Isoform Selectivities of Novel 4-Hydroxycoumarin Imines as Inhibitors of Myosin II

Abstract

Muscle myosin inhibition could be used to treat many medical conditions involving hypercontractile states, including muscle spasticity, chronic musculoskeletal pain, and hypertrophic cardiomyopathy. A series of 13 advanced analogs of 3-(N-butylethanimidoyl)ethyl)-4-hydroxy-2H-chromen-2-one (BHC) were synthesized to explore extended imine nitrogen side chains and compare aldimines vs. ketimines. None of the new analogs inhibit nonmuscle myosin in a cytokinesis assay. ATPase structure-activity relationships reveal that selectivity for cardiac vs. skeletal myosin can be tuned with subtle structural changes. None of the compounds inhibited smooth muscle myosin II. Docking the compounds to homology models of cardiac and skeletal myosin II gave rationales for the effects of side arm length on inhibition selectivity and for cardiac vs. skeletal myosin. Properties including solubility, stability and toxicity, suggest that certain BHC analogs may be useful as candidates for preclinical studies or as lead compounds for advanced candidates for drugs with cardiac or skeletal muscle myosin selectivity.

Graphical Abstract

INTRODUCTION

Myosins are a diverse class of motor proteins that produce force through the hydrolysis of adenosine triphosphate (ATP). They are biologically ubiquitous, highly conserved in eukaryotes and serve many load-sensing, motility, force-producing and structural functions in the cytoskeleton and muscle fibers.¹ Isoforms in skeletal, cardiac, and smooth muscle belong to a superfamily known as class II myosins. Three similar but distinct nonmuscle isoforms are included in the myosin II family: nonmuscle A (NM-IIA), B (NM-IIB) and C (NM-IIC).² Class II myosins are comprised of two homotrimers, each of which contains a heavy chain and two types of light chains (RLC and ELC: regulatory and essential). The heavy chain forms the N-terminal globular head domain that contains the actin and ATP-binding sites, followed by a lever arm domain that binds the two types of light chains. Two homotrimers stably interact through the C-terminal portions of the heavy chains to form a long coiled-coil domain. The coiled-coils interact electrostatically to form a filament, which is the physiological assembly that cyclically interacts with actin in the cell.

When myosin is activated through mechanical or chemical signaling, it cycles between actin-attached and detached states, as well as catalytic transition states that have different affinities for actin. When ATP binds acto-myosin, it greatly weakens the affinity of myosin for actin and causes detachment. Conformational changes attending the hydrolysis of ATP weaken the binding of myosin to actin. The release of inorganic phosphate (Pi) from the catalytic domain leads to a strongly bound adenine diphosphate (ADP) state, completion of the force-producing power stroke, and movement of the myosin filament relative to the actin filament.³

We are interested to develop isoform-specific class II myosin small molecule inhibitors as research tools and/or clinical candidates. Inhibition of muscle myosin could be used to treat many conditions that are known to be involved in hypercontractile states. For example, inhibition of skeletal muscle myosin could have clinical utility in the treatment of muscle spasticity following stroke⁴ and chronic musculoskeletal pain due to muscle spasm.⁵ The cardiac myosin inhibitor, mavacamten, has been tested on humans with encouraging results and is currently in clinical trials for hypertrophic cardiomyopathy,^6–8 while the cardiac myosin activator, omecamtiv mecarbil, has shown favorable outcomes over placebo for systolic heart failure.^9,10 Fasudil, an inhibitor of the upstream myosin effector Rho kinase,¹¹ has been approved for cerebral vasospasm and is a potent vasodilator, acting on myosin II in vascular smooth muscle cells.¹²

The myosin II inhibitor, blebbistatin, is frequently used in research, though its applications have been limited by lack of myosin II isoform selectivity as well as photoinstability, phototoxicity, and low solubility.^13–16 Blebbistatin and most of its characterized analogs inhibit nonmuscle myosin with moderate potency,¹³ preventing action potentials (APs) at neuromuscular junctions (NMJ) by blocking NM-IIB-mediated neurotransmitter release.¹⁷ Recently, a blebbistatin analog was developed that is selective for the skeletal isoform that does not appear to inhibit NM-IIA.¹⁸ A selective inhibitor can be used to investigate the structural and strain-sensing role of nonmuscle myosin affecting muscle myosin, as both isoforms have functional roles in smooth muscle cells.^{19– 23} For example, independent of smooth muscle myosin, NM-IIA and NM-IIB mediate aortic smooth muscle cell stiffness through separate mechanisms and at different locations within the cell.²⁴

We previously synthesized a series of 4-hydroxycoumarin myosin II inhibitors with aliphatic and aryl substituents and characterized their activity toward cardiac and skeletal myosin.²⁵ The lead compound, 3-(N-butylethanimidoyl)-4-hydroxy-2H-chromen-2-one (BHC), blocks skeletal muscle contraction without affecting neurotransmitter release and resulting APs at the NMJ,²⁶ a process dependent upon nonmuscle myosin activity. Several lines of evidence suggested that BHC and its analogs bind to the same site on myosin as blebbistatin and have similar mechanisms of action. Structure-activity relationships (SARs) and modeling identified several regions of the BHC scaffold that tune the inhibitory potencies and therefore confer selectivity for skeletal versus cardiac muscle myosins.²⁵

Reported here is a new series 4-hydroxycoumarin analogs, which we have synthesized to better explore and define SARs related to muscle myosin II inhibition, cytokinesis inhibition, and cytotoxicity. We have characterized the activities of these new analogs and our previously reported compounds²⁵ as inhibitors for chicken gizzard smooth muscle myosin, bovine ventricular cardiac myosin, and fast rabbit skeletal myosin. We have also investigated nonmuscle myosin II activity by cytokinesis inhibition, a NM-IIB-dependent process,²⁷ and determined cytotoxicity by a cell viability assay. Based upon these findings, we present several hydroxycoumarin imine analogs as promising small molecules for medicinal applications, since they are selective for cardiac and skeletal muscle myosin over smooth muscle and nonmuscle myosin II, display low cytotoxicity, and have favorable physical and chemical properties, such as solubility and hydrolytic stability. Our SAR studies have also uncovered subtle structural factors that alter selectivity between cardiac vs. skeletal myosin inhibition.

RESULTS AND DISCUSSION

Synthesis.

To expand our understanding of the SARs of BHC analogs with respect to myosin II inhibition, thirteen new derivatives were synthesized. We previously reported that BHC analogs possessing longer linear aliphatic or arylmethyl imino groups showed better myosin inhibition activity than those with short alkyl side chains.²⁵ This trend is exemplified by molecules 1 and 2 (Figure 1), which can be considered lead compounds for further analogs. However, as can be inferred by the structures, the addition of elongated hydrocarbon chains or ring systems is also accompanied by decreased solubility. To further improve the drug-like properties of the BHC analogs and study the effects of adding polar groups to the imino side chain on potency, we synthesized a series of analogs with heteroatom-containing linear chains or heteroatom-substituted aromatic moieties.

Figure 1.

Examples of relatively potent BHC analogs²⁵ with long aliphatic (1; JB052) or arylmethyl (2; JB002) imino substituents (R groups; see Scheme 1).

Based on previously reported synthetic routes²⁵ a series of BHC analogs with oxygen-containing imino side chains were prepared, as shown in Scheme 1. Accordingly, intermediate 3-acetyl-4-hydroxycoumarin (4) was prepared by acetylation of 4-hydroxycoumarin (3).^25,28 A series of six Schiff base BHC analogs (5a-d and 6a,b) were obtained by condensation of ketone 4 with commercially available primary amines.²⁹ In previous studies we also found that elongation of the acetyl group of 4 to propanoyl or butanoyl decreased the myosin inhibition potency of the corresponding ketimine.²⁵ To determine if reducing the size of the acyl moiety might increase potency, aldehyde 7 was prepared by Vilsmeier-Haack reaction of 3 with phosphorus oxychloride and dimethylformamide,³⁰ then it was condensed with primary amines to produce aldimine analogs 8a and 8b.

Scheme 1.

In silico ligand docking studies on skeletal myosin homology models based on the X-ray crystal structure of the blebbistatin complex of Dictyostelium discoideum myosin II (PDB 1yv3)¹³ suggested that elongation of the imino chain could increase interactions by occupying a channel in the myosin binding site. To test the calculations, several elongated analogs were designed with ether linkages to reduce hydrophobicity. Analogs with elongated imino chains were synthesized by treating primary alcohols 9a-c with thionyl chloride to give the corresponding alkyl chlorides 10a-c (Scheme 2).^31,32 The chlorinated products then underwent nucleophilic substitution with 3-(hydroxypropyl)amine (11), which was first treated with sodium hydride to deprotonate the hydroxyl group to make the oxygen more nucleophilic than the amine nitrogen.³³ The resulting products 12a-c were verified as primary amines by proton NMR spectroscopy, showing that the resonances of the terminal methylene hydrogens of the alkyl chloride reactants were shifted further downfield by O-substitution (~4.4 ppm) than expected for N-substitution (~3.8 ppm).³⁴ Condensation of 12a-c with 3-acetyl-4-hydroxycoumarin (4) gave desired elongated imino chain analogs 13a-c.

Scheme 2.

In addition to the ether-linked elongated imino chains, in silico studies also suggested that elongated chains incorporating tertiary amino moieties could show myosin inhibitory activity. Secondary amine 15 was synthesized by reductive amination of 3-methoxybenzaldehyde (14) by reaction with butylamine followed by sodium borohydride, as shown in Scheme 3. Butylamine was chosen as the amino component because the computational results suggested that four carbons would give an alkyl chain of ideal length. The secondary amino group was then alkylated by reaction with N-(3-bromopropyl)phthalimide (16) to produce tertiary amine 17. The terminal amino group was deprotected with hydrazine,³⁵ and the resulting primary amine 18 was condensed with 3-acetyl-4-hydroxycoumarin (4) to form ketimine 19 bearing an elongated imino group containing a tertiary amine.

Scheme 3.

As Schiff bases and enamines are known to undergo metabolism through hydrolysis in biological systems,³⁶ a new, hydrolytically stable structural motif was also investigated. A chalcone moiety was used to simulate the benzylimino side chain pharmacophore component of BHC analog 2. Accordingly, ketone 4 underwent aldol condensation with vanillin to produce the chalcone 20, as is also shown in Scheme 3.

Structural characterization.

