Scientific Rationale

Bacteria utilize quorum sensing (QS) to assess the population density and species complexity of their environment and adapt their physiological behavior to the prevailing conditions. A large number of bacterial species utilize QS as a means to produce context specific gene products, such as virulence factors. Therefore, the ability to prevent virulence factor production could have a huge impact on human health. An example of a QS bacterium is the human pathogen Vibrio cholerae. The World Health Organization estimates that reported cases represent only 5–10% of the actual number of cholera incidents. There are an estimated 3 to 5 million cases that occur every year and approximately 120,000 deaths [6]. These high case numbers persist despite the development of a vaccine and a simple treatment regimen. In addition to V. cholerae, there are at least six other Vibrio species that are human pathogens. It has been determined that V. cholerae, as well as other bacteria, rely on a two-component system to relay exogenous information to the cell interior. A transmembrane histidine kinase (i.e. sensor) and an internal response regulator are the primary constituents of these two-component systems [7]. V. cholerae possess two parallel QS pathways. One pathway is responsible for interspecies communication wherein LuxPQ is the transmembrane receptor. A second, genus-specific pathway is controlled by the CqsS sensor [8]. Both pathways converge upon a common intermediary, LuxO [9]. V. cholerae is distinct from a number of bacterial species where activation of the QS pathway leads to inhibition of virulence factor production and not production of virulence factors.

The endogenous ligand of CqsS is the α-hydroxy ketone 1 (Figure 1), also known as cholera autoinducer 1 (CAI-1) [9-10]. Through careful studies of mutant CqsS receptors, it has been determined that CAI-1 inhibits His194 auto-phosphorylation of CqsS [11]. The auto-phosphorylation of His194 is believed to initiate a phosphorelay sequence involving the intermediary transfer protein LuxU and terminates at the response regulator LuxO (Figure 2). LuxO is a member of the NtrC family of response regulators that require interaction with a σ-54 factor to activate transcription [12]. LuxO binds DNA upstream of RNA polymerase and contains an inherent ATPase activity that is needed for changes to the transcriptional complex which involves DNA looping and subsequent formation of an open promoter complex [13]. LuxO integrates the CqsS and AI-2 signaling pathways and a homologue has been found in all Vibrio species examined. There are species-specific differences on the number of pathways that feed into LuxO.

Figure 1

The Endogenous Ligand of the Vibrio cholerae Sensor Histidine Kinase CqsS, CAI-1. Cholera autoinducer-1 (CAI-1)

Figure 2

The V. cholerae CqsS/CAI-1 Quorum Sensing Phosphorelay system. At low cell density, there are insufficient amounts of CAI-1 to adequately bind CqsS. This leads to the production of AphA and the adoption of individual cell behaviors. At high cell density, (more...)

When CAI-1 is present in minute quantities, the phosphorelay system favors the phosphorylation of LuxO at Asp47 (Figure 2) [11]. Phosphorylated LuxO up-regulates transcription of the small regulatory RNAs qrr1-4, which subsequently promote translation of AphA while simultaneously suppressing translation of the HapR regulator. AphA regulates genes that are beneficial to individual behaviors, while HapR regulates genes that promote group (quorum) behaviors. Another level of regulating the RNA is the RNA chaperone Hfq that base pairs with and destabilizes the hapR mRNA [12]. When dense populations of V. cholerae synthesize and secrete sufficient quantities of CAI-1, this ligand will bind to CqsS and promote the production of HapR. The induction of HapR expression then allows the colonizing V. cholerae population to detach from the gut epithelium and vacate its host. Among the genes regulated by HapR is the expression of haemagglutinin protease (HapA) that facilitates the expulsion of V. cholerae from its human host [14]. In most QS bacterial species, QS induces virulence factor production but Vibrio cholerae is distinct from other non-vibrio species because QS results in repression of virulence factor and biofilm production [2]. From a therapeutic perspective, the artificial activation of CqsS or inhibition of LuxO or Hfq to promote HapR expression could lead to a new generation of anti-cholera treatments. A potential combinatorial therapeutic approach targeting multiple targets in the QS pathway could ensure effective and a genus-specific antimicrobial response.

Antagonism of LuxO has not been well studied in the literature thus far. The only known small molecule inhibitors of LuxO were reported after the initiation of this project [2]. Derived from an aza-uracil scaffold, these compounds show moderate inhibition of LuxO activity (IC₅₀'s range from 5 to 20 µM) in cellular assays. Further investigation of the most potent compound, (Figure 3, Aza-U), determined that these LuxO antagonists block ATPase hydrolysis and prevent subsequent transcriptional activation. Aza-U was selected as the positive control for all cell-based SAR assays in the post-HTS medicinal chemistry phase.

Figure 3

Aza-uracil-derived antagonists of LuxO ATPase activity. Aza-U (4) and its analogs have been shown to inhibit the ATPase activity of LuxO. Quorum-sensing modulation was measured in V. cholera strains BH1578 and BH1651 (see Section 2) and their respective (more...)

