Hydrophilic, Potent, and Selective 7-Substituted 2-Aminoquinolines as Improved Human Neuronal Nitric Oxide Synthase Inhibitors

Abstract

Neuronal nitric oxide synthase (nNOS) is a target for development of anti-neurodegenerative agents. Most nNOS inhibitors mimic L-arginine and have poor bioavailability. 2-Aminoquinolines showed promise as bioavailable nNOS inhibitors, but suffered from low human nNOS inhibition, low selectivity versus human eNOS, and significant binding to other CNS targets. We aimed to improve human nNOS potency and selectivity and reduce off-target binding by (a) truncating the original scaffold or (b) introducing a hydrophilic group to interrupt the lipophilic, promiscuous pharmacophore and promote interaction with human nNOS-specific His342. We synthesized both truncated and polar 2-aminoquinoline derivatives and assayed them against recombinant NOS enzymes. Although aniline and pyridine derivatives interact with His342, benzonitriles conferred the best rat and human nNOS inhibition. Both introduction of a hydrophobic substituent next to the cyano group and aminoquinoline methylation considerably improved isoform selectivity. Most importantly, these modifications preserved Caco-2 permeability and reduced off-target CNS binding.

Graphical abstract

Introduction

Neurodegenerative disorders (such as Alzheimer's and Parkinson's diseases, among others) result in severe neurological, cognitive, and motor deficits. The prevalence of these diseases has increased over the past century, and this upward trend is predicted to continue, especially with an increasingly aged population,¹ making the development of therapeutic agents to treat these disorders a priority. In addition, the neuronal damage observed in these diseases can also be observed in patients with neuropathic pain, stroke, cerebral palsy, and head trauma.

We have been investigating the enzyme neuronal nitric oxide synthase (nNOS) as a potential target for anti-neurodegenerative therapeutics. In the brain, nitric oxide (NO, produced by nNOS) is needed for neuronal signaling, but, under conditions of neurodegeneration, NO levels are high, owing to overexpressed or unregulated nNOS. This increased NO can form reactive species such as peroxynitrite, leading to neuronal cell damage.² Indeed, nNOS^3,4,5,6 has been implicated in the pathogenesis of neurodegeneration and peripheral nerve dysfunction (neuropathic pain), and inhibition of nNOS has shown promise in treating or preventing neuronal damage in animal models.^7,8

NO is formed when homodimeric nNOS converts L-arginine to L-citrulline, which occurs in the enzyme's oxygenase domain. Electrons (from NADPH) are shuttled through redox cofactors (FAD and FMN) in the reductase domain, and then move from one monomer's reductase domain to the other's oxygenase domain,⁹ where electron transfer proceeds to (6R)-5,6,7,8-tetrahydrobiopterin (H₄B), and finally to heme, which then oxidizes the bound L-arginine.¹⁰ Along with nNOS, two other isoforms of NOS (inducible NOS, or iNOS, and endothelial NOS, or eNOS) make NO in humans. iNOS is involved in the immune response, whereas the NO produced by eNOS regulates smooth muscle tone and blood pressure.

Over the years, several classes of nNOS inhibitors have been developed in our laboratories (1-6 are representative examples, Figure 1).^{11,12,13,14,15} These compounds are all competitive with L-arginine; compound 3 is also a heme-iron coordinator.¹² A general challenge in development of arginine-mimetic inhibitors is that molecules that resemble arginine often possess chemical properties (basicity, H-bonding potential, and high polar surface area, or PSA) that impede oral bioavailability and passage through the blood-brain barrier (BBB; compounds 1, 2, and 6, for example, have poor Caco-2 permeability).

Figure 1. Previously designed human nNOS inhibitors.

In addition to having good BBB permeability, nNOS inhibitors must be selective for the neuronal isoform over iNOS and eNOS. For example, eNOS inhibition can cause cardiovascular problems and hypertension.¹⁶ Nonspecific iNOS inhibition could interfere with immune defense, but its role in the brain is more complex; one study indicated that mice lacking iNOS develop enhanced Alzheimers' pathology,¹⁷ while others indicate that iNOS inhibition could actually be neuroprotective.¹⁸ A more recent challenge has been designing inhibitors with high potency against human nNOS (hnNOS). Historically, many structure-based design efforts have utilized rat nNOS (rnNOS), so many compounds (such as most shown in Figure 1) are selective for the rat enzyme, leading to weaker hnNOS inhibitors with low selectivity over human eNOS (heNOS).

The 2-aminoquinolines (compounds 4-6) have shown considerable promise as a scaffold for further nNOS inhibitor development. Compounds 4 and 5 have good potency, selectivity, and cellular permeability.¹⁴ They also exhibit good oral bioavailability in mice (4) and brain penetration in rats (5). Unfortunately, compound 5, considered our best lead for further development, had an unfavorable safety profile in rats, where it caused neurological and cardiovascular side effects at the doses required for effective nNOS inhibition. Counter-screening against a variety of CNS targets via the Psychoactive Drug Screening Program's (PDSP's) radioligand binding assay¹⁹ revealed that 5, with its GPCR-ligand-like pharmacophore, is a promiscuous binder (affinity < 100 nM) at serotonin, opioid, and histamine receptors, and its scaffold also resembles known human ether-a-go-go-related-gene (hERG) channel blockers. It is imperative that specific CNS-acting NOS inhibitors not strongly bind numerous other targets in the brain, as this could cause neurological, psychological, or other side effects. Additionally, compound 5 (and 4) has poor hnNOS activity and low hnNOS/heNOS (hn/he) selectivity.

One strategy for reducing the off-target binding of these compounds involved rearranging the pharmacophore of 5 to disrupt the GPCR ligand-like core, leading to compounds such as 6,¹⁵ where the position of the polar amine and hydrophobic aryl group are “inverted” relative to 5. Unfortunately, this compound had much lower cellular permeability and its hnNOS activity, while improved from 5, was still low, as was its hn/he selectivity. We therefore, in addition to our newer studies on phenyl-ether linked aminoquinolines^15,20, sought alternative (and parallel) strategies for reducing the off-target binding of 5 and improving its hnNOS activity and hn/he selectivity.

One such strategy, detailed therein, is based on a key structural difference between rnNOS and hnNOS. While rnNOS possesses an isoform-specific hydrophobic pocket, consisting of Tyr706, Leu337, and Met336, hnNOS lacks the leucine residue, which is replaced by His342.²¹ This pocket in hnNOS is smaller and more polar,²² and preliminary crystallography studies indicated that the hydrophobic haloaryl tail of 4, responsible for much of this compound's stabilization when bound to the hydrophobic pocket of rnNOS¹⁴, is not well-accommodated in the analogous region of hnNOS (possibly because of the bulky and polar histidine), resulting in the loss of hydrophobic contacts and much lower hnNOS activity. Our initial strategy consisted of simply truncating the linker chain between the quinoline and the hydrophobic aryl ring, hoping that lack of repulsion between the aryl ring and His342 might improve the selectivity for hnNOS over rnNOS (as observed for truncated 2-aminopyridines).²³

To this end, compounds 7-11 (Figure 2), bearing substituents on the aryl ring capable of a wide variety of interactions, were prepared and assayed against rnNOS and hnNOS. Although less active against these isoforms compared to compounds 4-6 (Table 1, vide infra), the good rat/human (rn/hn) selectivity of dimethylamines 10 and 11 encouraged the development of a second strategy – to directly interact with His342 itself.