Schiff bases derived from 3-acetyl-4-hydroxycoumarin (4) and 3-formyl-4-hydroxycoumarin (7) can undergo tautomerization between the imine and enamine tautomers, and the enamine tautomer can exist either as the E or Z stereoisomer (Figure 2). While both tautomers of the ketimines derived from 4 (Figure 2A) are stabilized by intramolecular hydrogen bonding, there is compelling evidence from X-ray crystallography,^37,38 in addition to ¹H and ¹³C NMR spectroscopic analysis³⁹ and density functional theory energy calculations,^40,41 that the ketimines exist primarily or exclusively as the E-enamine isomer. The ¹³C NMR spectra calculations,^40,41 that the ketimines exist primarily or exclusively as the E-enamine isomer. The ¹³C NMR spectra of ketimines 5a-d, 6a-b, 13a-c, and 19 lack a ketone carbonyl resonance at ~δ 180; it is presumed that they exist primarily as E-enamines, as observed for all previously investigated condensation products between 4 and primary amines. Aldimines derived from 7 also favor the enamine tautomer, but in this case the E and Z diastereomers exist in equilibrium (Figure 2B), splitting the ¹H NMR signals for the amino hydrogen (~δ 11.6 and 10.4 ppm) and vinylic hydrogen (~δ 8.5 and 8.4 ppm).⁴² The NMR spectra of 8a and 8b closely match those of similar reported structures. The E/Z equilibrium ratios calculated from proton NMR integrals are 60:40 and 70:30 for 8a and 8b, respectively. To maintain consistency with nomenclature in previous literature, compounds 5a-d, 6a-b, 8a-b, 13a-c, and 19 are named as their imine tautomers in the Experimental Section.

Figure 2.

Structures of imine and enamine tautomers of A) ketimines derived from 3-acetyl-4-hydroxycoumarin, and B) aldimines derived from 3-formyl-4-hydroxycoumarin.

Structure-activity relationships.

The half-maximal concentration (IC₅₀) for reduction of steady-state actin-activated Mg²⁺ ATPase activity for skeletal, cardiac, and smooth muscle myosin was determined to assess relative potency and selectivity (see ref. 25 for background). Since this ATPase activity represents the rate-limiting step in the ATPase cycle of myosin II, BHC analogs most likely stabilize the actin-unbound myosin−ADP−Pi state, as has been previously shown for blebbistatin. The data analysis and ATPase assay protocol, which directly measure P_i formation from enzymatic hydrolysis of ATP to ADP, were performed as previously described.²⁵ Importantly, all the myosins used in these studies were full-length native proteins purified from tissue, not soluble subfragments. Full length myosins contain two head domains and are able to form filaments in vitro. In this study, we assayed the 13 new analogs listed in Table 1 and the 25 coumarins we previously reported²⁵ for potency to inhibit cardiac, skeletal, and smooth muscle myosin ATPase activity. The only differences between these compounds are the side chains attached to the imine nitrogen and the group (H or methyl) attached to the imine carbon: R or R’, respectively, in the general structure shown at the top of Table 1.

Table 1.

Summary of IC₅₀ values, selectivity, cytotoxicity, and lipophilicity data for the new analogs.

Analog	Structure	Potency to inhibit muscle myosins; IC50 (μM)1			Selectivity	Cytotoxicity		Lipo- philicity
General structure		Cardiac (x)	Skeletal (y)	Smooth	x/y	CC502 (μM)	CC103 (μM)	ClogP4
5a (JB008)		52 ± 55	90 ± 30	≥ 100	0.6	> 200	286 ± 28	1.55, 1.17
5b (JB057)		24 ± 6	19 ± 2	≥ 100	1.3	> 1,000	184 ± 85	1.91, 1.53
5c (JB059)		38 ± 9	11 ± 1	≥ 100	3.5	> 1,000	339 ± 91	1.61, 1.23
5d (JB058)		8.1 ± 1.9	2.5 ± 0.2	≥ 100	3.2	980 ± 240	53 ± 27	1.97, 1.59
6a (JB061)		4.4 ± 0.5	9.1 ± 1.4	≥ 100	0.5	> 200	39 ± 38	3.17, 2.78
6b (JB062)		5.4 ± 0.7	1.6 ± 0.2	≥ 100	3.4	> 200	36 ± 35	3.17, 2.78
8a (JB100)		≥ 100	3.3 ± 2.0	≥ 100	≥ 30	> 1,000	135 ± 28	1.48, 1.51
8b (JB104)		≥ 100	≥ 100	≥ 100	–	–	–	2.84, 2.87
13a (JB102)		≥ 100	≥ 100	≥ 100	–	–	–	1.52, 1.13
13b (JB080)		≥ 100	50 ± 5	≥ 100	≥ 2.0	> 200	46 ± 9	3.18, 2.80
13c (JB092)		≥ 100	29 ± 3	≥ 100	≥ 3.4	> 200	110 ± 54	2.96, 2.58
19 (JB099)		≥ 100	≥ 100	≥ 100	–	–	–	4.57, 4.19
20 (JB077)		≥ 100	≥ 100	≥ 100	–	–	–	3.26

¹Inhibitor concentration at half-maximal (50%) decrease in ATPase activity. For details see ref. 25.

²Inhibitor concentration at half-maximal (50%) decrease in COS-7 cell viability. For details see Experimental Section.

³Inhibitor concentration at 90% decrease in COS-7 cell viability. For details see Experimental Section.

⁴ClogP values were calculated as described in the Experimental Section. In cases where imine/enamine tautomerism is possible, ClogP values of the imine and enamine forms are listed first and second, respectively. Clog P values are the same for E and Z enamines. For amine 19, ClogP was calculated for the protonated form.

⁵Uncertainty is the standard error of the mean from the fit to three independent titrations from 0.01 to 30 μM for each analog.

None of the 48 hydroxycoumarin analogs inhibit phosphorylated smooth muscle myosin.

While many inhibited skeletal and/or cardiac myosin, all analogs had an IC₅₀ ≥ 100 μM for phosphorylated chicken gizzard smooth muscle myosin. Unlike skeletal or cardiac myosin, the smooth muscle myosin isoform requires regulatory light chain phosphorylation for activated ATPase activity. There was evidence for weak inhibition for several compounds at 100–200 μM, but effects of the analogs on protein aggregation and/or analog solubility prevented calculation of reliable IC₅₀ values from fits to the binding curves. Therefore, in general, the hydroxycoumarin imine scaffold is highly selective for skeletal and cardiac muscle myosin over smooth muscle myosin. Until analogs are found that show significant potency for smooth muscle myosin, we are unable to speculate further about SARs vis-a-vis this isoform.

For the ketimines (R” = methyl), the chain length of the imine nitrogen R group modulates potency for both skeletal and cardiac myosins (Table 1).

The structures of the ketimines fall nicely into three classes according to the type of R side chain: aliphatic, ether, and aromatic. Table 2 groups the 9 new ketimines and 10 of the previously reported hydroxycoumarin ketimines according to these three classes, shows the side chain structures, and gives the number of non-hydrogen atoms (NHAs) in the longest linear substructure of each side chain. For both skeletal and cardiac myosin, the linear non-hydrogen atom (NHA) chain length of the R group (n) tunes potency (Figure 3). Interestingly, the pattern for IC₅₀ values of aliphatic, ether, and aromatic R groups versus n is very similar for skeletal and cardiac myosin. The highest potency is found when n is between 5–7, suggesting that the R group must be long enough to maximize protein-ligand interaction surface area but short enough to fit within the binding pocket. At n > 7, another trend appears. Although there are no analogs with n between 7 and 10 NHA, the longest aromatic analogs (n = 10 –11) retain measurable activity for skeletal myosin but not cardiac myosin. This is an interesting aspect of the hydroxycoumarin analog SAR that could be explored further to develop highly selective skeletal myosin inhibitors. Table 2 also identifies the selectivities (cardiac vs. skeletal) of each inhibitor and identifies those with IC₅₀ values equal to or less than 10 μM. As seen in Figure 3, potency for the aliphatic series increases consistently from n = 0 – 6 and skeletal is consistently favored by a factor of ca. 2–3. In the ether series, n = 4 gives weak potency and selectivity for cardiac, then both potencies peak at n = 6 with skeletal selectivity. The aromatic series is most interesting in that potencies are high for n = 5–7 and selectivities depend on substituent position on the benzene ring of benzyl analog JB002 (2). The meta methyl and methoxy analogs JB069 and JB061 (6a) are cardiac-selective, while para analogs JB060 and JB062 (6b) are skeletal-selective. These studies reveal 2/6b and 6a as potential lead compounds for developing skeletal-selective and cardiac-selective drugs, respectively.

Table 2.

Summary of structures of new and previously reported 4-hydroxyquinoline ketimines, longest linear NHA chain length of their imine R groups (n), relative selectivities for inhibition of cardiac vs. skeletal myosin, and identification of analogs with IC₅₀ values ≤ 10 μM. See Figure 3 for plot of IC₅₀ values vs. n.

Type	Compound	Structure (R)	n 1	Selectivity	IC50 value ≤ 10 μM ?
Aliphatic	JB0322	H	0	–	No
	JB0102	methyl	1	–	No
	JB0092	ethyl	2	skeletal	No
	JB0112	propyl	3	skeletal	No
	BHC2	butyl	4	skeletal	Yes
	JB005	pentyl	5	skeletal	Yes
	JB0522 (1)	hexyl	6	skeletal	Yes
Ether	JB008 (5a)	CH2CH2OMe	4	cardiac	No
	JB057 (5b)	CH2CH2OEt	5	skeletal	No
	JB059 (5c)	CH2CH2CH2OMe	5	skeletal	Yes
	JB058 (5d)	CH2CH2CH2OEt	6	skeletal	No
	JB102 (13a)	CH2CH2CH2O(CH2CH2O)2Me	11	–	No
Aromatic	JB002 (2)	CH2ph	5	skeletal	Yes
	JB0692	CH2(m-MePh)	5	cardiac	Yes
	JB0602	CH2(p-MePh)	6	skeletal	Yes
	JB061 (6a)	CH2(m-MeOPh)	6	cardiac	Yes
	JB062 (6b)	CH2(p-MeOPh)	7	skeletal	Yes
	JB080 (13b)	CH2CH2CH2OCH2(p-MeOPh)	11	skeletal	No
	JB092 (13c)	CH2CH2CH2OCH2Ar3	10	skeletal	No

¹Number of NHAs in longest linear chain attached to imine nitrogen R group.

²Ref. 25.

³See Table 1 for structure.

Figure 3.

Skeletal (A) and cardiac (B) myosin inhibition potencies (IC₅₀) vs. linear R group chain length in non-hydrogen atoms for selected analogs with aliphatic, ether, or aromatic R groups. Aliphatic analogs plotted are JB032, JB010, JB009, JB011, BHC, JB005, and 1. Ether analogs plotted are 5a, 5b, and 5d. Aromatic analogs plotted are 2, JB060, 6b, 13b and 13c.