It is hoped that the identification of additional small molecule modulators of LuxO could elucidate the mechanisms of the latter steps underpinning the V. cholerae QS phosphorelay system. The additional chemical matter will also assist in the structural characterization of potential binding pockets and, in combination with a clearer understanding of how the transcriptional activation occurs, could guide development of new QS agonists as potential anti-cholera agents. Since all vibrios possess a LuxO homologue, LuxO inhibitors should be efficacious against all pathogenic vibrios. Aza-U compounds were found to inhibit LuxO activity in V. harveyi and V. parahaemolyticus [2].

Materials and Reagents

• CellTiter-Glo^® Luminescent Cell Viability Assay was purchased from Promega (Catalog No. G7573; Madison, WI)

Bacterial strains & Cell Lines

The following cell lines were used in this study:

• Vibrio cholerae BH1578; a genetically modified strain lacking LuxS and CqsA autoinducer synthases that was provided by the Bassler lab. This strain was used in the primary assay.
• Vibrio cholerae BH1651; a genetically modified strain with constitutively active LuxO that was provided by the Bassler lab.
• Vibrio cholerae DH231; a genetically modified strain lacking the CqsS receptor that was provided by the Bassler lab.
• Vibrio cholerae WN1103; a genetically modified strain lacking the LuxQ receptor that was provided by the Bassler lab.
• HeLa obtained from ATCC (Catalog Number CCL-2; Manassas, VA) is a human epithelial adenocarcinoma cell line used for mammalian cytotoxicity profiling

Note: All of the V. cholerae strains also carry the heterologous V. harveyi luxCDABE luciferase operon

2.1. Assays

A summary listing completed assays and their corresponding PubChem AID numbers is provided in Appendix A. Refer to Appendix B for the detailed assay protocols.

2.2. Probe Chemical Characterization

After preparation as described in Section 2.3, the probe ML366 was analyzed by UPLC, ¹H and ¹³C NMR spectroscopy, and high-resolution mass spectrometry. The data obtained from NMR and mass spectroscopy are consistent with the structure of the probe, and UPLC indicates an isolated purity of greater than 98%. The complete synthetic protocol is provided in Appendix C, and associated spectral data are provided in Appendix E.

The solubility of ML366 was experimentally determined to be less than 1 µM in phosphate buffered saline (PBS) with 1% (v/v) DMSO. Solubility was also measured in the Luria Broth (LB) media used in the primary assay. Table 2 summarizes the solubility of ML366 and several analogs in both PBS and LB media. In general, LB media is a better solvent than PBS for these compounds.

Table 2

Solubility of ML366 and Analogs in PBS and LB Assay Media.

The probe is stable in PBS solution (>99% remaining after a 48-hour incubation at 23 °C). The data from the PBS stability assay is provided in Figure 4. The probe is also stable to human plasma with greater than 75% remaining after incubation at 37 °C for 5 hours. Murine plasma degrades the probe rapidly as less than 1% of compound was remaining after a 5-hour incubation. No significant reaction with glutathione (GSH) was observed for ML366. Table 3 summarizes the various stability assays performed. Experimental procedures for all analytical assays are provided in Appendix D. The physical properties of probe ML366 are summarized in Table 4.

Figure 4

Stability of Probe ML366 in PBS Buffer (pH 7.4, 23°C).

Table 3

Plasma Stability and Plasma Binding of ML366.

Table 4

Summary of Probe Properties Computed for ML366.

2.3. Probe Preparation

ML366 and most associated analogs were prepared according to the route outlined in Scheme 1. Commercially available 2-amino-oxadiazole 6 was reacted with an excess of cyclohexylcarboxylic acid under peptide coupling conditions. Upon complete consumption of starting material, the product was selectively precipitated by the addition of water.

Scheme 1

Synthesis of Probe ML366.

Full experimental details for the preparation of ML366 are provided in Appendix C.

Probe attributes

• Induces light production in V. cholerae strain BH1578 with EC₅₀ ≤ 10 µM
• Induces light production in V. cholerae strains BH1651, DH231 & WN1103
• Non-toxic to HeLa cells with IC₅₀ ≥ 30 µM

The current project utilizes several genetically modified strains of the human pathogen Vibrio cholerae in an effort to identify potential activators of the V. cholerae quorum sensing (QS) network. The MLSMR collection of 352,083 compounds was examined for possible QS activators, using V. cholerae BH1578 as the primary test strain. In the V. cholerae BH1578 strain, the V. harveyi luxCDABE (luciferase) operon was introduced as a cosmid, and the operon is activated by endogenous mechanisms to elicit light upon reaching a quorum. In addition, the BH1578 strain is a cqsA, luxS double deletion mutant that lacks the ability to produce any endogenous autoinducers of QS. Therefore, light production can only occur in response to an activating screening compound. Secondary assays utilizing additional V. cholerae mutants were included in order to triage active compounds by their potential mode of action. Based on their behavior in the various secondary assays, compounds were categorized as an agonist of either CqsS or LuxQ; an inhibitor of LuxO; or as an inhibitor of Hfq.