Figure 2. Overall strategy for design of hydrophilic 2-aminoquinolines based on lead 5.

Table 1. Inhibition of NOS enzymes by aminoquinoline analogues 7-25^a.

Compound	Ki (μM)a				Selectivity
	rnNOS	hnNOS	miNOS	beNOS	rn/mi	rn/be
4	0.049	0.318	44.0	11.2	899	228
5	0.066	0.440	28.4	7.24	431	110
6	0.058	0.295	27.7	12.5	478	216
7	0.529	2.05	15.3	7.95	29	15
8	0.575	0.768	NT	NT	ND	ND
9	1.32	2.17	NT	NT	ND	ND
10	4.83	4.30	NT	NT	ND	ND
11	1.82	1.10	NT	NT	ND	ND
12	0.421	1.02	30.1	6.53	71	16
13	0.522	0.764	42.2	8.42	81	16
14	0.085	0.076	13.3	NT	156	ND
15	0.036	0.274	10.0	1.57	278	44
16	0.038	0.108	11.2	1.04	295	27
17	0.090	0.130	13.8	0.562	153	7
18	0.050	0.073	17.2	1.86	344	37
19	0.216	0.164	84.2	NT	390	ND
20	0.041	0.050	25.0	0.273	609	7
21	0.037	0.032	21.3	0.581	575	16
22	0.021	0.020	10.3	0.092	492	4
23	0.031	0.021	5.15	0.115	166	4
24	0.019	0.052	4.70	1.22	247	64
25	0.025	0.030	4.83	0.468	193	19

^aCompounds 7-25 were assayed in vitro against four purified NOS isoforms: rat nNOS, human nNOS, bovine eNOS, and murine iNOS, using known literature methods, and K_i values are calculated directly from IC₅₀ values using the Cheng-Prusoff equation (see Experimental Section for details). IC₅₀ values are the average of at least two replicates from 7-9 data points; all experimental standard error values are less than 16%, and all correlation coefficients are > 0.83. Selectivity values are ratios of respective K_i values. NT = not tested; ND = not determined.

The pyridine of compound 3 was reported to H-bond with His342 (discovered via X-ray crystallographic analysis),¹³ improving this scaffold's hnNOS potency. To this end, we prepared several series of compounds with hydrophilic groups that could act as H-bond acceptors: homologated N,N-dimethylamino derivatives (12-15), pyridine derivatives (16-19) and nitriles 20-21. Finally, structure-based modifications were performed to improve the hn/he selectivity of nitrile 21. The o-methyl group (of 22) and o-chlorine (of 23) were installed on the benzonitrile ring to make additional contact with Met336 (in rnNOS; Met341 in hnNOS). This residue is absent in eNOS isoforms (both bovine and human eNOS), replaced by a smaller valine, and contact between methionine residues (vs. valine residues) and inhibitors has previously been implicated in improved n/e selectivity.^14,15,22 Finally, two active nitrile-containing molecules (22 and 23) were methylated at the 4-position of the quinoline. This modification was previously reported to improve potency (and sometimes n/e selectivity) for the analogous 2-aminopyridines by the interaction of the 4-methyl group with a small hydrophobic pocket located near the “back wall” of the heme-binding pocket.²⁴

In addition to possibly improving hnNOS activity, it was hypothesized that the addition of more hydrophilic groups (such as amines, pyridines, or nitriles) could reduce off-target binding. For example, the addition of a nitrile (relative to an aromatic chloride, as in 5) slightly increases the PSA and decreases the overall hydrophobicity of the molecule,²⁵ which may a) disfavor binding to GPCRs and other CNS targets, many of which depend on extensive hydrophobic interactions for ligand binding, and b) decrease non-specific hydrophobic interactions with other proteins. As very high PSA (>100 Ų) could hamper brain permeation or oral bioavailability, however, we were careful to keep the PSA of all analogues below 80 Ų.

All synthesized compounds were assayed against rnNOS and hnNOS to determine their selectivity for the human isoform (Table 1). To determine selectivity, murine iNOS and bovine eNOS were used, and select compounds were also assayed against human eNOS, in an attempt to begin translating our work to more “humanized” systems (Table 2). Finally, the most selective compound (25) was assayed in a Caco-2 assay to approximate cellular permeability and in the PDSP counter-screen to determine the effects of the new modifications on off-target binding.

Table 2. Inhibition of hnNOS and heNOS by selected compounds^a.

Compound	Ki (μM) a		Selectivity (hn/he)
	Human nNOS	Human eNOS
4	0.318	9.49	30
5	0.440	11.8	27
6	0.295	7.41	25
18	0.073	1.58	22
21	0.032	1.03	32
22	0.020	2.08	104
23	0.021	2.70	129
24	0.052	5.79	111
25	0.030	5.76	192

^aCompounds 18 and 21-25 were assayed in vitro against human eNOS, using known literature methods and K_i values are calculated directly from IC₅₀ values using the Cheng-Prusoff equation (see Experimental Section for details). IC₅₀ values are the average of at least two replicates from 7-9 data points; all experimental standard error values are less than 15%, and all correlation coefficients are > 0.94. Selectivity values are ratios of respective K_i values; human nNOS K_i values are from Table 2.

Chemistry

Preparation of 7-substituted 2-aminoquinolines 7-13 proceeded through bromide 26, which was prepared according to literature procedures.^14,15 To prepare truncated analogues 7-11 (Scheme 1) bromide 26 was treated with an excess of commercially available anilines 27-31 and catalytic potassium iodide under microwave conditions²⁶ to afford the acetamide-protected intermediates 32-36 in good yields. These intermediates were not characterized extensively, but were immediately deacetylated with K₂CO₃ in refluxing methanol²⁷ followed by treatment with methanolic HCl to yield desired analogues 7-11 as hydrochloride salts. The homologated compounds (12 and 13) were prepared by treatment of 26 with commercially available benzylamines 37 or 38 in the presence of Cs₂CO₃ (Scheme 2).²⁸ The desired products were Boc-protected at the secondary amine to aid in purification, providing protected intermediates 39 (from 37) and 40 (from 38). Subsequent deacetylation, workup, and cleavage of the Boc group under acidic conditions afforded 12 and 13, respectively.

Scheme 1.

Reagents and conditions: (a) cat. Kl, MeCN, 110 °C, μ-wave; (b) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

Scheme 2.