Interestingly, the effect of R′ can depend upon R, and this may be a promising approach to tuning selectivity. This can be seen when comparing 5d and 8a, both of which contain the same aliphatic ether R group. Analog 5d with R′ = methyl is a good inhibitor for both skeletal and cardiac myosin, however 8a with R′ = H retains potency only for skeletal myosin (IC₅₀ = 2.5 ± 0.2 μM) and is not potent toward cardiac myosin (IC₅₀ ≥ 100 μM). Analog 8a is so far the most highly selective inhibitor for skeletal myosin (Table 1, factor of ~30) we have characterized. The oxygen of the aliphatic ether group may be important to the structural basis of skeletal selectivity. Note that analogs with optimum R group length containing an aliphatic ether and R′ = methyl tend to be selective for skeletal myosin, such as 5b, 5c, 5d, 13b, and 13c. In principle, replacing an electron donating methyl group with hydrogen near an enamine should increase its NH hydrogen-bond donor ability, so the R′ = H in 8a may promote hydrogen bonding to the aliphatic ether oxygen lone pair electrons. The effect of a hydrogen in the R′ position warrants further investigation with different functional groups at the R position.

Computational Studies.

To investigate the possibility of ether side chain hydrogen bonding and to explore determinants of skeletal myosin selectivity, 13c was docked to a known structure of skeletal muscle myosin and homology models of cardiac and smooth muscle myosin. Compound 13c has a long R group with n = 10 NHA including an ether moiety and is relatively selective for skeletal myosin, with an IC₅₀ ≥ 100 μM in cardiac and smooth muscle myosin and displaying no cytokinesis inhibition at 40 μM (see below). Although 13c has only moderate potency for skeletal myosin (29 ± 3 μM), docking may give insight into the trend of longer R groups and presence of the ether group in retaining activity in skeletal but not in cardiac muscle myosin.

The binding pocket for the hydroxycoumarins, which is shared by blebbistatin,²⁵ is in a highly dynamic cleft between the upper and lower 50 kDa domains with conformational changes occurring during actin binding and release.³ The binding pocket is highly conserved between myosin isoforms, except for two residues. Figure 4A shows the energy minimized binding pose for 13c (yellow) along with non-conserved residues surrounding the coumarin bicyclic moiety, which includes phenylalanine F654 in skeletal and histidine H651 in cardiac myosin. All three human non-muscle isoforms (NM-IIA, NM-IIB, and NM-IIC) and in smooth muscle myosin contain tyrosine Y663. This raises the possibility that tyrosine and histidine side chains at this position stabilize interactions that diminish potency, whereas phenylalanine does not. The hydroxyl group of tyrosine may be changing the shape of the pocket because it can form hydrogen bonds with other residues or solvent, leading to steric constraints. The magnitude of this steric effect could be less with the smaller histidine side chain.

Figure 4.

Docking results for the Z-enamine tautomer of 13c in its highest binding pose (yellow) in skeletal muscle myosin (green; PDB: 6YSY). Also highlighted are the differing residues derived from cardiac and smooth isoform homology models in orange and magenta, respectively. Non-conserved pocket residues for the different isoforms are highlighted in licorice representation. A) Aspect emphasizes the non-conserved residues surrounding the coumarin bicyclic moiety. B) Aspect emphasizes the non-conserved residues surrounding the hydroxycoumarin R group. Residue identity and protein sequence number are indicated.

The enamine NH can act as a hydrogen bond donor to the carbonyl oxygen at the coumarin 4 position and to the ether oxygen in the R group of 13c (Figure 4A). This could lead to an intramolecular hydrogen bonded six-membered ring that might stabilize the ligand in an effective conformation with respect to inhibition of myosin ATPase. The polar histidine in cardiac myosin could destabilize intramolecular hydrogen bonding, leading to lower potency relative the skeletal myosin, as seen with most of the ether-containing analogs. Three carbon atoms separating the ether oxygen from the enamine nitrogen are required to form this hydrogen-bonded six-membered ring. Analogs with an ether group in the side chain at this position (5c, 5d, 8a, 13b, and 13c) all display selectivity for skeletal over cardiac myosin. The addition of other ether groups along the side chain may lead to competition for enamine hydrogen bonding and therefore weaken the intramolecular bonding network and decreasing potency, as seen with 13a. The other non-conserved residue in the binding pocket is shown in Figure 4B. Leucine (L476) is unique to skeletal myosin, while the phenylalanine (F473) is conserved in cardiac, smooth, and NM-IIB. The presence of the bulkier phenylalanine may shorten and change the shape of the pocket in this region specifically affecting potency for analogs with longer R groups. This steric hindrance may also explain the differences between cardiac and skeletal inhibition with analogs containing the longer R groups. The complex interplay between intramolecular hydrogen bonding and imine/enamine tautomerization make the hydroxycoumarin imines a unique and versatile scaffold for myosin II inhibition. Small changes to the inhibitor structure can have drastic effects on potency and selectivity that could in part be due to shifts in the conformational equilibrium in the local environment of the binding pocket. The position of the ether oxygen along the R group side chain appears to be an important modification of hydroxycoumarins for tuning potency and selectivity.

Cytokinesis assay for inhibition of NM-IIB.

Class II myosins include non-muscle as well as muscle myosin and there is a high degree of sequence and structural homology within this class.^43,44 NM-II is important for many vital cellular mechanisms,^{17,19,23,24,45} so inhibiting this isoform could limit medicinal potential by affecting many necessary biological processes. We assayed all 48 analogs including the 13 new compounds reported here and those in Ref. 25 for NM-IIB inhibition using cytokinesis in Cos-7 cells as an indirect measure of NM-IIB activity. NM-IIB is an integral and required component of the contractile ring that forms the cleavagefurrow and splits the cytoplasm into two daughter cells during cytokinesis.^27,46 When cytokinesis is inhibited, the nuclei can replicate, but the cleavage-furrow cannot form and split the cytoplasm. Therefore, inhibition of NMIIB is sufficient but not necessary to inhibit cytokinesis. Blebbistatin is a cytokinesis inhibitor via NM-IIB inhibition, but this process can be inhibited through other signaling or structural targets such as the actin modulator cytochalasin, the aurora B kinase inhibitor hesperidin, or the calmodulin antagonist W7.²⁷

Figure 5 summarizes the ability of all forty-eight analogs synthesized to date to inhibit cytokinesis in Cos-7 cells. Blebbistatin and para-aminoblebbistatin were used as positive controls. Both compounds inhibited cytokinesis as evidenced by relatively high nuclei to cell ratios (N/C) compared to the two negative controls, media alone and 1 % DMSO in media. None of the analogs, including those we previously reported,²⁵ showed N/C ratios significantly above baseline values comparable to the negative controls. This lack of effect could be explained if the analogs were not membrane permeant. However, this is unlikely because we have shown that selected analogs were able to inhibit skeletal muscle contraction in physiologic studies with intact muscle tissue.^25,26 In addition we have calculated octanol/water partition coefficients (ClogP values) for each of the new compounds (Table 1) and find that they lie in the ideal range of for bioavailability (0–4),⁴⁷ if controlled primarily by gut cell membrane permeability. In summary, our data suggest that the hydroxycoumarin analogs do not significantly inhibit NM-IIB under these conditions.

Figure 5.

Ability of analogs to inhibit cytokinesis in Cos-7 cells (40 μM, 24 h, 37 °C). Nuclei to cell ratio (N/C) is shown for two negative controls, media and 1% DMSO in media, as well as two positive controls, para-aminoblebbistatin (paB) and blebbistatin (blebb). Results are shown for all the analogs reported here as well as those from Ref. 25. The average standard deviation was 4% of the N/C (error bars not shown).

Toxicity, solubility and chemical stability.

These are some of the key “drug-like properties” that often supersede potency for practical success of a drug. Cytotoxicity was assessed for the analogs with detectable myosin inhibitory activity in Cos-7 mammalian fibroblast cells based on measuring cell viability using an MTT assay (Table 1). All of the hydroxycoumarins were found to have such low toxicity up to the maximum concentrations tested (200–1,000 μM) that the concentration killing 50% of the cells (CC₅₀ value) could not be calculated for all except 5d (ca. 1 mM). In order to compare compound toxicities, CC₁₀ values for only 10% cell mortality were calculated (Table 1). Only 5d, 6a, 6b, and 13b showed CC₁₀ values in the two-digit micromolar range. Analog 6b and several other hydroxycoumarin imines with structural similarities to the analogs we characterized in our study were previously found to be cytotoxic in a normal human cell line, BEAS-2B, and two human cancer cell lines, MCF-7 and A549, in the low μM range.⁴¹ Compound 6b was reported to be among the most cytotoxic of those compounds. Hydroxycoumarin imines that are cytotoxic to human cancer cells but not normal cells like those characterized in our study warrant further investigation as potential chemotherapeutic agents. The cytotoxicity of 6b does not appear to be mediated by inhibition of NM-IIB as this analog did not inhibit cytokinesis at a concentration of 40 μM (Figure 5), well above the cytotoxicity IC₅₀ in three different human cell lines.⁴¹ The low cytotoxicity of analogs that are potent and selective inhibitors of skeletal muscle myosin, such as 8a, is encouraging when evaluating the medicinal potential of these analogs.

In our previous study,²⁵ we determined the kinetic solubilities of three analogs having interesting potencies and selectivities for muscle myosins. The results are shown in Table 3 for comparison with the newly-determined solubilities of analogs 5d, 13b, and 13c. All of these measurements were made by addition of DMSO stock solutions to pH 7.4 PBS, detecting turbidity by nephelometry, with final DMSO concentrations of 2.5%. The least soluble compound was found to be N-benzyl analog JB002 in the previous series,²⁵ but 100 μM is more than 10-fold higher than the skeletal myosin potency of this compound. Increasing the polarity of the aromatic ring to pyridine in JB031 greatly increases solubility, as does incorporation of one ether oxygen in the 6 NHA side chain of 5d. Compounds 13b and 13c with extended side chains containing oxygen-substituted aromatic rings have lower kinetic solubilities, but they are still in a practical range for a drug.

Table 3.

Kinetic solubilities of selected hydroxycoumarin muscle myosin inhibitors.

Compound	Structure	Kinetic solubility (μm)1
BHC2		607 ± 2
2 JB0022		103 ± 2
JB0312		> 400
5d (JB058)		> 1250
13b (JB080)		103 ± 3
13c (JB092)		122 ± 2

¹Determined by nephelometry in PBS (pH 7.4) containing 2.5% DMSO.