Analysis of the primary screening data revealed that approximately 500 compounds could activate the V. cholerae QS pathway. Re-evaluation of these compounds in a dose-response format led to the selection of 26 compounds for subsequent examination as commercially-made dry powders. Based on their cellular potency in BH1578 and low toxicity towards HeLa cells, two scaffolds (CID 4443990 and CID 665390) were selected for further development as LuxO inhibitors. The optimization of CID 665390 to ML370 was undertaken by the University of Kansas MLPCN Chemistry Center, and the results of those studies will be reported separately.

The substituted 2-amino-oxadiazole CID 4443990 (Figure 5) is an activator of V. cholerae QS with a measured AC₅₀ = 5.1 µM for the commercial material. Re-synthesized CID 4443990 is equipotent to the commercial sample (AC₅₀ = 6.1 µM), and neither demonstrated observable toxicity to HeLa cells (IC₅₀ >35 µM). Analogs of CID 4443990 were designed, prepared, and evaluated for QS activation. The results of the structure-activity relationship (SAR) studies are summarized in Section 3.4.

Figure 5

Representative Dose Response Curves for Samples of CID 4443990. CID 4443990 was tested across a range of concentrations up to 35 µM in the primary assay and the BH1651 assay. Concentration response curves were generated with Genedata Screener (more...)

3.1. Summary of Screening Results

As described above, the Molecular Libraries and Small Molecules Repository (MLSMR) collection of 352,083 unique compounds were tested at a single concentration (20 µM) for their ability to induce the production of luciferase by BH1578, thus indicating successful activation of the QS signaling pathway. 553 compounds were active at 20 µM, representing a nominal hit rate of 0.15%.

The active compounds were then re-tested in a dose-response format, wherein only 2 compounds actually met or exceeded the project requirement of AC₅₀ ≤ 1.0 µM. However, the low success rate of the primary assay prompted us to expand the primary assay cut-off to include substances with AC₅₀ ≤ 10 µM.

Subsequently, dry powders for 26 compounds were procured for re-testing. After purity analysis and structural confirmation by NMR spectroscopy, these substances were screened according to the workflow outlined in Figure 6.

Figure 6

Critical Path for Probe Development.

The assay progression outlined in Figure 6 assessed the activation of QS pathways in several V. cholerae strains: BH1578, BH1651, and SLS353. CID 4443990 emerged as a promising LuxO inhibitor and was subsequently prioritized for development. A number of analogs were prepared and assayed for QS activation. The results of the SAR investigation are presented in detail in Section 3.4.

3.2. Representative Dose Response Curves for Probe ML366

Figure 7ML366 Induces Quorum Sensing via LuxO ATPase Inhibition without Toxicity

ML366 was tested across a range of concentrations up to 35 µM in the primary BH1578 assay, the SLS353 qrr4:GFP reporter assay, the in vitro LuxO ATPase assay and a HeLa cytotoxicity assay. Concentration response curves were generated with Genedata Screener Condoseo and show normalized percent activity for the individual doses. DH231 assay (AID 686931), IC₅₀ = 2.1 µM (A); SLS353 assay (AID 686942) (B); ML366 in vitro LuxO ATPase assay (AID 686936) (C); ML366 HeLa CellTiter-Glo (AID 686928), IC₅₀ ≥ 35 µM (D). ○ = replicate 1, △ = replicate 2. RFU=relative fluorescence unit.

3.3. Scaffold/Moiety Chemical Liabilities

ML366 possesses no obvious chemical liabilities, and a phosphate-buffered saline solution of the probe showed no significant decomposition over a 48-hour period. In addition, glutathione did not react with the probe appreciably after six hours. Exposure to human plasma did not affect the probe, but it decomposed rapidly upon treatment with murine plasma. These results are summarized in Section 2.2.

A search of PubChem for ML366 shows that this compound has been evaluated in a total of 583 assays. There are 15 assays unrelated to this project in which ML366 showed a reported AC₅₀ below 20 µM. Table 5 lists the 15 assays, the PubChem activity outcome, and the measured activity of ML366. It should be noted that the assay results include active, inactive or inconclusive categories. Inconclusive data are cases where data has too few data points or no dose dependent response. By viewing the dose curves, the term ‘active’ proves to be misleading for some of the results. For example, the NPC activator assay (AID 485313) dose curve reveals a partial curve with only 4 data points where only the top concentration shows an appreciable increase in signal that does not meet 100% activity for that assay. The Rab9 assay (AID 484297) shows a potent activation of signal for this luciferase reporter but there has been no subsequent data reported for this compound within this particular project nearly three years after this data was submitted to PubChem (as of 23 July 2013).