Reagents and conditions: (a) i. Cs₂CO₃, DMF, r.t., ii. Boc₂O, THF, r.t.; (b) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

To synthesize aminoquinolines possessing more than one methylene unit between the aryl ring and the secondary amine (analogues 14-16, 20, and 21), the requisite primary amines 44, 47, 50, and 53 were first prepared. To prepare phenethylamine 44, nitroarene 41 (Scheme 3A) was first hydrogenated to yield aniline 42²⁹, followed by Eschweiler-Clarke methylation and hydrogenation of 43 to afford 44. To prepare isomeric amine 47 (Scheme 3B), commercially available aldehyde 45 was condensed with nitromethane to yield unstable nitrostyrene 46³⁰, which was immediately reduced to 47 with LiAlH₄.³¹ Pyridinepropanamine 50 (Scheme 3C) was prepared by the method of Mukherjee et al.,¹³ starting with the Mitsunobu conversion of primary alcohol 48 to azide 49, followed by Staudinger reduction to the amine (and acidic hydrolysis to yield dihydrochloride salt 50).

Scheme 3.

Reagents and conditions: (a) H₂, Pd/C, MeOH, r.t.; (b) formalin, HCO₂H, DMF, 0 °C-reflux; (c) H₂, Raney Ni, MeOH/NH3:EtOH (1:1), r.t.; (d) NO₂CH₃, NH₄OAc, AcOH, reflux; (e) LiAIH₄ in THF, THF, 0 °C-reflux; (f) PPh₃, diphenylphosphoryl azide, DIAD, THF, 0 °C-r.t.; (g) i. P(OEt)₃, benzene, reflux, ii. MeOH/HCI, r.t.; (h) Boc₂O, THF, r.t.; (i) i. K₄[Fe(CN)₆], t-BuXPhos Pd G₃, t-BuXPhos, dioxane, KOAc (0.1 M in H₂O), reflux, ii. MeOH/HCI, r.t. (after isolation).

Finally, cyanophenethylamine 53 (Scheme 3D) was prepared by Boc-protection of commercially available 3-bromophenethylamine (51) followed by palladium-catalyzed cyanation of intermediate carbamate 52. Removal of the Boc group yielded hydrochloride 53.³²

From these amines, as well as commercially available 4-cyanophenethylamine (55), analogues 14-16, 20, and 21 were prepared by an indirect reductive amination (Scheme 4) where aldehyde 54 (prepared by literature procedures)¹⁵ was treated with the required primary amine under mildly acidic, dehydrating conditions (cat. AcOH and anhydrous Na₂SO₄), and the resulting imine was reduced with NaBH₄ (and the secondary amine protected with Boc₂O to aid in purification as described above) to yield intermediates 56-60. These were deprotected to yield final compounds 14-16, 20, and 21.

To prepare substituted pyridine compounds 17-19, the necessary pyridinepropanamines were prepared from bromopyridines 61-63 (Scheme 5). Sonogashira coupling with N-Boc-propargylamine yielded alkynes 64-66, which were hydrogenated and deprotected to give dihydrochloride salts 67-69. Subsequent indirect reductive amination between these amines and 54, followed by reduction and protection, gave the intermediate acetamides 70-72. Deacetylation and Boc group removal afforded pyridine derivatives 17-19 as water-soluble trihydrochloride salts.

The substituted cyanophenethylamines (81 and 82) and their respective aminoquinoline derivatives 22 and 23 were prepared as shown in Scheme 6. Starting with commercially available methyl 4-bromo-3-methylbenzoate (73, for 81) and 4-bromo-3-chlorobenzoic acid (74, for 82), reduction with either LiAlH₄ or BH₃-THF, respectively, resulted in primary alcohols 75 and 76.^33,34 Subsequent bromination of the alcohol under Appel conditions and displacement of the bromide with KCN resulted in benzonitrile intermediates 77³⁵ and 78 in good yield. BH₃-THF reduction³⁶ and protection with Boc₂O afforded 79 and 80.

Scheme 6.

Reagents and conditions: (a) (for 73) LiAIH₄ in THF, THF, -15 °C; (b) (for 74) BH₃-THF, THF, 0 °C-r.t.; (c) i. CBr₄, PPh₃, DCM, 0 °C, ii. KCN, Bu₄NBr, CH₂CI₂, H₂O, r.t.; (d) i. BH₃-THF, THF, reflux, ii. Boc₂O, THF, r.t.; (e) (for 78) K₄[Fe(CN)₆], t-BuXPhos Pd G3, t-BuXPhos, dioxane, KOAc (0.1 M in H₂O), 100 °C; (f) (for 79) CuCN, DMF, reflux; (g) MeOH/HCI, r.t. (after isolation); (h) i. 54, AcOH, Na₂SO₄, CHCI₃, r.t., ii. NaBH₄, MeOH, 0 °C-r.t., iii. Boc₂O, THF, r.t.; (i) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

Palladium-catalyzed cyanation of 79 followed by deprotection (vide supra) yielded the hydrochloride salt 81. Likely because of the bulk of the o-chloro substituent, the analogous cyanation was very low-yielding, and 82 was instead prepared by reaction with CuCN in refluxing DMF followed by acidic deprotection (to yield the crude hydrochloride salt). Subsequent indirect reductive amination with 54 and either 81 or 82, followed by NaBH₄ reduction and Boc-protection, yielded protected intermediates 83 and 84.

Deprotection, as described above, afforded 22 and 23 in moderate yield. Synthesis of 4-methylated aminoquinolines³⁷ 24 and 25 (Scheme 7) began with Doebner-Miller condensation of 3-bromoaniline 85 and 3-buten-2-one to yield 86, and oxidation with m-CPBA resulted in N-oxide 87. Deoxygenative amination (and debutylation of the intermediate t-butylaminoquinoline with TFA)³⁸ yielded 88. Acetylation with N-acetylimidazole yielded 89, followed by Pd-catalyzed formylation³⁹ to yield 90. A similar indirect reductive amination between 90 and either 55 or 81, followed by reduction and protection, afforded protected intermediates 91 and 92, which were then deprotected to yield 24 and 25 as the dihydrochloride salts.

Scheme 7.

Reagents and conditions: (a) 3-buten-2-one, FeCI₃ × 6 H₂O, AcOH, 60 °C-140 °C; (b) m-CPBA, CH₂CI₂, r.t.; (c) i. t-BuNH₂, Ts₂O, PhCF₃/CH₂CI₂, 0 °C, ii. TFA, reflux; (d) N-acetylimidazole, THF, reflux; (e) N-formylsaccharin, Et₃SiH, Pd(OAc)₂, dppb, Na₂CO₃, DMF, 75 °C; (f) i. 55 or 81, AcOH, Na₂SO₄, CHCI₃, r.t., ii. NaBH₄, MeOH, 0 °C r.t., iii. Boc₂O, THF, r.t.; (g) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

Results and Discussion

Historically, rat nNOS (rnNOS) and bovine eNOS (beNOS) were the first NOS enzymes to be expressed and crystallized, and to date have provided a majority of the structural data for isoform-selective inhibitor design. Therefore, compounds 7-25 were first assayed against rnNOS, murine macrophage iNOS (miNOS), and beNOS, using the previously described hemoglobin capture assay (see Experimental Section). These isoforms are also used to approximate isoform selectivity for rnNOS, because the amino acids that generally interact with our inhibitors are identical in rat and murine iNOS, as well as in rat and bovine eNOS, and are unlikely to result in significant species differences. Because of our two-fold goal of decreasing the off-target binding and improving the human nNOS (hnNOS) activity of the previous leads, the studied compounds were also assayed against purified hnNOS, and six of the most potent compounds against hnNOS were also assayed against purified human eNOS (heNOS). Table 1 summarizes the apparent K_i values and isoform selectivities for 7-25, and the activity and selectivity for hnNOS and heNOS are summarized in Table 2. Values for 4-6 are included for comparison. We have found human iNOS very difficult to express and purify in our laboratories, so it is not included in this discussion. As such, most of the following discussion will focus on nNOS and eNOS, as achieving high n/e selectivity was a primary goal of this study. Additionally, iNOS isoforms do not easily grow diffraction-quality crystals, the role of iNOS in neurodegeneration is complicated, and n/e selectivity tends to be much harder to achieve than n/i selectivity for 2-aminoquinolines.