²Ref. 25.

Given the potential for hydrolysis of imine and enamine functional groups in aqueous media, we previously determined the half-life for hydrolysis of JB031 in 99:1 (v/v) PBS/methanol to be 3.5 ± 0.5 days at room temperature.²⁵ Aldimines, such as new analogs 8a and 8b, may hydrolyze more rapidly than ketimines. We were particularly concerned that N-benzyl aldimine 8b had no detectable potency for muscle myosin inhibition, while the corresponding ketimine 2 was found to have IC₅₀ values of 8 and 1 μM for cardiac and skeletal myosin, respectively! Therefore, we carried out parallel hydrolysis studies of 8b and 2 in the buffer used for the ATPase assay. Ketimine 2 was found to have a half-life of > 6 days and aldimine 8b had a half-life of 2.5 ± 0.5 days at room temperature.²⁵ Aldimines, such as new analogs 8a and 8b, may hydrolyze more rapidly than ketimines. We were particularly concerned that N-benzyl aldimine 8b had no detectable potency for muscle myosin inhibition, while the corresponding ketimine 2 was found to have IC₅₀ values of 8 and 1 μM for cardiac and skeletal myosin, respectively! Therefore, we carried out parallel hydrolysis studies of 8b and 2 in the buffer used for the ATPase assay. Ketimine 2 was found to have a half-life of > 6 days and aldimine 8b had a half-life of 2.5 ± 0.5 days at room temperature. Since the duration of exposure to aqueous conditions for the ATPase assays was less than 2 h, it is unlikely that appreciable hydrolysis occurred in this time. This suggests that the differences in potencies of 8b and 2 are due to the structures of the compounds, not their rates of hydrolysis.

SUMMARY AND CONCLUSIONS

We have synthesized and structurally characterized 13 new analogs of the lead compound, 3-(N-butylethanimidoyl)-4-hydroxy-2H-chromen-2-one (BHC), most of which contain extended and functionalized side chains attached to the imine nitrogen atom. BHC analogs all apparently bind to a site on myosin II that is distinct from the ATPase catalytic site and inhibit release of inorganic phosphate in the power stroke of this motor molecule. We have shown by means of a cytokinesis assay that none of the compounds inhibit function of nonmuscle myosin and we also explored SARs for inhibition of cardiac, smooth and skeletal muscle myosin II by means of an ATPase assay. While none of the analogs inhibited smooth muscle myosin, the studies have uncovered subtle structural factors that affect cardiac vs. skeletal myosin inhibition selectivity, including structures of ether-containing side chains and positions of substituents on aromatic rings in side chains. A new variation in structure was also explored involving removal of the ketimine methyl group to produce aldimines with altered activity and selectivity. Computational studies involving docking the analogs to homology models of the structures of cardiac and skeletal myosin II rationalize the observed dependence of potency on side chain length and produce hypotheses for cardiac vs. skeletal selectivity. In particular, differences in amino acid residues in the binding sites of isoforms are proposed to affect hydrogen bonding networks for analogs with ethercontaining side chains. These studies may lead to useful strategies for fine tuning isoform selectivity in hydroxycoumarin inhibitors of myosin II. Cytotoxicities determined by an MTT assay showed that compounds with the hydroxycoumarin scaffold and with the substituents explored have low toxicities in the concentration range for myosin inhibition. Studies of solubilities and hydrolytic stabilities of selected structures showed that these properties are adequate for drug applications.

Clinical Relevance.

Inhibition of muscle myosin could be used to treat many conditions that involve hypercontractile states, such as muscle spasticity following stroke and chronic musculoskeletal pain caused by muscle spasm. Clinical trials on cardiac myosin inhibitors have focused on hypertrophic cardiomyopathy and systolic heart failure. An important potential application of inhibitors myosin II in vascular smooth muscle cells is vasodilation for treatment of cerebral vasospasm. For clinical applications, small-molecule inhibitors should be specific for skeletal, cardiac, or smooth muscle myosin and not affect nonmuscle myosins. None of the analogs we examined have any detectible activity toward nonmuscle IIB or smooth muscle myosin but some show promising selectivity between cardiac and skeletal myosin, particularly JB002 (2), JB061 (6a), and JB062 (6b). Moreover, these compounds generally exhibit good drug-like properties, including solubility, chemical stability, and very low cytotoxicity. They might be good candidates for initial preclinical studies or they could serve as lead compounds for development of advanced candidates.

EXPERIMENTAL SECTION

Chemistry: General Methods.

Reagents and solvents purchased from Acros Organics, Sigma-Aldrich, or Fisher Scientific were of ACS reagent grade or better and were used without purification. Anhydrous tetrahydrofuran obtained by distillation from sodium/benzophenone under nitrogen. Anhydrous acetonitrile obtained by purification with an LC Technology Solutions SPBT-1 solvent purification system. All reactions were performed under air unless otherwise noted. “Overnight” periods are ca. 16 h. “Column chromatography” was performed with silica gel (40–63 μm particle size, 230–400 mesh). “Flash column chromatography” was performed on a Yamazen AKROS flash chromatograph using prepackaged neutral silica (40 μm particle, 60 Å pore size). Melting points were measured on a Thomas-Hoover capillary melting point apparatus and are uncorrected. ¹H NMR (400 MHz or 500 MHz) and ¹³C NMR (101 MHz or 126 MHz) spectra were acquired on a Varian 400 or Varian Unity +500 spectrometer. All chemical shifts (δ) are reported in ppm units relative to solvent resonance, as follows: ¹H, CDCl₃/TMS = 0.00, DMSO-d₆/TMS = 0.00; ¹³C, CDCl₃ = 77.16, DMSO-d₆ = 39.52. Infrared spectra (IR) were recorded neat on a Nicolet 6700 FTIR spectrometer. Low resolution mass spectra (MS) were acquired on a Waters Micromass ZQ electrospray ionization quadrupole mass spectrometer with positive ion detection (capillary voltage = 3.5 kV). UV-visible spectra (UV/Vis) were recorded on a Shimadzu UV-2550 UV-Vis spectrophotometer; molar extinction coefficients (ε, M⁻¹cm⁻¹) were determined by Beer-Lambert law plots of 4–5 data points in the absorbance range of 0.2–0.8 AU. Samples for elemental analysis were dried at 78 °C (0.1 mm) for 2 d and microanalysis was performed by NuMega Resonance Labs, Inc. The purities of all samples submitted for biological testing were greater than 95%, as shown by combustion microanalysis.

Synthesis.

Schiff bases 5a-d, 6a-b, 8a-b, 13a-c and 19 were synthesized from primary amines²⁹ and purified by one of the following three methods, unless stated otherwise. Method A. A 0.2–0.8 M solution of 3-acetyl-4-hydroxy-2H-chromen-2-one (4) and 1.1–2 eq. of an amine in the stated solvent was stirred and boiled under reflux 4–16 h, then cooled to room temperature. The resulting precipitate was collected by filtration and recrystallized from ethanol. Method B. A 0.2–0.8 M solution of 3-acetyl-4-hydroxy-2H-chromen-2-one (4) and 1.1–1.2 eq. of an amine in the stated solvent was stirred and boiled under reflux 4–16 h, then cooled to room temperature. The solution was concentrated to dryness by rotary evaporation and the product was purified by recrystallization from ethanol. Method C. A 0.2–0.8 M solution of 3-acetyl-4-hydroxy-2H-chromen-2-one (4) and 1.1–1.2 eq. of an amine in the stated solvent was stirred and boiled under reflux 4–16 h, then cooled to room temperature. The solution was concentrated to dryness by rotary evaporation and the product was purified by flash chromatography on silica, eluting with the described solvent(s).

4-Hydroxy-3-(N-(2-methoxyethyl)ethanimidoyl)-2H-chromen-2-one (5a, JB008).

Following method B, condensation of 0.49 g (2.4 mmol) of 4 with 2-methoxyethylamine in ethanol gave 0.28 g (36%) of 5a as white crystals, mp 96 °C. ¹H NMR (500 MHz, DMSO-d₆) δ 13.80 (s, 1 H, NH), 7.95 (dd, 7.8, 1.7 Hz, 1 H, H5), 7.63 (ddd, 8.0, 7.2, 1.7 Hz, 1 H, H7), 7.29 (td, 7.6, 1.1 Hz, 1 H, H6), 7.25 (dd, 8.2, 1.0 Hz, 1 H, H8), 3.79 (m, 2 H, NCH₂), 3.64 (t, 5.1 Hz, 2 H, CH₂O), 3.37 (s, 3 H, OCH₃), 2.67 (s, 3 H, NCCH₃). ¹³C NMR (126 MHz, DMSO-d₆) δ 179.6, 176.3, 161.9, 153.0, 133.9, 125.6, 123.5, 120.2, 116.1, 96.1, 69.5, 58.2, 43.7, 18.4. IR (cm⁻¹) 2864 (w), 2360 (w), 1693 (m), 1593 (m), 1571 (m), 1462 (s), 1371 (m), 1336 (s), 1304 (m), 1226 (w) 1199 (w), 1073 (m), 1025 (w), 981 (m), 906 (w), 758 (s). MS (ESI⁺) m/z 262.22 (MH⁺). Anal. Calcd. for C₁₄H₁₅NO₄: C, 64.36; H, 5.79; N, 5.36. Found: C, 64.14; H, 6.18; N, 5.54. UV (DMSO) λ_max, nm (ε): 324 (2.93 × 10⁴).

3-(N-(2-Ethoxyethyl)ethanimidoyl)-4-hydroxy-2H-chromen-2-one (5b, JB057).

Following method B, condensation of 0.46 g (2.2 mmol) of 4 with 2-ethoxyethylamine in ethanol gave 0.50 g (81%) of 5b as white crystals, mp 82–83 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 13.77 (s, 1 H, NH), 7.96 (ddd, 7.8, 1.7, 0.5 Hz, 1 H, H5), 7.63 (ddd, 8.2 Hz, 7.2, 1.7 Hz, 1 H, H7), 7.23–7.31 (m, 2 H, H6,8), 3.78 (m, 2 H, NCH₂), 3.67 (m, 2 H, NCH₂CH₂), 3.54 (m, 2 H, OCH₂CH₃), 2.67 (s, 3 H, NCCH₃), 1.17 (t, 7.0 Hz, 3 H, OCH₂CH₃). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.5, 176.3, 161.9, 153.0, 133.9, 125.7, 123.6, 120.2, 116.2, 96.1, 67.3, 65.7, 44.0, 18.5, 15.0. IR (cm⁻¹) 2864 (w), 1692 (m), 1570 (m), 1461 (m), 1371 (m), 1335 (m), 1302 (m), 1227 (w), 1156 (w), 1090 (m), 1025 (m), 981 (m), 903 (m), 752 (s), 668 (m). Anal. Calcd. for C₁₅H₁₇NO₄: C, 65.44; H, 6.22; N, 5.09. Found: C, 65.04; H, 6.47; N, 5.37.