Table 5

PubChem Assays Reporting Dose-Dependent Activity of ML366.

The mild inhibition of firefly luciferase (Table 5, entry 4) by ML366 is not a concern for this project. Despite its nominal potency of 2.7 µM against purified enzyme, ML366 only achieves ~35% maximum inhibition of function. There is no indication of firefly luciferase inhibition in a cellular context, and the V. cholerae assays studied in the current project express the naturally occurring V. harveyi luciferase protein. An alignment of the firefly versus V. harveyi protein sequences using an online protein alignment tool, CLUSTAL Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/), indicates that the eukaryotic and prokaryotic luciferases do not possess significant homology (data not shown). In addition, the primary assay relies upon a gain-of-signal readout; potent luciferase inhibitors would inherently be removed by the assay design. Finally, our HeLa cytotoxicity assay uses CellTiter Glo, which utilizes firefly luciferase and there is no reduction in signal observed, even at the top concentration of 35 µM (Figure 7D).

ML366 was screened for toxicity towards HeLa cells, and no adverse effects were observed below 35 µM concentration. The reported activity against ELG1 in modified HEK293T cells (Table 5, entry 10) does not suggest mammalian toxicity because the treatment leads to an increase in luciferase signal.

A structure-based search (80% similarity to ML366) in SciFinder reveals that several related 2-amino-thiadiazoles are potent and selective EP₃ receptor antagonists [17]. The function of the EP₃ receptor has not been fully elucidated, but it is implicated in gastric acid secretion as well as uterine, bladder, and smooth muscle contractions [17]. Figure 8 shows representative aminothiadiazoles.

Figure 8

Aminothiadiazole-containing Inhibitors of the EP₃ Receptor.

A similar search (80% structural similarity) of the Thomson Reuters Integrity database did not identify any publications or clinical trials describing the probe or related compounds.

3.4. SAR Tables

In order to investigate the activity of the hit compound CID 4443990, 29 structurally related analogs were synthesized and evaluated for their ability to initiate the QS pathway in the BH1578 V. cholerae strain. The biological assay data of these analogs are presented in Tables 6-8. Characterization data (¹H NMR spectra and UPLC chromatograms) of these analogs are provided in Appendix G.

Table 6

Examination of the Benzodioxane System.

Table 8

Different Combinations of Amides and Benzodioxane Replacements.

The initial oxadiazole hit CID 4443990 shows acceptable potency in the primary BH1578 cell line and is capable of light induction superior to the Aza-U positive control (Table 6, entry 1). Enhancing cellular potency and aqueous solubility (cf. Section 2.2) were prioritized goals for SAR exploration. The murine plasma instability, tentatively attributed to the amide functionality, was deemed a secondary objective for optimization.

It was hypothesized that the 1,4-benzodioxane heterocycle is a contributing factor for the poor aqueous solubility of the hit. Accordingly, less substituted aromatic rings were evaluated in place of the larger ring system. Reducing this group to an anisole (Table 6, entries 2-4) or chlorobenzene (Table 6, entries 5-7) affords minor gains in solubility, with the ortho-substituted analogs being the most soluble compounds. Presumably, disrupting the planarity of the overall structure contributes to the higher solubility. Two compounds from this phenyl series elicit only a partial response in BH1578 cells (Table 6, entries 2 & 4), and the remaining four are inactive. Introduction of pyridyl and cyclopentyl rings produces soluble, but inactive, compounds (Table 6, entries 8-9).

Examination of the cyclohexyl amide was undertaken next and the results are summarized in Table 7. While retaining the benzodioxane portion, a variety of cycloalkane amides were prepared (Table 7, entries 2-8). The cycloheptyl amide (Table 7, entry 5) was the largest ring evaluated and is the only compound of this amide series displaying any cellular activity. Surprisingly, reducing the ring size significantly increases aqueous solubility (Table 7, entries 2-3). This finding, as well as the high solubility of the piperidine and tetrahydropyran analogs (Table 7, entries 6-7), suggests that the benzodioxane is not the primary contributor to the hit's poor solubility. The aliphatic amides and amine prepared (Table 7, entries 9-11) are inactive in the primary assay as are the phenyl sulfonamides (Table 7, entries 12-13). Imine intermediates and the free amine were also evaluated (Table 7, entries 14-16), but none of these compounds show measurable induction of V. cholera quorum sensing.

Table 7

Biological Activity of Various Amides and Amine Substituents.

Several compounds bearing various combinations of amides and benzodioxane replacements were synthesized (Table 8). While most have increased aqueous solubility, none are as efficacious as the original hit.

The SAR is flat with regards to cellular potency; modification of the cyclohexyl amide or the benzodioxane region strongly diminishes quorum sensing agonism. However, aqueous solubility is readily obtained via reducing the carbon content of the amide. Because the more soluble compounds underperform in the primary assay, the original hit (CID 4443990) was selected as probe ML366.