Our initial attempts at improving the hnNOS activity of the 2-aminoquinoline scaffold involved truncation of the aminoalkyl linker to remove the clash between His342 in hnNOS and the bulky haloaryl groups of 4 and 5. The truncated derivatives (7-10), however, were significantly less potent than 4 and 5 against rnNOS, with the most potent derivative (7) showing approximately tenfold less activity than 4 and 5 (Table 1).

Initial Structural Analysis of Truncated and Aniline Analogues

The X-ray crystal structure of compound 8, (which was used for crystallographic analysis because of the poor solubility of 7) reveals the reason for the K_i increase for these truncated inhibitors (Figure 3A). While the aminoquinoline portion mimics arginine and binds to Glu592 (rnNOS), there are no interactions with the nNOS-specific hydrophobic pocket (Leu337, Met336, and Tyr706) that provided a major source of stabilization upon binding of 4 and 5. Instead, the fluorophenyl ring points upward toward the roof of the active site near Arg481, where the fluorine interacts with Gln478. This “bent-up” binding mode was previously observed for aniline-linked aminoquinoline derivatives¹⁵, reflecting the geometry required for the aniline to interact with the nearby heme propionate. The larger methoxyl group of 9 and the N,N-dimethylamine of 10 should be similarly positioned, but owing to the larger size of these substituents (relative to the fluorine in 8), the potency for 9 and 10 is lower.

Figure 3.

Active site structures of rnNOS-8 (A) and hnNOS-8 (B). For clarity, the residues on the roof of the rnNOS active site, Glu478 and Arg481, are not shown. In this figure and all the following structural figures, major H-bonds are depicted with dashed lines. The omit Fo – Fc electron density maps for the bound inhibitor is displayed at 2.5 σ contour level. All structural figures were prepared with PyMol (www.pymol.org).

Not too surprisingly, 8 has similar potency against rnNOS and hnNOS, given that the binding mode is the same (see the hnNOS-8 structure in Figure 3B). The rnNOS/hnNOS selectivity approaches 1:1, and in addition to an identical binding mode to rnNOS, there are no direct interactions with His342 in the hnNOS-8 structure. Despite their low potency, similar rnNOS/hnNOS activity is also observed for 10 and 11, and this, along with the increased hydrophilicity, prompted further investigation of the N,N-dimethylaniline-containing scaffold.

Further modification of 10 and 11 (via homologation) resulted in meta-substituted compound 12 and para-substituted 13. Both compounds displayed improved potency against rnNOS and hnNOS, with similar K_i values between the two nNOS enzymes, although a greater increase in potency was observed for homologation of 10 to 12. The rnNOS-13 and hnNOS-13 structures (Figures 4A and 4B, respectively; the rnNOS-12 and hnNOS-12 structures are shown in SI Figures S1A and S1B, respectively) indicate that homologation of 11 to 13 does result in additional interactions between the inhibitor and the enzyme.

Figure 4.

Active site structures of rnNOS-13 (A) and hnNOS-13 (B). Note that only in hnNOS can the linker amine of 13 make H-bonds with the heme propionates; however, the N,N-dimethylaniline is more disordered in hnNOS than in rnNOS (indicated by density quality).

For example, although the electron density is weak in this area, one N-methyl group of 13 can potentially interact with Leu337 and Met336 in rnNOS (Figure 4A). In hnNOS, the N,N-dimethylamino group prefers an orientation facing away from the polar and bulkier His342 (Figure 4B), instead making contacts with Met341. The linker amine H-bonds with the heme propionates in hnNOS but bends away from the propionates in rnNOS. This difference is likely the result of the different nature of the interactions between the tail N,N-dimethylamino group and either the polar His342 (hnNOS) or non-polar Leu337 (rnNOS). Nonetheless, there are some deleterious interactions that may be present as well. The steric crowding around the linker amino group and repulsion of the heme propionates by the nearby phenyl ring make this ring and its substituent partially disordered as evidenced by the rather poor density in the region (Figure 4B).

Unfortunately, this singly homologated scaffold does not appear to offer significant advantages, potency- or selectivity-wise, over a compound such as 8 (Table 1). Compared to the truncated (8-11) or singly homologated (12 and 13) analogues, the doubly homologated compounds 14 and 15 are significantly improved rnNOS and hnNOS inhibitors. In the rnNOS-15 structure (Figure 5A), the good potency of 15 is visible in the clean electron density throughout the inhibitor except for the N,N-dimethylamino group. The main reasons for increased potency are that the elongated tail relieves the clash and crowding around the linker amino group, allowing the H-bond with heme propionate D, and there are non-polar interactions between the aryl ring and N,N-dimethylamino group and Met336 and Trp306. The hnNOS-15 structure is very much the same, except the aryl ring moves over slightly so that it is farther away from Met341 (> 4.2 Å) while the N,N-dimethylamino group interacts with His342. Compound 15 thus provides another example where the His342/Leu337 difference between hnNOS and rnNOS, respectively, controls differences in affinity. Compound 15 makes better nonpolar contacts with Leu337 and Met 336 in rnNOS than His342 and Met341 in hnNOS, which explains why it is still a better rnNOS inhibitor.

Figure 5.

Active site structures of rnNOS-15 (A), hnNOS-15 (B), and beNOS-15 (C). Important van der Waals contacts are marked with red arrows, and distances are labeled in Å. The Trp residue from the other subunit is labeled with a B in parentheses. The glycerol in beNOS is labeled as Gol.

Compound 15 has the best rn/be selectivity of the N,N-dimethylaniline compounds tested. Like the nNOS structures, the beNOS-15 structure is well ordered (Figure 5C) and retains the H-bonds from the linker amine to both heme propionates. The 44-fold rn/be selectivity is mainly the result of the smaller Val106 in beNOS that replaces Met336 of rnNOS. A glycerol molecule is bound next to Val106 (common in many eNOS structures). The aryl ring moves slightly to avoid the glycerol, and as such, any contacts less than 5.2 Å between the N,N-dimethylamine and the small Val106 cannot be made, although some hydrophobic contacts with Trp76, Leu107, and Val106 are retained with the aryl ring itself. In contrast, the close hydrophobic contacts from both the aryl ring and N,N-dimethylamino group to the nearby Trp306 and bulkier Met336 are more extensive and at shorter distances in rnNOS. Therefore, the rn/be selectivity is likely a true isoform difference (due to Met/Val variation) despite the binding being slightly disturbed by glycerol. Compound 15 also has the highest rn/mi selectivity out of the aniline series. Although structural data are not available for miNOS-15, miNOS has an asparagine (Asn115) in place of Leu337 in rnNOS, and this residue was previously shown to be a crucial determinant of rn/mi selectivity⁴⁰, as it strongly repels hydrophobic moieties such as aryl rings (or possibly 15's aniline methyl groups).

