most promising compound in the series. SPL207 was shown to compromise membrane integrity across all tested Gram-negative species. Notably, it exhibited strong synergistic activity in combination with colistin against colistin-resistant strains. MD simulations provided atomistic insights into SPL207–membrane interactions, supporting the proposed mechanism of OM destabilization. Further studies are warranted to elucidate the structural determinants required to optimize the activity of these adarotene analogs against Gram-negative pathogens.