3.5. Cellular Activity

The primary assay and several secondary assays were performed with intact Vibrio cholerae bacteria and compounds were capable of inhibiting an intracellular kinase, LuxO. This cell-based activity suggests that the compounds are able to penetrate the cell and act on an intracellular target. In addition, compounds were tested in a cytotoxicity screen utilizing mammalian HeLa cells without apparent cytotoxicity. An overview of these assays is provided in Section 2.1, and full experimental details can be found in Appendix B. The probe ML366 satisfies the cellular activity criteria specified for this project (refer to Section 4, Table 9).

Table 9

Comparison of the performance of ML366 to Project Criteria.

3.6. Profiling Assays

Because the tentative mode of action for ML366 is ATPase inhibition, it was also tested at 10 µM in Millipore's KinaseProfiler for potential cross-reactivity towards human kinases. ML366 was remarkably clean against the 59 human kinases tested. The most inhibition seen was at 39% inhibition with CK1γ1 and 29% inhibition of mTOR (See Appendix H, Table H1). In addition, ML366 was submitted to Eurofins Panlabs for evaluation in their LeadProfilingScreen to determine possible off-target effects against a broad panel of receptors, ion channels, and transporters. ML366 had a very clean profile with no binding in most assays and weak activity against Adenosine A3 (38%), Dopamine D_4.2 (36%) and Melatonin MT1 (34%). Data from the kinase profiling assays are available in Appendix H, Table H2.

Table H1

Human kinase profiling results for ML366.

Table H2

Radioligand binding assay results for ML366.

4.1. Comparison to existing art and how the new probe is an improvement

A recently identified series of aza-uracil-containing compounds are the only known, reported inhibitors of V. cholerae LuxO [2]. The most potent member of this family, designated Aza-U (cf. Figure 3), was synthetically prepared and used as the positive control for all cell-based assays. ML366 is more potent than Aza-U in the primary assay when compared directly (EC₅₀ = 6.1 µM and >35 µM, respectively). Consistently across all of the assays, ML366 performed better than Aza-U (Figure 5C, Figure 7B, Table 10 & 11). In BH1578, BH1651, and DH231, only the positive control dose of 40 µM induces light production and the lower doses do not generate a consistent dose-dependent response (see Figures 5C & 7B). In the WN1103 assay, ML366 is more potent compared to Aza-U (2.1 to 21 µM, respectively, Table 11). In addition to Aza-U, ML366 was compared to the natural QS ligand, CAI-1, as well as two other small molecule CqsS agonists previously reported by our group in late 2012 and the results of this comparison are summarized in Table 10. It is anticipated that the current LuxO probes ML366 and ML370 will provide valuable information regarding available binding pockets of V. cholera LuxO and guide the development of the next generation of quorum sensing focused anti-vibrio therapeutics.

Table 10

Comparison of Project Probes to Natural Ligand and Prior Art.

Table 11

Probe Activation of Quorum Sensing in Vibrio cholerae Mutant Strains. ML366 and several analogs were evaluated in alternative V. cholerae mutants to establish possible modes of action. The reported EC₅₀ values are the average of at least two independent (more...)

Before the development of these chemical probes, CAI-1, the endogenous ligand of CqsS was the best small molecule available for inhibiting quorum sensing in Vibrio cholera. The long hydrophobic tail of CAI-1 leads to significant micelle formation making it a less attractive chemical tool. All the probes developed in this project (Table 10) including ML366 do not have structural features that promote micelle formation and we expect these probes to have broad utility.

4.2. Mechanism of Action Studies

ML366 was evaluated for its ability to activate the quorum sensing pathways of additional V. cholerae strains (Table 11). Strain BH1651 expresses a constitutively active LuxO^D47E mutant (PubChem AID 651847) as described in Section 2.1.3. For this particular bacterial strain, LuxO is locked into an active conformation and thus the signaling processes downstream of LuxO are actively diverted to promote AphA expression. Consequently, HapR and luciferase expression is continuously repressed independent of CqsS or LuxPQ signaling. The ability of ML366 to promote light production in BH1651 suggests the compound acts upon or downstream of LuxO, consistent with the putative activity of LuxO inhibitors. V. cholerae DH231 and WN1103 are mutant strains lacking CqsS and LuxQ membrane receptors, respectively. If ML366 is working downstream of either CqsS or LuxPQ, the DH231 and WN1103 cells should activate light production with compound treatment. This predicted phenotype is precisely what is observed (Table 11, entries 3-4). This activity profile appears to be conserved across both LuxO inhibitors (ML366 and ML370); QS agonists like CAI-1, ML343 and ML344 activate V. cholerae QS only if the cells express the CqsS receptor. Compared to DMSO, ML366 significantly decreases LuxO ATPase activity in vitro (cf. Figure 7C). This suggests that ML366 inhibits LuxO in a similar manner to Aza-U by directly targeting LuxO itself [2].