Despite the increased hnNOS activity of these compounds, it is worth noting that there are several drawbacks associated with N,N-dimethylanilines such as 14 and 15. Often flagged as metabolic liabilities in drug discovery (because of potential oxidation into toxic species)⁴¹, anilines can also be air or light sensitive. Although we handled all anilines carefully, storing stocks and solutions in the dark and cold and assaying all compound dilutions within 48 hours of preparation, it was found that solutions of 14 lost their nNOS inhibitory activity when stored for longer periods of time (ca. one week), and a N,N-diethyl analogue of 11 (not reported here) darkened rapidly in solution. Both of these findings indicated possible instability of the aniline and highlighted the need for other histidine-interacting groups.

Pyridine Analogues

As alternative hydrophilic groups, pyridine analogues 16-19 were designed to mimic 3, which was shown by X-ray crystallography to H-bond with His342 and resulted in good hnNOS inhibiton for pyrimidinylimidazole inhibitors.¹³ Compound 16 is not only an excellent rnNOS inhibitor (36 nM), but its hnNOS inhibitory activity is improved fourfold relative to 5. In both the rnNOS-16 (Figure 6A) and hnNOS-16 structures (Figure 6B), the linker amino group H-bonds with heme propionate D, leaving the tail pyridino group near a mainly hydrophobic pocket. In both structures, the tail pyridino group shows flexibility, indicated by weaker electron density. However, the position of the pyridine ring can be identified at lower contour levels, and the crystal structures indicate that the pyridine can have two functions. In the rnNOS-16 structure, the pyridine acts as a hydrophobic group, with its aryl hydrogens making van der Waals contacts with Met336, Leu337, and Trp306 (Figure 6A), while in the hnNOS-16 structure, the pyridine nitrogen forms a 2.8 Å H-bond with His342, as well as hydrophobic contacts with Met341 and Trp311 (Figure 6B). The same binding mode is maintained in beNOS (Figure 6C). As with other inhibitors, the rn/be selectivity (27-fold) may stem from Val106 versus Met336 in rnNOS, as contacts with Val106 in beNOS are less extensive and farther away than those with Met336 in rnNOS.

Figure 6.

Active site structures of rnNOS-16 (A), hnNOS-16 (B), and beNOS-16 (C). Major van der Waals contacts are marked with red arrows. Gol is a glycerol molecule bound next to H₄B in beNOS.

The substituted pyridines (17 and 18) were designed to strengthen the H-bond acceptor potential of the pyridine, whereas 19 was meant to mimic the haloaryl group of 4 and 5. However, neither the rnNOS nor hnNOS activities correlate with the electronics of the pyridine, suggesting other factors are at play, as revealed by crystallography. In the rnNOS-17 structure (SI Figure S2A), the extra methyl group in 17 contacts Met336, Leu337, and Trp306. In the hnNOS-17 structure (SI Figure S2B), the pyridino group cannot adopt the same orientation, owing to potential close contacts with His342, and thus moves closer to the H₄B site. Similarly, the pyridine ring is oriented differently in the hnNOS-18 and rnNOS-18 structures (Figure 7). In the hnNOS-18 structure, the methoxyl group orients away from His342 and interacts with Met341, whose side chain is repositioned to maximize interactions with the methoxypyridino group (Fig. 7A). As 18 is a good hnNOS inhibitor, this is obviously a favorable binding mode. In the rnNOS-18 structure the methoxyl group is oriented in the opposite direction, where it forms favorable contacts with Leu337 (Fig. 7B).

Figure 7.

Active site structures of hnNOS-18 (A) and rnNOS-18 (B). The distance from the tail pyridino group to His342 is more than 5.0 Å. The methoxyl group makes hydrophobic contact (red arrow) with Leu337 in rnNOS.

Inhibitor 19 contains a fluorine, and in the hnNOS-19 structure (Figure S3B), this fluorine points toward Met341, while in rnNOS the fluorine faces the opposite direction, toward Leu337. This further underscores that interactions with Met341 are favored in hnNOS while interactions with Leu337 are favored in rnNOS. As in 16, a H-bond exists (3.0 Å between the pyridine and His342), although it may be a weaker interaction than in the hnNOS-16 structure because of the fluorine. Unfortunately, the n/e selectivity for these compounds remains quite low. Additionally, compound 18 has very similar K_i values between beNOS and heNOS, indicating that pyridines may bind similarly to the eNOS enzymes of these two species. The pyridine series, however, has good n/i selectivity, suggesting that the pyridino group may also behave as a hydrophobic group in iNOS, clashing with Asn115. Compound 19 is an extremely weak iNOS inhibitor that closely resembles 4, an analogue that is a similarly poor binder to iNOS.

Benzonitrile Analogues

We also investigated the use of a nitrile as a potential hydrogen-bond acceptor from His342. Compounds 20 and 21 are both potent dual rnNOS and hnNOS inhibitors. However, the hnNOS-21 (Figure 8A) and rnNOS-21 structures (Figure 8B) are nearly identical. While the electron density for the cyano group is weak, precluding precise positioning, it is clear that the cyano group does not prefer an orientation that enables strong H-bonding with His342. Instead, in both the rnNOS and rhNOS the benzonitrile group points toward and contacts Trp311 (hnNOS) or Trp306 (rnNOS). We do, however, observe electron density consistent with a minor conformer where the benzonitrile faces toward His342 (hnNOS) or Leu337 (rnNOS). Computer modeling of the other conformation confirmed that, because of the length of the linker and the bulkiness of the benzonitrile in 21, even when the cyano group points toward His342 in hnNOS, a H-bond between the two cannot be easily established.

Figure 8.

Active site structures of hnNOS-21 (A), rnNOS-21 (B), and beNOS-21 (C). Close van der Waals contacts are marked with red arrows.

Unfortunately, the rnNOS/beNOS selectivity for 20 and 21 is, like the pyridines, still fairly low. This is mainly because of the tight binding of 21 to beNOS (Table 1). To our surprise, in beNOS the benzonitrile ring actually makes more extensive close contacts with Trp76, Leu107, as well as the smaller Val106 (Figure 8C). These benzonitrile compounds may simply bind to beNOS too strongly to allow for high rn/be selectivity.

Given the high potency of 20 and 21, we decided to try several structure-based modifications to improve the n/e selectivity. The o-methyl group of 22 and o-chlorine of 23 were introduced to provide possible contacts with Met336/Met341 (rnNOS/hnNOS). Interestingly, while this modification does result in a slight improvement in rnNOS and hnNOS potency, it does not improve selectivity for rnNOS over beNOS relative to the unsubstituted benzonitrile compounds (20 and 21), but it does substantially improve selectivity for hnNOS over heNOS. Both 22 and 23 exhibit hn/he selectivity >100-fold because they bind >20-fold weaker to heNOS than beNOS (Table 1). Examination of the crystal structures revealed a potential reason for this disparity.