4-Hydroxy-3-(N-(3-methoxypropyl)ethanimidoyl)-2H-chromen-2-one (5c, JB059).

Following method B, condensation of 0.50 g (2.5 mmol) of 4 with 3-methoxypropylamine in ethanol gave 0.32 g (47%) of 5c as white crystals, mp 78–79 °C. ¹H NMR (400 MHz, CDCl₃) δ 8.05 (dd, 7.9, 1.7 Hz, 1 H, H5), 7.53 (ddd, 8.2, 7.3, 1.7 Hz, 1 H, H7), 7.19–7.25 (m, 2 H, H6,8), 3.67 (td, 6.9, 5.5 Hz, 2 H, NCH₂), 3.52 (t, 5.7, 2 H, CH₂O), 3.38 (s, H, OCH₃), 2.75 (s, 3 H, NCCH₃), 2.02 (tt, 6.8, 5.7 Hz, 2 H, NCH₂CH₂). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.5, 176.0, 161.8, 153.0, 133.7, 125.6, 123.4, 120.1, 116.07, 96.0, 68.9, 58.0, 41.2, 28.3, 18.0. IR (cm⁻¹) 2947(w), 2864 (w), 1694 (m), 1598 (m), 1570 (m), 1461 (m), 1379 (w), 1228 (m), 1103 (m), 1052 (w), 947 (m), 886 (m), 752 (s). Anal. Calcd. for C₁₅H₁₇NO₄: C, 65.44; H, 6.22; N, 5.09. Found: C, 65.49; H, 6.10; N, 4.69.

3-(N-(3-Ethoxypropyl)ethanimidoyl)-4-hydroxy-2H-chromen-2-one (5d, JB058).

Following method B, condensation of 0.46 g (2.3 mmol) of 4 with 3-ethoxypropylamine in ethanol gave 0.47 g (74%) of 5d as white crystals, mp 62–63 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 13.73 (s, 1 H, NH), 7.95 (ddd, 7.8, 1.7, 0.5 Hz, 1 H, H5), 7.63 (ddd, 8.2, 7.2, 1.7 Hz, 1 H, H7), 7.22–7.32 (m, 2 H, H6,8), 3.67 (q, 6.4 Hz, 2 H, NCH₂), 3.39–3.54 (m, 4 H, CH₂OCH₂), 2.66 (s, 3 H, NCCH₃), 1.85–1.96 (m, 2 H, NCH₂CH₂), 1.12 (t, 7.0 Hz, 3 H, OCH₂CH₃). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.5, 176.0, 161.8, 153.0, 133.8, 125.6, 123.5, 120.2, 116.1, 96.0, 66.8, 65.5, 41.4, 28.6, 18.1, 15.0. IR (cm⁻¹) 2868 (w), 1702 (s), 1571 (s), 1483 (m), 1459 (s), 1371 (w), 1343 (m), 1292 (m), 1229 (w), 1159 (w), 1104 (m), 1047 (w), 958 (m), 872 (m), 764 (s), 666 (w). Anal. Calcd. for C₁₆H₁₉NO₄: C, 66.42; H, 6.62; N, 4.84. Found: C, 66.16; H, 6.99; N, 5.13.

4-Hydroxy-3-(N-(3-methoxybenzyl)ethanimidoyl)-2H-chromen-2-one (6a, JB061).

Following method C, 0.52 g (2.5 mmol) of 4 was condensed with m-methoxybenzylamine in ethanol. Elution with 1:1 (v/v) hexanes/EtOAc gave 0.52 g (64%) of 6a as white crystals, mp 146–147 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 13.97 (s, 1 H, NH), 7.93 (ddd, 7.8, 1.7, 0.5 Hz, 1 H, H5), 7.64 (ddd, 8.2, 7.3, 1.8 Hz, 1 H, H7), 7.36 (ddd, 8.1, 7.4, 0.5 Hz, 1 H, anisyl), 7.24–7.31 (m, 2 H, H6,8), 6.92–7.02 (m, 3 H, anisyl), 4.85 (d, 5.4 Hz, 2 H, NCH₂), 3.77 (s, 3 H, OCH₃), 2.71 (s, 3 H, NCCH₃). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.7, 176.3, 161.8, 159.6, 153.0, 137.4, 134.1, 130.1, 125.7, 123.6, 120.1, 119.8, 116.2, 113.6, 113.3, 96.3, 55.1, 47.5, 18.6. IR (cm⁻¹) 2943 (w), 1696 (s), 1612 (m), 1570 (m), 1467 (m), 1376 (w), 1341 (m), 1308 (m), 1270 (m), 1236 (w), 1149 (m), 1072 (w), 1042 (m), 967 (m), 901 (m), 767 (m), 702 (m). Anal. calcd for C₁₉H₁₇NO₄ • 0.25 H₂O: C, 69.61; H, 5.38; N, 4.27. Found: C, 69.50; H, 5.25; N, 4.64.

4-Hydroxy-3-(N-(4-methoxybenzyl)ethanimidoyl)-2H-chromen-2-one (6b, JB062).

Following method C, 0.51 g (2.5 mmol) of 4 was condensed with p-methoxybenzylamine in ethanol. Elution with 1:1 (v/v) hexanes/EtOAc gave 0.52 g (64%) of 6b as white crystals, mp 162–163 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 13.90 (s, 1 H, NH), 7.91 (ddd, 7.8, 1.7, 0.5 Hz, 1 H, H5), 7.63 (ddd, 8.2, 7.3, 1.7 Hz, 1 H, H7), 7.34–7.40 (m, 2 H, anisyl), 7.23–7.31 (m, 2 H, H6,8), 6.96–7.03 (m, 2 H, anisyl), 4.80 (d, 5.3 Hz, 2 H, NCH₂), 3.77 (s, 3 H, OCH₃), 2.72 (s, 3 H, NCCH₃). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.7, 175.9, 161.83, 159.0, 153.0, 134.0, 129.4, 127.7, 125.6, 123.6, 120.1, 116.2, 114.4, 96.2, 55.1, 47.1, 18.8. IR (cm ⁻¹) 2836 (w), 1704 (s), 1608 (s), 1569 (s), 1512 (m), 1465 (s), 1360 (w), 1335 (w), 1303 (m), 1249 (m), 1174 (w), 1137 (w), 1064 (w), 1028 (m), 953 (m), 823 (m), 760 (s), 686 (w). Anal. calcd for C₁₉H₁₇NO₄: C, 70.58; H, 5.30; N, 4.33. Found: C, 70.36; H, 5.14; N, 4.46.

4-Hydroxyl-2H-chromen-2-one-3-carboxaldehyde (7).³⁰

A solution of 5.3 g (33 mmol) of 4-hydroxycoumarin and 6.6 mL of POCl₃ in 60 mL of DMF was stirred at 50 °C for 18 h under nitrogen, then poured into 100 mL of a saturated aqueous solution of sodium acetate. The precipitate was collected by filtration then recrystallized from ethanol to give 4.5 g (72%) of 7 as white crystals, mp 135–137 °C (lit.^48,49 135–137, 138–140 °C). ¹H NMR (400 MHz, DMSO-d₆) δ 9.90 (s, 1 H, CHO), 7.89 (dd, 7.7, 1.6 Hz, 1 H, H5), 7.51 (m, 1 H, H7), 7.11–7.23 (m, 2 H, H6,8). ¹³C NMR (126 MHz, DMSO-d₆) δ 188.0, 177.0, 164.3, 154.4, 132.6, 125.4, 122.8, 122.3, 116.3, 102.5.

3-(N-(3-Ethoxypropyl)methanimidoyl)-4-hydroxy-2H-chromen-2-one (8a, JB100).

Following method C, 0.53 g (2.8 mmol) of 7 was condensed with 3-ethoxypropylamine in ethanol. Elution with 15:5 (v/v) hexanes/EtOAc gave 0.37 g (49%) of 8a as white crystals, mp 106–107 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 11.65 (s, 0.6 H, NH), 10.41 (s, 0.4 H, NH), 8.54 (d, 15.5 Hz, 0.3 H, NCH), 8.42 (d, 14.7 Hz, 0.7 H, NCH), 7.93 (m, 1 H, H5), 7.65 (m, 1 H, H7), 7.27–7.35 (m, 2 H, H6,8), 3.67 (q, 6.4 Hz, 2 H, NCH₂), 3.38–3.43 (m, 4 H, CH₂OCH₂), 1.87 (quint, 6.2 Hz, 2 H, NCH₂CH₂), 1.14 (t, Hz, 3 H, CH₃). ¹³C NMR (126 MHz, DMSO-d6) δ 179.3, 162.9, 162.3, 160.7, 154.2, 134.3, 125.3, 124.0, 117.0, 95.6, 67.2, 65.6, 48.6, 29.3, 15.0. IR (cm⁻¹) 3178 (w), 2967 (w), 2874 (w), 1713 (s), 1623 (s), 1567 (m), 1451 (s), 1438 (m). 1373 (m), 1332 (m), 1265 (m), 1199 (s), 1078 (s), 843 (m), 751 (s). Anal. Calcd. for C₁₅H₁₇NO₄ • 0.25 H₂O: C, 64.39; H, 6.30; N, 5.01. Found: C, 64.35; H, 6.30; N, 5.16.

3-(N-Benzylmethanimidoyl)-4-hydroxy-2H-chromen-2-one (8b, JB104).