4.3. Planned Future Studies

In addition to V. cholerae, there are at least six Vibrio species that cause disease in humans and are contracted through contaminated water or ingested during consumption of seafood [17]. Aza-U inhibits LuxO in multiple Vibrio species, including V. harveyi and V. haemolyticus [2]. ML366 will be tested against these and multiple vibrios for the ability to shutdown virulence factor production. Combinatorial therapies have proven efficacious in a number of infectious disease treatment regimens and could also be applicable in treating cholera. ML366 will be tested in combination with several CqsS agonists and the natural CAI-1 ligand. ML366 and the other probes will also be tested in animal models of cholera. Since LuxO is a kinase, ML366 was tested against a panel of 60 human kinases representing multiple branches of the kinome to gauge potential host-pathogen specificity (see Appendix H, Table H1). Identification and characterization of the binding site(s) occupied by ML366 will be undertaken. These studies will include analysis with several LuxO mutants in cells and using in vitro kinase and ATPase assays. Other experiments could include structural efforts. This structural information will undoubtedly be beneficial in the development of additional LuxO inhibitors as potential anti-cholera therapeutics. Co-crystallization of LuxO with the probe is the most attractive possibility for successful crystallization, but protein NMR binding studies are also an option. Other methods to determine probe-protein interaction, albeit with less precision, are biophysical experiments to determine direct binding of the compound to the LuxO protein. Differential scanning fluorimetry (DSF) or thermal shift is a means of detecting binding of a ligand to purified protein. This technique is routinely used in our laboratory and could be applied to LuxO.

ML366 would benefit from additional chemical optimization to improve solubility and murine plasma stability. Also, evaluation of ML366 in a panel of in vitro ADME assays and potential optimization of ADME properties to enable in vivo studies would be beneficial. From a structural perspective, investigating additional heteroaromatic rings may be necessary (Figure 10). It is likely that the lipophilic cyclohexyl amide cannot be readily manipulated, but it may be possible to replace the central oxadiazole with more soluble heterocycles such as pyridine or pyrazine. Additional five-membered heterocycles should also be considered. Reduction of the eastern benzodioxane was not tolerated, so bicyclic replacements will be attempted. Substituted quinolines or purines may provide the desired potency and solubility. Following the example of Hilfiker and coworkers [18], benzomorpholines and benzopiperazines may also be viable alternatives.

Figure 10

Possible directions for future SAR investigation of ML366.

V. cholerae BH1578 primary bioluminescence assay (2132-01) (AID 588346, AID 602243, AID 624270, AID 651854, AID 651809, AID 651816, AID 652239, AID 686929)

Materials and Reagents

• Bacterial strain: BH1578 (V. cholerae ΔcqsA ΔluxS carrying pBB1 cosmid, which contains the V. harveyi luxCDABE luciferase operon)
• LB Medium: Dissolve in 10 g/L Tryptone, 5 g/L Yeast Extract, and 10 g/L NaCl in distilled water, autoclave, store at room temperature
• Tetracycline (10 mg/mL): Dissolve 10 mg tetracycline in 1 mL 100% ethanol, store at -20 °C, protect from light.
• LB/tet: add 1 mL tetracycline (10 mg/mL) to 1 L of LB medium. Final concentration of tetracycline is 10 µg/mL. Make it fresh for every use.
• CAI-1 stock: Dissolve CAI-1 in DMSO to 50 mM (10.7 mg/mL), store at -20 °C

Procedure

1. Grow up BH1578 reporter strain in 50 mL LB/Tet at 30 °C for >16 hours with shaking (200 rpm). The final OD₆₀₀ of each culture should be > 3.0.
2. Adjust culture to OD₆₀₀ = 0.3, mix well. (Note: a low speed centrifugation (200 rpm for 1 min) removed most biofilm aggregates).
3. Add 20 µL LB/tet per well with Thermo Combi MultiDrop fluid dispenser into a 384 well plate. (Greiner black, clear bottom plates were used for HTS while Corning 8867BC white opaque 384 plates were used for subsequent assays).
4. Pin transfer 100 nL compound or 1 µM CAI-1 as positive control per well from compound source plate to assay plate (pinning volume was 150 nL for the primary HTS).
5. Dispense 10 µL of diluted culture into each well of a 384 well plate.
6. Incubate the plates at 30 °C without shaking for 6 hours.
7. Measure bioluminescence (Lum(384)) and OD₆₀₀ in a Perkin-Elmer Envision Multilabel Reader

For dose retests, compounds were arrayed so that 2 rows of DMSO were located between each test compound. This minimized any signal from a neighboring well yielding a false positive due to an extremely high bioluminescence signal.