In the rnNOS-22 structure (Figure 9A), the o-methyl group turns inward toward Leu337, the phenyl ring and cyano group are approximately 4.2 - 4.5 Å from Met336, and the cyano group maintains close contacts with Trp306. In the beNOS-22 structure (Figure 9B), the aryl ring flips over by 180°, where the aryl group contacts (∼ 4.0 Å) Val106 while the cyano group makes similar contacts as in the rnNOS-22 structure. The similarity in these non-bonded contacts might explain why 22 is the strongest beNOS inhibitor (92 nM) in the series, with a rn/be selectivity of only 4 (Table 1). Similar to observations in other structures, in the hnNOS-22 structure (Figure 9C), the benzonitrile has turned inward to form a weak H-bond with His342, which places both the aryl ring and extra o-methyl group 3.9-4.2 Å from Met341. In the heNOS-22 structure (Figure 9D), not only is the His342-benzonitrile interaction missing (as His342 is replaced by Phe105), but there is also no interaction between Val104 and the o-methyl group of the inhibitor, although the aryl ring does make some contact with Val104. The only obvious difference between heNOS and beNOS that could explain why 22 is a much poorer inhibitor of heNOS than of beNOS is the Phe105 (heNOS) vs. Leu107 (beNOS) difference. A similar phenomenon could occur with compound 23 and account for its different potency against beNOS and heNOS (Table 2, SI Figure S4). These results suggest that differences in a single residue among the same isoform from different species, such as Leu337/His342 in rnNOS vs. hnNOS (and Leu107/Phe105 in beNOS vs. heNOS), or even, more importantly, between different isoforms, as Met336/Val104 in hnNOS vs. heNOS, can have significant effects on binding of the same compound.

Figure 9.

Active site structures of rnNOS-22 (A), beNOS-22 (B), hnNOS-22 (C), and heNOS-22 (D). Van der Waals contacts discussed in the text are marked with red arrows. The weak H-bond mentioned in the text for hnNOS has a distance of 3.4 Å in one subunit but is farther in the other.

The Effect of Methylation

Finally, a methyl group was introduced at the 4-position of the aminoquinoline to yield compounds 24 and 25 (methylated versions of 21 and 22, respectively). Previously, this modification was reported to improve potency for 2-aminopyridines²⁴ by providing an extra hydrophobic contact along the back wall of the heme-binding pocket (an area known as the “S-pocket”). Interestingly, possibly because of the larger size and increased interacting surface area of 2-aminoquinoline in comparison of 2-aminopyridine, the 4-methyl has little effect on potency of these 2-aminoquinoline compounds for nNOS despite the extra hydrophobic contact visible in the crystal structures, such as in the hnNOS-25 crystal structure (Figure 10A; for the rnNOS-25 structure see SI Figure S5A).

Figure 10.

Active site structures of hnNOS-25 (A) and heNOS-25 (B). To illustrate the different binding modes of the 4-methylated vs. unmethylated 2-aminoquinolines in heNOS, the structural model of 25 (yellow) is overlaid (C) with that of 22 (cyan). To better visualize the inhibitors, the panel C has been rotated relative to panel B about an axis perpendicular to the figure plane by 180°.

How can the 4-methyl group not improve potency with this extra contact? In the crystal structures of most of these compounds, a stabilizing H-bond is present between the linker amino group and nearby heme propionate.¹³ With the 4-methyl group present, as in the hnNOS-25 structure, the aminoquinoline's orientation is more parallel with the heme than that in an unmethylated analogue such as 22 (Figure 10C), which possibly strains the linker chain and weakens the H-bond between the amine and propionate. At least for these 2-aminoquinoline compounds, any favorable extra contact with the 4-methyl group may be offset by the weakened H-bond between the linker amine and propionate.

It is worth noting that the 4-methyl group increases n/e selectivity for rnNOS and beNOS, and to a more significant degree, hnNOS and heNOS (the hn/he selectivity for 21 and 22 more than tripled and nearly doubled upon methylation in 24 and 25, respectively). This increase in selectivity was also reported for 2-aminopyridines²⁴ although it is not obvious why. For this 2-aminoquinoline scaffold the improved selectivity is mainly the result of poorer eNOS binding upon adding the 4-methyl group (22 vs. 25, Table 1). Comparing the heNOS-25 with the heNOS-22 structure (Figure 10C), indicates that in addition to the weakened H-bonds with the propionates and the extra contact made by the 4-methyl group (vide supra), there is a difference in the orientation of the o-methyl group on the benzonitrile ring, which points toward Phe105 for 25 but away from this residue for 22. When the 4-methyl group is present, the position of the entire inhibitor (as in 25) is more sequestered inside the NOS active site due to the extra contact with the protein and this could introduce steric hindrance and torsional strain in the linker chain or hinder flexibility required to achieve maximal contact. The global effects of all these structural differences lead to poorer binding of the methylated compounds to eNOS, resulting in higher n/e selectivity in general. The trend of increasing selectivity is observed from 21 to 24 as well. The rnNOS-24, beNOS-24, and heNOS-24 structures, and the superimposition of 24 and 21 in beNOS, are shown in SI Figure S6.

Off-Target Profiling and Cellular Permeability Assay

To assess the efficacy of our modifications in reducing off-target binding, the compound with the highest hn/he selectivity (25) was tested in the PDSP (Table 3). In this assay,¹⁹ compounds are tested for binding to 45 different CNS targets and receptors via radioligand displacement. Initially, compounds are tested at a primary high dose (10 μM) and then a secondary K_i determination is performed for compounds showing >50% binding in the primary assay. We have classified off-target binding using the following rubric: concerning (K_i <100 nM, or <∼2× nNOS K_i value), moderate (100-300 nM, or ∼2-5× nNOS K_i value), weak (>300 nM, or >∼5× nNOS K_i value, typically ∼1 μM), and insignificant (<50% at 10 μM). The off-target profile for 5 reveals concerning or moderate binding at 15/45 targets (mostly serotonin and histamine receptors). 22/45 Targets were classified as weak, and 8/45 as insignificant for this compound. Conversely, for rearranged compound 6,¹⁵ the fraction flagged as concerning or moderate decreased, with a concomitant increase in “insignificant” binding. The PDSP results for 25 indicate that the installation of the polar nitrile group is also effective at reducing off-target binding (despite two extra hydrophobic methyl groups on the molecule). Although it is not as effective as 6's structural rearrangement, 25 shows only three concerning targets (the 5-HT1a receptor, alpha1A adrenergic receptor, and dopamine D3 receptor) and less overall binding to serotonin receptors, indicating that the polar group may be successfully interrupting the GPCR-ligand-like the pharmacophore of 5 or decreasing nonspecific hydrophobic binding. Indeed, it is reported that the installation of polar nitriles reduce overall ligand lipophilicity⁴², and by proxy, may reduce promiscuity and negative toxicological observations relative to isosteric aryl chlorides.²⁵

Table 3. PDSP binding summary for selected compounds^a.