Following method C, 0.57 g (3.0 mmol) of 7 was condensed with benzylamine in ethanol. Elution with 15:5 (v/v) hexanes/EtOAc gave 0.46 g (56%) of 8b as a white solid, mp 162.0–162.5 °C (lit.⁴² 165–167 °C). ¹H NMR (500 MHz, DMSO-d₆) δ 11.93 (s, 0.7 H, NH), 10.68 (s, 0.3 H, NH), 8.72 (d, 14.8 Hz, 0.3 H, NCH), 8.62 (d, 14.3 Hz, 0.7 H, NCH), 7.92 (m, 1 H, H5), 7.66 (m, 1 H, H7), 7.38–7.45 (m, 4 H, Ph), 7.26–7.32 (m, 3 H, H6,8, H4’), 4.83 (d, 5.9 Hz, 2 H, NCH₂Ph). ¹³C NMR (126 MHz, DMSO-d₆) δ 179.4, 162.7, 162.2, 160.6, 154.2, 136.7, 134.4, 128.8, 128.0, 125.3, 123.9, 120.2, 116.9, 96.0, 53.1. IR (cm⁻¹) 3242 (w), 3060 (w), 2962 (w), 1702 (s), 1687 (s), 1630 (s), 1598 (s), 1465 (m), 1355 (m), 1259 (m), 1225 (m), 1107 (m), 857 (m), 753 (s), 695 (m). Anal. calcd for C₁₇H₁₃NO₃ • 0.17 H₂O: C, 72.33; H, 4.76; N, 4.96. Found: C, 72.39; H, 4.43; N, 5.11.

1-(2-Chloroethoxy)-2-methoxyethane (10a).³¹

A solution of 5.0 mL (42 mmol) of diethylene glycol monomethyl ether and 3.4 mL (42 mmol) of pyridine in 20 mL of DCM was stirred under nitrogen as a solution of 3.4 mL (0.13 mol) of SOCl₂ was added at rate of 0.2 ml/min, then the resulting solution was boiled at reflux under nitrogen for 18 h. After cooling to 25 °C, the solution was diluted with 10 ml DCM and the organic layer was washed with 30 ml sat. NaHCO₃, dried with Na₂SO₄ and concentrated to dryness by rotary evaporation. The pale yellow residue underwent a simple vacuum distillation to collect 4.3 g (73%) of 10a as a clear oil (bp 70 °C, 1 mm). ¹H NMR³¹ (400 MHz, CDCl₃) δ 3.75 (m, 2 H, OCH₂CH₂O), 3.65 (m, 4 H, CH₂OCH₂CH₂O), 3.55 (m, 2 H, CH₂Cl), 3.39 (s, 3 H, CH₃). ¹³C NMR (101 MHz, CDCl₃) δ 71.9, 71.4, 70.6, 59.0, 42.6.

1-(Chloromethyl)-4-methoxybenzene (10b).

A solution of 5.0 g (36 mmol) of 4-methoxybenzyl alcohol and 8.2 mL (73 mmol) of thionyl chloride in 5 mL of anhydrous THF was stirred at 25 °C for 12 h under nitrogen, then 10 mL of water was added. The resulting solution was extracted with DCM (2 × 20 mL) and the combined extracts were washed with 10 mL of a saturated aqueous solution of NaHCO₃, dried (Na₂SO₄), filtered and concentrated to dryness by rotary evaporation, giving 4.7 g (82%) of 10b as a yellow oil, which was used in the next step without further purification. ¹H NMR⁵⁰ (400 MHz, CDCl₃) δ 7.31 (m, 2 H, H2,6), 6.86 (m, 2 H, H3,5), 4.55 (s, 2 H, CH₂), 3.79 (s, 3 H, CH₃).

5-(Chloromethyl)-1,3-benzodioxole (10c).

A solution of 5.0 g (33 mmol) of piperonol and 7.5 ml (66 mmol) of thionyl chloride in 5 mL of anhydrous THF was stirred at 25 °C under nitrogen for 12 h, then 10 mL of water was added. The resulting solution was extracted with DCM (2 × 20 mL) and the combined extracts were washed with 10 mL of a solution of saturated aqueous NaHCO₃, dried (Na₂SO₄), filtered and concentrated to dryness by rotary evaporation, giving 5.2 g (92%) of 10c as a yellow oil, which was used in the next step without further purification. ¹H NMR⁵¹ (400 MHz, CDCl3) δ 6.86 (dd, 1.8, 0.4 Hz, 1 H, H4), 6.83 (ddt, 7.9, 1.8, 0.5 Hz, 1 H, H6), 6.76 (dd, 7.9, 0.5 Hz, 1 H, H7), 5.96 (s, 2 H, OCH₂O), 4.51 (s, 2 H, CH₂Cl).

General procedure for synthesis of primary amine intermediates 12a-c. A suspension of 1 mole equivalent of NaH (60% in mineral oil) in 10 ml of anhydrous THF was stirred at 0 °C under nitrogen as 1 mole equivalent of 3-aminopropanol (11) was added slowly, then the mixture was boiled under reflux for 2 h. The mixture was cooled to 0 °C and stirred under nitrogen as 0.9 mol equivalents of the alkyl chloride was added slowly, then the reaction mixture was boiled under reflux for 2 h and cooled to room temperature. The solvent was removed by rotary evaporation then 10 mL of a 1 N aqueous HCl solution was stirred with the residue. The resulting mixture was washed with 10 mL of DCM then basified to pH ~11 with 10% aqueous NaOH solution and extracted with DCM (2 × 20 mL). The combined extracts were dried (Na₂SO₄), filtered and concentrated to dryness by rotary evaporation and the residue was used in the next step without further purification.

3-(2-(2-Methoxyethoxy)ethoxy)propylamine (12a).

Reaction of 0.20 g (8.3 mmol) of NaH (60% in mineral oil), 0.64 mL (8.3 mmol) of 11, and 1.0 g (7.3 mmol) of 10a gave 12a as a yellow oil.

3-(4-Methoxybenzyloxy)propylamine (12b).

Reaction of 0.44 g (11 mmol) of NaH (60% in mineral oil), 0.83 mL (11 mmol) of 11, and 1.5 g (9.7 mmol) of 10b gave 0.99 g (52%) of 12b as a yellow oil. ¹H NMR⁵² (400 MHz, CDCl3) δ 7.25 (m, 2 H, H2,6), 6.86 (m, 2 H, H3,5), 4.42 (s, 2 H, PhCH₂O), 3.78 (s, 3 H, CH₃), 3.51 (t, 6.2 Hz, 2 H, OCH₂CH₂), 2.79 (t, 7.0 Hz, 2 H, CH₂NH₂), 1.72 (m, 2 H, OCH₂CH₂). 1.41 (s, 2 H, NH₂).

3-(1,3-Benzodioxol-5-ylmethoxy)propylamine (12c).

Reaction of 1.4 g (34 mmol) of NaH (60% in mineral oil), 2.6 mL (34 mmol) of 11, and 5.2 g (30 mmol) of 10c gave 4.2 g (66%) of 12c as a yellow oil. ¹H NMR (400 MHz, CDCl₃) δ 6.72–6.86 (m, 3 H, H4,6,7), 5.93 (s, 2 H, OCH₂O), 4.39 (s, 2 H, OCH₂Ar), 3.52 (t, 6.2 Hz, 2 H, CH₂CH₂O), 2.82 (m, 2 H, NCH₂), 1.74 (quint., 6.6 Hz, 2 H, CH₂CH₂O).

4-Hydroxy-3-(N-(3-(2-(2-methoxyethoxy)ethoxypropyl)ethanimidoyl)-2H-chromen-2-one (13a, JB102).

Following method C, 0.15 g (0.72 mmol) of 4 was condensed with 12a in ethanol. Elution with 3:97 (v/v) methanol/DCM gave 0.17 g (13% over 2 steps) of 13a as a yellow oil. ¹H NMR (400 MHz, CDCl₃) δ 8.04 (m, 1 H, H5), 7.53 (m, 1 H, H7), 7.18–7.26 (m, 2 H, H6,8), 3.57–3.73 (m, 10 H, NCH₂, CH₂OCH₂CH₂OCH₂), 3.51–3.57 (m, 2 H, CH₂OCH₃), 3.37 (s, 3 H, OCH₃), 2.75 (s, 3 H, NCCH₃), 2.04 (m, 2 H, NCH₂CH₂). ¹³C NMR (126 MHz, CDCl₃) δ 181.1, 176.6, 162.9, 153.6, 133.7, 126.0, 123.5, 116.5, 97.1, 72.0, 70.6, 70.5, 67.6, 59.1, 41.2, 29.2, 18.2. IR (cm⁻¹) 2871 (m), 1700 (s), 1606 (s), 1572 (m), 1464 (s), 1337 (s), 1305 (w), 1231 (w), 1202 (w), 1102 (s), 1029 (w), 954 (w), 899 (w), 759 (s), 668 (w). Anal. Calcd. for C₁₉H₂₅NO₆: C, 62.80; H, 6.93; N, 3.85. Found: C, 62.56; H, 6.98; N, 4.07.

4-Hydroxy-3-(N-(3-(4-methoxybenzyloxy)propylethanimidoyl)-2H-chromen-2-one (13b, JB080).

Following method B, condensation of 0.31 g (1.6 mmol) of 4 with 12b in ethanol gave 0.39 g (67%) of 13b as white crystals, mp 92–93 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 13.73 (s, 1 H, NH), 7.94 (dd, 7.8, 1.8 Hz, 1 H, H5), 7.63 (ddd, 8.2, 7.2, 1.7 Hz, 1 H, H7), 7.2–7.33 (m, 4 H, H6,8, H2’,6’), 6.85 (m, 2 H, H3’,5’), 4.41 (s, 2 H, OCH₂Ar), 3.69 (s, 3 H, OCH₃), 3.66 (q, 6.2 Hz, 2 H, NCH₂), 3.52 (t, 5.9 Hz, 2 H, NCH₂CH₂CH₂O), 2.64 (s, 3 H, CCH₃), 1.93 (p, 6.5 Hz, 2 H, CH₂CH₂CH₂). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.5, 176.1, 161.8, 158.6, 153.0, 133.9, 130.2, 129.2, 125.6, 123.6, 120.2, 116.2, 113.5, 96.0, 71.6, 66.1, 54.9, 41.3, 28.5, 18.1. Anal. Calcd. for C₂₂H₂₃NO₅• 0.25H₂O: C, 68.47; H, 6.24; N, 3.63. Found: C, 68.32; H, 6.33; N, 3.68.

3-(N-(3-(1,3-Benzodioxol-5-ylmethoxy)-4-hydroxypropylethanimidoyl)-2H-chromen-2-one (13c, JB092).