V. cholerae BH1651 secondary LuxO inhibitor bioluminescence assay (2132-03) (AID 624254, AID 624269, AID 651847, AID 652289)

Materials and Reagents

• Bacterial strain: BH1651 (V. cholerae luxO^D47E carrying pBB1 cosmid, which contains the V. harveyi luxCDABE luciferase operon)
• LB Medium: Dissolve in 10 g/L Tryptone, 5 g/L Yeast Extract, and 10 g/L NaCl in distilled water, autoclave, store at room temperature
• Tetracycline (10 mg/mL): Dissolve 10 mg tetracycline in 1 mL 100% ethanol, store at -20 °C, protect from light.
• LB/tet: add 1 mL tetracycline (10 mg/mL) to 1 L of LB medium. Final concentration of tetracycline is 10 µg/mL. Make it fresh for every use.

Procedure

1. Grow up BH1651 reporter strain in 50 mL LB/Tet at 30 °C for >16 hours with shaking (200 rpm). The final OD₆₀₀ of each culture should be > 3.0.
2. Adjust culture to OD₆₀₀ = 0.3, mix well. (Note: a low speed centrifugation (200 rpm for 1 min) removed most biofilm aggregates).
3. Add 20 µL LB/tet per well with Thermo Combi MultiDrop fluid dispenser into a 384 well plate. (Greiner black, clear bottom plates were used for HTS while Corning 8867BC white opaque 384 plates were used for subsequent assays).
4. Pin transfer 100 nL compound per well from compound source plate to assay plate.
5. Dispense 10 µL of diluted culture into each well of a 384 well plate.
6. Incubate the plates at 30 °C without shaking for 6 hours.
7. Measure bioluminescence (Lum(384)) and OD₆₀₀ in a Perkin-Elmer Envision Multilabel Reader

For dose retests, compounds were arrayed so that 2 rows of DMSO were located between each test compound. This minimized any signal from a neighboring well yielding a false positive due to an extremely high bioluminescence signal.

V. cholerae DH231 secondary sensor mechanism bioluminescence assay (2132-04) (AID 624281, AID 651808, AID 686931)

Materials and Reagents

• Bacterial strain: DH231 (V. cholerae ΔcqsS ΔluxS carrying pBB1 cosmid, which contains the V. harveyi luxCDABE luciferase operon)
• LB Medium: Dissolve in 10 g/L Tryptone, 5 g/L Yeast Extract, and 10 g/L NaCl in distilled water, autoclave, store at room temperature
• Tetracycline (10 mg/mL): Dissolve 10 mg tetracycline in 1 mL 100% ethanol, store at -20 °C, protect from light.
• LB/tet: add 1 mL tetracycline (10 mg/mL) to 1 L of LB medium. Final concentration of tetracycline is 10 µg/mL. Make it fresh for every use.
• CAI-1 stock: Dissolve CAI-1 in DMSO to 50 mM (10.7 mg/mL), store at -20 °C

Procedure

1. Grow up reporter strain in 50 mL LB/Tet at 30 °C for >16 hours with shaking (200 rpm). The final OD₆₀₀ of each culture should be > 3.0.
2. Adjust culture to OD₆₀₀ = 0.3, mix well. (Note: a low speed centrifugation (200 rpm for 1 min) removed most biofilm aggregates).
3. Add 20 µL LB/tet per well with Thermo Combi MultiDrop fluid dispenser into a 384 well plate. (Corning 8867BC white opaque 384 plates were used).
4. Pin transfer 100 nL compound per well or 1 µM CAI-1 as positive control from compound source plate to assay plate.
5. Dispense 10 µL of diluted culture into each well of a 384 well plate.
6. Incubate the plates at 30 °C without shaking for 6 hours.
7. Measure bioluminescence (Lum(384)) and OD₆₀₀ in a Perkin-Elmer Envision Multilabel Reader

For dose retests, compounds were arrayed so that 2 rows of DMSO were located between each test compound. This minimized any signal from a neighboring well yielding a false positive due to an extremely high bioluminescence signal.

V. cholerae WN1103 Secondary sensor mechanism bioluminescence assay (2132-05) (AID 624275, AID 651841, AID 686930)

Materials and Reagents

• Bacterial strain: WN1103 (V. cholerae ΔcqsA ΔluxQ carrying pBB1 cosmid, which contains the V. harveyi luxCDABE luciferase operon)
• LB Medium: Dissolve in 10 g/L Tryptone, 5 g/L Yeast Extract, and 10 g/L NaCl in distilled water, autoclave, store at room temperature
• Tetracycline (10 mg/mL): Dissolve 10 mg tetracycline in 1 mL 100% ethanol, store at -20 °C, protect from light.
• LB/tet: add 1 mL tetracycline (10 mg/mL) to 1 L of LB medium. Final concentration of tetracycline is 10 µg/mL. Make it fresh for every use.
• CAI-1 stock: Dissolve CAI-1 in DMSO to 50 mM (10.7 mg/mL), store at -20 °C