Compound	Concerning	Moderate	Weak	Insignificant	Total
5	8	7	22	8	45
6	3	3	17	22	45
25	3	6	22	14	45

^aOff-target binding is classified into four categories: concerning (K_i <100 nM, or <∼2× nNOS K_i value), moderate (100-300 nM, or ∼2-5× nNOS K_i value), weak (>300 nM, or >∼5× nNOS K_i value, typically ∼1 μM), and insignificant (<50% bound at 10 μM), for a total of 45 receptors as assayed by the PDSP's “comprehensive screen” (see reference 17).

Finally, compound 25 was tested for permeability in a Caco-2 assay (Table 4) to estimate its membrane permeability (such as through the intestinal lumen or the blood-brain barrier, although the efficacy of using Caco-2 to approximate passage through the latter is usually less accurate⁴³ due to fundamental differences in the two membranes).

Table 4. Caco-2 permeability data for selected compounds^a.

Compound	Apparent Permeability (Papp, 10-6 cm s-1)b		Efflux ratioc	Recovery
	Mean A-->B	Mean B-->A		A-->B	B-->A
5	30.3	24.5	0.81	98.0%	67.0%
25	20.8	20.1	0.97	58.5 %	83.0 %
Warfarind	23.2	22.5	0.97	-	-
Ranitidinee	0.15	1.2	8.0	-	-
Talinololf	0.066	3.4	51.5	-	-

^aAll assays were performed over 2 h at a concentration of 10 μM. See Experimental Section for details.

^bP_app: apparent permeability rate value.

^cEfflux ratio: P_app (B➔A)/P_app (A➔B).

^dHigh permeability control.

^eLow permeability control.

^fHigh efflux control.

Compound 5 is extremely permeable in this assay, has high compound recovery values, and a low efflux ratio [ratio of membrane permeability (A➔B) to efflux (B➔A), < 3 is considered favorable]. Pleasingly, despite the hydrophilic modification and increase in total polar surface area (from 50.4 Ų for 5 to 74.2 Ų for 25), compound 25 maintains very good permeability, good recovery, and low efflux (indicating that it is unlikely to be a significant substrate for P-gp or other drug efflux transporters).

Conclusions

In summary, we sought to optimize our first generation of potent and selective 2-aminoquinolines (4 and 5) to both (a) improve their binding to hnNOS and selectivity over heNOS, and (b) reduce off-target binding in CNS counter-screening assays by disrupting the GPCR-ligand-like pharmacophore with a hydrophilic group. Although truncation of the scaffold to remove repulsion between the haloaryl moiety and the hnNOS-specific residue His342 did not result in improved potency or selectivity against rnNOS or hnNOS, N,N-dimethylaniline containing inhibitors had very similar activities against rat and human nNOS. Singly homologating the truncated inhibitors improved this activity slightly, and double homologation significantly improved human nNOS potency. Crystal structures indicated that the N,N-dimethylaniline could form a weak H-bond with His342. Further improvements in potency were achieved by introduction of a 3-pyridine-containing tail, which could act as both a hydrophobic group (in rnNOS) and H-bond acceptor for His342 (hnNOS), although alternative binding modes were also observed in the crystal structures. The most potent dual rat/human nNOS inhibition (with selectivity of approximately 1:1 between the nNOS enzymes of these two species) was achieved by installation of a nitrile, and selectivity for hnNOS over heNOS was improved to >100-fold by placement of a hydrophobic substituent ortho- to the nitrile to interact with Met341 (which is replaced by a smaller valine in heNOS). Interestingly, this modification did not improve selectivity for rnNOS over beNOS. Finally, we found that introduction of a methyl group at position 4 of the aminoquinoline of our most potent benzonitrile-containing leads, while not able to improve potency much toward nNOS, was able to enhance n/e selectivity (both rnNOS/beNOS and hnNOS/heNOS). One compound, the doubly methylated benzonitrile (25), had nearly 200-fold hnNOS/heNOS selectivity, although, based on the crystal structures, interpretation of this observation is not that straightforward. Compound 25, when assayed in the PDSP screen, displayed a marked reduction in off-target binding (relative to 5) without compromising the excellent Caco-2 permeability of the first-generation compounds, indicating that cyanated tail moieties of this type may be useful for designing new bioavailable dual rnNOS/hnNOS inhibitors with improved safety profiles.

Experimental Section

General Procedures

Anhydrous solvents (MeOH, DMF, CH₂Cl₂, and THF) were distilled prior to use. All remaining solvents and reagents were purchased from commercial vendors and were used without further purification. Methanolic HCl (3 M, for Boc-deprotection and ammonium salt formation), was prepared fresh by the reaction of acetyl chloride and anhydrous MeOH at 0 °C. Air- or moisture-sensitive reactions were run under an atmosphere of dry argon. An Agilent 971-FP automated flash purification system with SiliaSep (Silicycle, 40-63 μm, 60 Å, 12-80 g) pre-packed silica cartridges was used in purification of all compounds that required it. Melting points were determined in capillary tubes using a Buchi Melting Point B-540 apparatus and are uncorrected. ¹H- and ¹³C-NMR spectra were determined using a Bruker Avance-III 500 (direct cryoprobe) instrument at 500 and 126 MHz, respectively. Low-resolution ESIMS was performed using a Bruker AmazonSL Ion Trap mass spectrometer. High-resolution mass spectral data were obtained at the Integrated Molecular Structure Education and Research Center (Northwestern University) on an Agilent 6210A TOF mass spectrometer in positive ion mode using ESI, with an Agilent G1312A HPLC pump and an Agilent G1367B autoinjector. Data were processed using MassHunter software version B.02.00. Analytical HPLC was performed using an Agilent Infinity 1260 HPLC system with an injection volume of 10 μL. A Phenomenex Luna 5 μm C-8(2) 100 Å column, 50 × 4.60 mm, was used for all analytical HPLC experiments. The purity of all final target compounds was found to be ≥95% by HPLC, using a 10-minute gradient of 95% H₂O/5% acetonitrile + 0.05% TFA to 95% acetonitrile/5% H₂O + 0.05% TFA, at 1.5 mL/min. Preparative HPLC was performed at the Center for Molecular Innovation and Drug Discovery ChemCore laboratory (Northwestern University), using an Agilent 1200 Series HPLC, an Agilent 6120 Quadrupole Mass Spectrometer (API-MS mode), and a Phenomenex Luna 5 μ m C-8(2) 100 Å preparative HPLC column, 150 × 21.2 mm (gradients used are described under subheadings of individual compounds). Microwave chemistry was performed using a Biotage Initiator Sixty research microwave, in Biotage vials. Analytical thin-layer chromatography was performed on Silicycle extra-hard 250 μm TLC plates. Intermediates and final analogues were visualized with short-wavelength UV light, KMnO₄ and ninhydrin stain, where relevant. Compounds 26,¹⁵ 42,²⁹ 46-50,^13,30,31 52,³⁷ 54,¹⁵ 75-77,^33-35 and 85-90,^37,20 were prepared by known literature procedures and their spectral data are consistent with those data reported for the same; their preparations are not reported here. The preparations of 43, 44, 53, 64-69 and 78-82 are described in the Supporting Information.