Following method C, 0.77 g (3.8 mmol) of 4 was condensed with 12c in ethanol. Elution with 7:10 (v/v) hexanes/EtOAc gave 0.59 g (40%) of 13c as a yellow solid, mp 89–91 °C. ¹H NMR (400 MHz, CDCl₃) δ 8.04 (dd, 7.8, 1.7 Hz, 1 H, H5), 7.53 (ddd, 8.2, 7.3, 1.7 Hz, 1 H, H7), 7.18–7.26 (m, 2 H, H6,8), 6.70 – 6.83 (m, 3 H, H4’,6’,7’), 5.90 (s, 2 H, OCH₂O), 4.42 (s, 2 H, OCH₂Ar), 3.67 (td, 6.8, 5.5 Hz, 2 H, NCH₂CH₂CH₂O), 3.59 (t, 5.9 Hz, 2 H, NCH₂CH₂CH₂O), 2.73 (s, 3 H, CCH₃), 2.03 (m, 2 H, CH₂CH₂CH₂). ¹³C NMR (101 MHz, DMSO-d₆) δ 179.5, 176.1, 161.8, 153.0, 147.2, 146.5, 133.9, 132.1, 125.6, 123.5, 121.2, 120.2, 116.1, 108.2, 107.8, 100.8, 96.0, 71.8, 66.2, 41.3, 28.5, 18.1. Anal. Calcd. for C₂₂H₂₁NO₆: C, 66.83; H, 5.35; N, 3.54. Found: C, 66.69; H, 5.31; N, 3.58.

N-Butyl-N-3-methoxybenzylamine (15).

A solution of 2.0 g (15 mmol) of m-anisaldehyde and 1.6 ml (16 mmol) of butyl amine in 10 mL of chloroform was boiled at reflux under nitrogen for 12 h then cooled to 25 °C. The solution was dried (Na₂SO₄), filtered and concentrated to dryness by rotary evaporation. A solution of the residue in 15 mL of EtOH was stirred at 0 °C as 0.56 g (15 mmol) of NaBH₄ was added, then stirred at 25 °C for 4 h. The solvent was removed by rotary evaporation and a solution of the residue in 20 mL of water was extracted with DCM (3 × 20 mL). The combined extracts were dried (Na₂SO₄), filtered and concentrated to minimum volume by rotary evaporation, giving 2.7 g (94%) of 15 as a yellow oil, was used without further purification. ¹H NMR⁵³ (400 MHz, CDCl₃) δ 7.23 (m, 1 H, H5), 6.87–6.91 (m, 2 H, H4,6), 6.78 (m, 1 H, H2), 3.79 (s, 3 H, OCH₃), 3.76 (s, 2 H, NCH₂Ar), 2.63 (m, 2 H, NCH₂CH₂), 1.49 (m, 2 H, NCH₂CH₂), 1.35 (m, 2 H, NCH₂CH₂CH₂), 1.25 (m, 1 H, NH), 0.91 (t, 7.3 Hz, 3 H, CH₂CH₃).

N’-Butyl-N’-(3-methoxybenzyl)-N-3-aminopropylphthalimide (17).

A solution of 3.4 g (17 mmol) of 15, 7.0 g (26 mmol) of N-3-bromopropylphthalimide, 0.29 g (1.7 mmol) of potassium iodide, and 2.6 g (19 mmol) of potassium carbonate in 40 ml of anhydrous CH₃CN was boiled at reflux under nitrogen for 48 h then the solvent was removed by rotary evaporation. A mixture of the residue and 40 mL of water was extracted with DCM (3 × 20 mL) and the combined extracts were dried (Na₂SO₄), filtered and concentrated to dryness by rotary evaporation. Purification by silica gel column chromatography eluting with 3:7 (v/v) EtOAc/hexanes gave 1.7 g (25%) of 17 as a white solid. ¹H NMR (400 MHz, CDCl₃) δ 7.84 (m, 2 H, H2,5), 7.70 (m, 2 H, H3,4), 7.16 (ddd, 8.2, 7.5, 0.4 Hz, 1 H, H5’), 6.87 – 6.94 (m, 2 H, H2’,6’), 6.73 (ddd, 8.2, 2.7, 1.0 Hz, 1 H, H4’), 3.80 (s, 3 H, OCH₃), 3.71 (m, 2 H, CONCH₂), 3.52 (s, 2 H, NCH₂Ph), 2.49 (t, 7.0 Hz, 2 H, CON(CH₂)₂CH₂N), 2.41 (t, 7.2 Hz, 2 H, NCH₂(CH₂)₂CH₃), 1.84 (m, 2 H, CONCH₂CH₂CH₂N), 1.44 (m, 2 H, NCH₂CH₂CH₂CH₃), 1.31 (m, 2 H, CH₂CH₃), 0.86 (t, 7.3 Hz, 3 H, CH₂CH₃). ¹³C NMR (101 MHz, CDCl₃) δ 168.5, 159.7, 141.8, 133.9, 132.3, 129.1, 123.3, 121.2, 114.1, 112.5, 58.7, 55.3, 53.6, 51.2, 36.6, 29.3, 26.3, 20.7, 14.2.

N-(3-Aminopropyl)-N-butyl-N-(3-methoxybenzyl)amine (18).

A solution of 1.5 g (3.8 mmol) of 17 and 0.38 g (7.7 mmol) of hydrazine monohydrate in 25 mL of ethanol was boiled at reflux under nitrogen for 18 h then the solvent was removed by rotary evaporation. A mixture of the residue with 20 mL of a 5% aqueous NaOH solution was stirred for 30 min, diluted with 25 mL of water, then extracted with DCM (3 × 30 mL). The combined extracts were dried (Na₂SO₄) and concentrated to minimum volume by rotary evaporation, giving 0.87 g (91%) of 18 as a colorless oil that was used without further purification. ¹H NMR (400 MHz, CDCl₃) δ 7.20 (m, 1 H, H5), 6.88–6.92 (m, 2 H, H2,6), 6.77 (m, 1 H, H4), 3.79 (s, 3 H, OCH₃), 3.51 (s, 2 H, NCH₂Ph), 2.71 (t, 6.8 Hz, 2 H, H₂NCH₂), 2.36–2.48 (m, 4 H, CH₂CH₂NCH₂, CH₂NCH₂CH₂), 1.59 (quint., 6.8 Hz, 2 H, H₂NCH₂CH₂), 1.47 (m, 2 H, NCH₂CH₂CH₂CH₃), 1.24–1.40 (m, 4 H, NH₂, CH₂CH₃), 0.88 (t, 7.3 Hz, 3 H, CH₂CH₃). ¹³C NMR (101 MHz, CDCl₃) δ 159.7, 142.1, 129.1, 121.2, 114.2, 112.2, 58.8, 55.2, 53.7, 51.5, 40.6, 31.0, 29.3, 20.7, 14.2.

3-(N-(N’-Butyl-N’−3-methoxybenzyl-3-aminopropyl)ethanimidoyl)-4-hydroxy-2H-chromen-2-one (19, JB099).

Following method C, 0.31 g (1.2 mmol) of 4 was condensed with 18 in ethanol. Elution with 8:2 (v/v) hexanes/EtOAc gave 0.39 g (58%) of 19 as yellow powder, mp 55–56 °C. ¹H NMR (400 MHz, CDCl₃) δ 8.02 (ddd, 7.8, 1.8, 0.5 Hz, 1 H, H5), 7.53 (ddd, 8.2, 7.2, 1.7 Hz, 1 H, H7), 7.13–7.25 (m, 3 H, H6,8, H5’), 6.83–6.88 (m, 2 H, H2’,6’), 6.69 (m, 1 H, H4’), 3.73 (s, 3 H, OCH₃), 3.48–3.51 (m, 4 H, CNCH₂(CH₂)₂N, NCH₂Ph), 2.67 (s, 3 H, NCCH₃), 2.56 (t, 6.4 Hz, 2 H, CN(CH₂)₂CH₂N), 2.46 (m, 2 H, NCH₂(CH₂)₂CH₃), 1.85 (quint., 6.7 Hz, 2 H, CNCH₂CH₂CH₂N), 1.49 (m, 2 H, NCH₂CH₂CH₂CH₃), 1.34 (m, 2 H, N(CH₂)₂CH₂CH₃), 0.90 (t, 7.3 Hz, 3 H, CH₂CH₃). ¹³C NMR (126 MHz, CDCl₃) δ 180.9, 176.2, 159.6, 153.6, 141.4, 133.6, 129.1, 125.9, 123.4, 121.1, 120.7, 116.5, 114.7, 111.7, 96.9, 58.7, 55.0, 55.0, 54.0, 50.3, 42.2, 29.3, 26.8, 20.6, 18.6, 14.1. Anal. Calcd. for C₂₆H₃₂N₂O₄: C, 71.53; H, 7.39; N, 6.42. Found: C, 71.20; H, 7.10; N, 6.51.

4-Hydroxy-3-(3-(4-hydroxy-3-methoxyphenyl)-1-oxo-2-propen-1-yl)-2H-1-Benzopyran-2-one (20, JB077).

A solution of 1.2 g (5.9 mmol) of 4, 0.90 g (5.9 mmol) of vanillin, and 0.35 mL (3.6 mmol) of piperidine in 10 mL of chloroform was boiled at reflux under nitrogen for 18 h then cooled to room temperature and the solvent was removed by rotary evaporation. A solution of the residue in 10 mL of chloroform was boiled under reflux with activated carbon for 1 h then vacuum filtered through a celite plug. Flash chromatography eluting with 7:3 (v/v) hexane/EtOAc gave 0.22 g (11%) of 20 as a yellow powder, mp 123–124 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.07 (s, 1 H, OH), 8.14 (d, 15.6 Hz, 1 H, H5), 7.98–8.09 (m, 2 H, H7, COCHCHPh), 7.82 (m, 1 H, COCHCHPh), 7.43 (m, 2 H, H6,8), 7.32 (m, 2 H, H4’,7’), 6.91 (d, 8.0 Hz, 1 H, H6’), 3.86 (s, 3 H, OCH₃). ¹³C NMR (126 MHz, DMSO-d₆) δ 190.5, 181.2, 159.6, 154.0, 151.2, 148.1, 148.0, 136.3, 125.9, 125.3, 124.5, 124.3, 118.1, 116.7, 116.3, 116.1, 112.7, 100.2, 55.6. MS (ESI+) m/z 339.10 (MH⁺). IR (cm⁻¹) 3170 (m), 2919 (w), 1672 (s), 1611 (m), 1588 (s), 1488 (s), 1418 (s), 1318 (m), 1153 (m), 1117 (m), 1028 (m), 977 (m), 898 (w), 843 (m), 759 (s), 670 (m). Anal. Calcd. for C₁₉H₁₄O₆ • 0.5 H₂O: C, 65.71; H, 4.35. Found: C, 65.67; H, 4.17.

Pharmacokinetic evaluation: ClogP.

Octanol/water partition coefficients were calculated with Chemaxon Marvinsketch 22.13 (https://chemaxon.com/products/marvin), according to a consensus model based on previous literature.^54–56

Kinetic solubility.