Procedure

1. Grow up reporter strain in 50 mL LB/Tet at 30 °C for >16 hours with shaking (200 rpm). The final OD₆₀₀ of each culture should be > 3.0.
2. Adjust culture to OD₆₀₀ = 0.3, mix well. (Note: a low speed centrifugation (200 rpm for 1 min) removed most biofilm aggregates).
3. Add 20 µL LB/tet per well with Thermo Combi MultiDrop fluid dispenser into a 384 well plate. (Corning white opaque 8867BC 384 plates were used).
4. Pin transfer 100 nL compound or 1 µM CAI-1 as positive control per well from compound source plate to assay plate.
5. Dispense 10 µL of diluted culture into each well of a 384 well plate.
6. Incubate the plates at 30 °C without shaking for 6 hours.
7. Measure bioluminescence (Lum(384)) and OD₆₀₀ in a Perkin-Elmer Envision Multilabel Reader

For dose retests, compounds were arrayed so that 2 rows of DMSO were located between each test compound. This minimized any signal from a neighboring well yielding a false positive due to an extremely high bioluminescence signal.

V. cholerae SLS353 secondary qrr4:GFP fluorescence reporter assay (2132-06) (AID 652219, AID 686942)

Materials and Reagents

• Bacterial strain: SLS353 (V. cholerae luxO^D47E carrying qrr4:GFP plasmid)
• LB Medium: Dissolve in 10 g/L Tryptone, 5 g/L Yeast Extract, and 10 g/L NaCl in distilled water, autoclave, store at room temperature
• Kanamycin (100 mg/mL): Dissolve 100 mg kanamycin in 1 mL water, store at -20 °C
• LB/Kan: add 1 mL Kanamycin (100 mg/mL) to 1 L of LB medium. Final concentration of Kanamycin is 100 µg/mL. Make it fresh for every use.

Procedure

8.

Grow up strain in 50 mL LB/Kan at 30 °C for >16 hours with shaking (200 rpm). The final OD₆₀₀ of each culture should be > 3.0.

9.

Adjust culture to OD₆₀₀ = 0.3, mix well. (Note: a low speed centrifugation (200 rpm for 1 min) removed most biofilm aggregates).

10.

Add 20 µL LB/Kan per well with Thermo Combi MultiDrop fluid dispenser into a 384 well plate. (Corning black opaque 384 plates were used).

11.

Pin transfer 100 nL compound per well from compound source plate to assay plate.

12.

Dispense 10 µL of diluted SLS353 culture into each well of a 384 well plate.

13.

Incubate the plates at 30 °C without shaking for 16 hours.

14.

Measure GFP fluorescence in a Perkin-Elmer Envision Multilabel Reader

LuxO in vitro ATPase assay (AID No. 686936)

Data Analysis

For the primary screen and other assays, negative-control (NC) wells and positive-control (PC) wells were included on every plate. The raw signals of the plate wells were normalized using the ‘Stimulators Minus Neutral Controls’ or the ‘Neutral Controls’ method (when no positive control was available) in GeneData Screener Assay Analyzer (v7.0.3 & v10.0.2). The median raw signal of the intra-plate NC wells was set to a normalized activity value of 0, while the median raw signal of the intra-plate PC wells was set to a normalized activity value of 100. Experimental wells were scaled to this range, resulting in an activity score representing the percent change in signal relative to the intra-plate controls. The mean of the replicate percent activities were presented as the final ‘PubChem Activity Score’. The ‘PubChem Activity Outcome’ class was assigned as described below, based on an activity threshold of 70%:

• Activity_Outcome = 1 (inactive), less than half of the replicates fell outside the threshold.
• Activity_Outcome = 2 (active), all of the replicates fell outside the threshold, OR at least half of the replicates fell outside the threshold AND the ‘PubChem Activity Score’ fell outside the threshold.
• Activity_Outcome = 3 (inconclusive), at least half of the replicates fell outside the threshold AND the ‘PubChem Activity Score did not fall outside the threshold.

Human kinase panel results

ML366 was profiled against a panel of 60 human kinases by Merck Millipore's KinaseProfiler™ service at a single concentration of 10 µM. Protein kinases are tested in a radiometric assay format and the raw data are measured by scintillation counting (in cpm).

Eurofins Radioligand binding assays

ML366 was screened against 67 targents with the LeadProfileScreen panel from Eurofins. Biochemical assay results are presented as the percent inhibition of specific binding or activity throughout the report. A significant response is considered anything greater than 50%. None of the assay results showed significant binding by ML366. Methods employed in this study have been adapted from the scientific literature to maximize reliability and reproducibility. Reference standards were run as an integral part of each assay to ensure the validity of the results obtained. Full assay conditions and references are available on the Eurofins website (http://www.eurofinspanlabs.com).