The following procedure⁵⁷ was used to determine kinetic solubility. Solutions of each compound with various starting concentrations in 2.5% DMSO with a final volume of 100 μL were prepared in quadruplicate by serial dilution on sterile 96-well microplates. Using a micropipette, the appropriate amount of a 20 mM DMSO stock solution for each compound tested was added into a vial contained 858 μL of PBS buffer (pH 7.4) and the appropriate amount of DMSO in order to keep the DMSO concentration constant. A second starting solution with lower concentration of the compound was prepared in the same manner. Mixing with a micropipette, serial dilutions were performed from the starting solution with each well containing a final volume of 100 μL. Each plate was placed in a Nepheloskan Ascent nephelometer with the lid off, shaken for 2 min, incubated for 1 h at 37 °C, and shaken again for 3 min before measuring each well. The raw data was adjusted by subtracting the average blank value and analyzed using Microsoft Excel.

Hydrolytic stability.

Hydrolysis experiments were performed at 30 °C in the same assay buffer used in the ATPase assay, 10 mM 3-morpholinopropane-1-sulfonic acid (MOPS), pH 7.0, 50 mM NaCl, 0.1 mM ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), 5 mM MgCl₂, 2 mM ATP (Sigma; A3377) and 1 mM dithiothreitol (DTT). A 10 μL aliquot of a 20 mM solution of each compound in DMSO was mixed thoroughly with 490 μL of DMSO and 24.5 mL of ATPase assay buffer (final concentration 8 μM, 2% DMSO). Immediately after preparation a 3 mL aliquot was removed and the UV spectrum was recorded on a Shimadzu UV-2550 UV-visible spectrophotometer in steps of 0.5 nm with a quartz cuvette (1 cm pathlength). The remaining solution was kept at 30 °C in a glass vial sealed with parafilm and every 24 h another aliquot was removed and used to record the UV spectrum. Reference solutions were prepared in ATPase assay buffer (8 μM of the compound, 2% DMSO). Immediately after preparation of each reference solution, a UV spectrum was recorded as described above. See Supporting Information for UV plots and further information.

Biological evaluation: Reagents.

All reagents unless specified otherwise were from Sigma-Aldrich. Specific product numbers are given, if important.

Proteins.

Skeletal muscle myosin was purified⁵⁸ from fresh rabbit psoas muscle and cardiac myosin was purified⁵⁹ from fresh bovine left ventricles. Both types of myosin were stored at −20°C in 50% aq. glycerol and were shown by urea gel electrophoresis to be unphosphorylated. Smooth muscle myosin was purified from fresh chicken gizzards (Pel-freeze) similar to a previously described⁶⁰ and phosphorylated as previously described.⁶¹ Smooth muscle myosin was never frozen, stored on ice, and used within 2–3 weeks. All smooth muscle myosin samples assayed for actin-activated Mg²⁺ ATPase activity were fully phosphorylated as assessed by urea gel electrophoresis. Actin was purified from rabbit psoas muscle⁶² and stored on ice as F-actin.

ATPase assays.

Steady-state actin-activated Mg²⁺ATPases were performed at 30 °C, as previously described.²⁵ The assay buffer was 10 mM 3-morpholinopropane-1-sulfonic acid (MOPS), pH 7.0, 50 mM NaCl, 0.1 mM ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), 5 mM MgCl₂, 2 mM ATP (Sigma; A3377) and 1 mM dithiothreitol (DTT). The actin concentration was 32 μM. The final myosin concentration was 0.1, 0.5, and 0.2 mg/mL for skeletal, cardiac, and smooth muscle myosin, respectively. Dilution series of all analogs were prepared in DMSO from a 25 mM stock solution and added such that the DMSO concentration in the reaction mixture was 2 %. All samples were protected from light and stored at −20 °C.

Assay for inhibition of cytokinesis in Cos-7 cells.

The cytokinesis inhibition screen was as previously described,⁶³ with minor modifications. Cos-7 cells (4,000 cells in 100 μL per well; ATCC Cat. # CRL-1651) were plated onto a 96-well tissue culture plate and incubated for 24 h with 5% CO₂ at 37 °C prior to adding 100 μL of a 2X solution (see below) of the hydroxycoumarin analogs (final concentration 40 μM) or positive controls, para-aminoblebbistatin (20 μM, Thermo; 226991) and (−)-blebbistatin (10 μM; Calbiochem CAS 856925–71). Negative controls were 1% DMSO in non-fluorescent DMEM and media alone. Stock solutions of compounds were prepared in DMSO, diluted 50-fold into growth media, then 2-fold after plating the compounds, resulting in a final DMSO concentration of 1%. For each independent experiment done on 96-well plate, the four controls, two negatives and two positives, were done with 6 replicate wells therefore two complete rows on the plate were dedicated to controls. Each compound was done with 3 replicate wells on each independent 96-well plate experiment. Each compound was assessed in at least two independent experiments and at least 600 cells were imaged in each well. The cells were incubated in the presence of compounds for 24 hr at 37 °C with 5% CO₂. A staining solution was then added to the cells containing fluorescein-diacetate (6 μM; Sigma; F7378), Hoechst 33342 (10 μM; Life Technologies; H3570), and propidium iodide (4 μM; Life Technologies; P3566) in media and incubated for 10 min. The staining solution was aspirated and replaced with non-fluorescent FluoroBrite™ DMEM media (Life Technologies; A1896701). Light exposure was limited for all steps after the addition of compounds. After staining, the cells were imaged and analyzed in CellProfiler⁶⁴ to determine the nuclei to cell ratio (N/C). A Discover Echo Revolve 4 fluorescence microscope was used for imaging with Discover Echo DAPI (ex: 380/30 nm; em 450/50 nm; dm: 425 nm), FITC (ex: 470/40 nm; em: 525/50 nm; dm: 495 nm), and TRITC (ex: 530/40 nm; em: 605/70 nm; dm: 560) filter cubes. The number of cells stained with FDA and the number of nuclei stained with Hoechst33342 were counted for each image of the cells. Nuclei that were positive for both Hoechst33342 and propidium iodide were not included in the N/C calculation as propidium iodide is a membrane-impermeable nuclear stain that cannot stain live nuclei with intact cell and nuclear membranes. When cytokinesis is inhibited the N/C will increase.

Cytotoxicity Assay.

Reagents purchased from Thermo Fisher, Fisher Scientific, and Acros Organics were of ACS reagent grade and was used after purification by 0.2 μm HPTFE filtration. Supplies purchased from Fisher Scientific and Nexday Science were used after sterilization by high temperature steam autoclave. Cell assay was incubated in Fisherbrand Isotemp oven at 37 °C in 5% CO₂ atmosphere. Cell density was determined with Nexcelom Bioscience Cellometer Auto T4 cell counter.

A 75 cm² flask containing Cos-7 cells ATCC (Cat. # CRL-1651)) incubated with complete growth media (10% fetal bovine serum, 5% non-essential amino acids, 5% penicillin-streptomycin, 80% Dulbeco’s modified eagle medium), henceforth labeled simply as media, was prepared for the MTT assay. The media was removed from the flask were rinsed with 4 mL of PBS, which is subsequently disposed. The cells were detached by adding 2 mL of 0.25% trypsin-EDTA media and incubating for 10 min. The detached cells were suspended with 2 ml media and transferred to a centrifuge test tube containing 7 mL of media. The cell suspension was centrifuged for 10 min at 8000 rpm and the resulting supernatant was decanted. The cells were re-suspended with 5 mL of media to create the mother solution and a 20 μL aliquot was diluted with 60 μL of PBS. Cell density was determined by mixing 10 μL of cell aliquot and 10 μL of trypan blue solution (0.2% w/v in water) and transferring the dyed solution to automatic cell counter. Cell suspensions were diluted with media to achieve a cell concentration to seed 2.5 × 104 cells per well in 200 μL on a 96 well plate. Plate layout included 4 wells for each compound, 4 wells seeded with cells that were not treated with compounds and 4 wells that were not cell seeded as controls. Plates were incubated for 24 h after which the media was replaced with 200 μL of media with 1% DMSO or 20, 40, 100, or 200 μM compound. For control wells, with and without cells, media was replaced with 200 μL of media containing 1% DMSO. After incubating for 24 h, the treatment media was replaced with 100 μL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution (1.2 mM in PBS), incubated for 2 h protected from light, and cells were lysed by adding 200 μL of 0.1% NP-40 (v/v) in isopropanol containing 0.04 M aq. HCl. Plates were agitated to dissolve the dark blue formazan crystals and absorbance at 564 nm was immediately measured on a BioTek synergy H1 microplate reader.

Homology models and docking.

The Oryctolagus cuniculus skeletal muscle myosin heavy chain 4 (PDB: 6ysy)¹⁸ was used as the template to build a homology model of the bovine cardiac β-myosin heavy chain (UniProt Q9BE39 (MYH7_BOVIN; gene: MYH7)), chicken gizzard smooth muscle myosin (UniProt P10587 (MYH11_CHICK; gene: MYH11)) using SWISS-MODEL.⁶⁵ See the Supporting Information for reference 25 for PDB files for homology models of bovine cardiac and rabbit skeletal myosins (Figure 4). AutoDock Vina⁶⁶ was used to dock 13c to the rabbit skeletal myosin structure (PDB: 6ysy) with all ligands removed. The entire protein was used as the search space. Since one structure (PDB: 6ysy) was used as a template for all homology models, the 3-dimensional coordinates of the protein backbone, especially in highly conserved regions, for the homology models were nearly identical. Therefore, the homology model structures were overlayed on the skeletal myosin structure (PDB: 6ysy) to allow for a direct comparison of the residue identity and conformational differences in highly conserved region of the heavy chain, specifically the blebbistatin and hydroxycoumarin binding pocket in the cleft between the upper 50 kDa and lower 50 kDa domains.

Abbreviations:

AP

action potential

BHC

(N-butylethanimidoyl)ethyl)-4-hydroxy-2H-chromen-2-one

blebb

blebbistatin

paB

para-aminoblebbistatin

DCM

dichloromethane

DMEM

Dulbecco’s Modified Eagle Medium

EGTA

ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid

FDA

fluorescein diacetate

MOPS

3-morpholinopropane-1-sulfonic acid

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

N/C

nuclei to cell ratio

NHA

non-hydrogen atom

NM-II

nonmuscle myosin II

NM-IIA

nonmuscle myosin IIA

NM-IIB

nonmuscle myosin IIB

NM-IIC

nonmuscle myosin IIC

NMJ

neuromuscular junction

OD

optical density

PI

propidium iodide

Pi

inorganic phosphate

SAR

structure-activity relationship