General Procedure 1: Synthesis and Deprotection of Truncated Aniline-Linked 2-Aminoquinolines²⁶

Step 1. A 5 mL sealable microwave vial was charged with intermediate 26 (1 eq.) and KI (10-15 mol%), which were then diluted with anhydrous MeCN (0.5-1 mL). A solution of the requisite aniline (3 eq.) in MeCN (1-1.5 mL) was added. The vial was sealed and the reaction stirred under microwave heating at 110 °C for 20-25 min. The resulting solution was diluted with CH₂Cl₂ (∼20-30 mL) and washed with sat. aq. NaHCO₃ (3 × 20 mL). The aqueous phase was extracted with EtOAc (3 × 25 mL) and the combined organics were washed with H₂O (20 mL) and sat. aq. NaCl (20 mL), dried over anhydrous sodium sulfate, and concentrated. The residue was purified by flash column chromatography (SiO₂; the gradient is described below for individual compounds) to yield the desired secondary amine. Step 2. The protected intermediate was immediately diluted with MeOH (9-10 mL) and K₂CO₃ (2 eq.) was added. The mixture was heated at reflux for 2-2.5 h, cooled, and concentrated, and the residue was partitioned between EtOAc (10 mL) and H₂O/sat. aq. NaCl (1:1, 10 mL). The layers were separated, the aqueous phase was extracted with EtOAc (3 × 20 mL), and the organics were combined, washed with sat. aq. NaCl (20 mL), dried over anhydrous sodium sulfate, and concentrated. Step 3. The resulting free 2-aminoquinoline was dissolved in MeOH (10 mL), and treated with methanolic HCl (∼3 M, 1.5 mL). The mixture was stirred at room temperature for 20 min, and ether (15 mL) was added, which afforded the desired compound after filtration, washing with ether, and drying in vacuo.

3-(((2-Aminoquinolin-7-yl)methyl)amino)benzonitrile Dihydrochloride (7)

Prepared from 26 (0.100 g, 0.36 mmol) and 3-aminobenzonitrile (27, 0.126 g, 1.07 mmol), using General Procedure 1, Step 1. After workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 35% EtOAc in CH₂Cl₂, intermediate 32 (0.094 g, 88%) was deprotected with K₂CO₃ (0.082 g, 0.59 mmol), using General Procedure 1, Step 2. Following workup, the free aminoquinoline was converted to the dihydrochloride salt, following General Procedure 1, Step 3, to give the title compound as a white solid (0.093 g, 90% from 32): mp 219-223 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.07 (s, 1 H), 9.14 (br s, 1 H), 8.34 (d, J = 9.3 Hz, 1 H), 8.17 (br s, 1H), 7.90 (d, J = 8.2 Hz, 1 H), 7.61 (br s, 1 H), 7.48 (dd, J = 8.2, 1.1 Hz, 1 H), 7.24 (t, J = 7.9 Hz, 1 H), 7.06 (d, J = 9.3 Hz, 1 H), 6.94-6.89 (m, 3 H), 4.53 (s, 2 H); the anilinium protons are broadened into residual water and appear as a broad hump at 5.78 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.5, 148.9, 145.5, 143.1, 136.1, 130.3, 129.1, 124.3, 120.0, 119.6, 119.5, 117.4, 114.7, 114.2, 113.3, 111.8, 45.8; ESIMS m/z (rel. intensity) 275 (MH⁺, 100); HRMS calcd for C₁₇H₁₅N₄, 275.1296; found, 275.1291.

7-(((3-Fluorophenyl)amino)methyl)quinolin-2-amine Dihydrochloride (8)

Prepared from 26 (0.100 g, 0.36 mmol) and 3-fluoroaniline (28, 0.119 g, 1.07 mmol), using General Procedure 1, Step 1. After workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 35% EtOAc in CH₂Cl₂, intermediate 33 (0.088 g, 84%) was deprotected with K₂CO₃ (0.078 g, 0.57 mmol), using General Procedure 1, Step 2. Following workup, the free aminoquinoline was converted to the dihydrochloride salt, following General Procedure 1, Step 3, to give the title compound as a white solid (0.084 g, 87% from 33): mp 209-212 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.09 (s, 1 H), 9.13 (br s, 1 H), 8.33 (d, J = 9.3 Hz, 1 H), 8.15 (br s, 1 H), 7.88 (d, J = 8.2 Hz, 1 H), 7.61 (s, 1 H), 7.47 (dd, J = 8.2, 0.9 Hz, 1 H), 7.07-7.02 (m, 2 H), 6.41 (dd, J = 8.2, 1.4 Hz, 1 H), 6.33-6.28 (m, 2 H), 4.47 (s, 2 H); the anilinium protons are broadened into residual water and appear as a broad hump at 5.65 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ (164.5 + 162.6, 1 C), 154.5, (150.5 + 150.4, 1 C), 145.9, 143.1, 136.1, (130.53 + 130.45, 1 C), 129.0, 124.3, 120.0, 114.8, 113.3, 108.9, (102.4 + 102.2, 1 C), (98.8 + 98.6, 1 C), 46.2; ESIMS m/z (rel. intensity) 268 (MH⁺, 100); HRMS calcd for C₁₆H₁₅FN₃, 268.1250; found, 268.1244.

7-(((3-Methoxyphenyl)amino)methyl)quinolin-2-amine Dihydrochloride (9)

Prepared from 26 (0.100 g, 0.36 mmol) and 3-methoxyaniline (29, 0.132g, 1.07 mmol), using General Procedure 1, Step 1. After workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 35% EtOAc in CH₂Cl₂, intermediate 34 (0.082 g, 71%) was deprotected with K₂CO₃ (0.070 g, 0.51 mmol), using General Procedure 1, Step 2. Following workup, the free aminoquinoline was converted to the dihydrochloride salt, following General Procedure 1, Step 3, to give the title compound as a cream-colored solid (0.069 g, 77% from 34): mp 237-240 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.00 (s, 1 H), 9.09 (s, 1 H), 8.34 (d, J = 9.3 Hz, 1 H), 8.16 (s, 1 H), 7.88 (d, J = 8.2 Hz, 1H), 7.63 (s, 1 H), 7.49 (dd, J = 8.2, 1.2 Hz, 1 H), 7.05 (d, J = 9.3 Hz, 1 H), 6.96 (t, J = 8.0 Hz, 1 H), 6.21-6.14 (m, 3 H), 4.46 (s, 2 H), 3.63 (s, 3 H); the anilinium protons are broadened into residual water and appear as a broad hump at 4.76 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 160.3, 154.2, 149.1, 146.1, 142.9, 135.9, 129.7, 128.8, 124.2, 119.8, 114.7, 113.1, 105.8, 101.9, 98.7, 54.6, 46.5; ESIMS m/z (rel. intensity) 280 (MH⁺, 40); HRMS calcd for C₁₇H₁₈N₃O, 280.1450; found, 280.1447.

N-((2-Aminoquinolin-7-yl)methyl)-N,N-dimethylbenzene-1,3-diamine Trihydrochloride (10)

Prepared from 26 (0.100 g, 0.36 mmol) and N,N-dimethylbenzene-1,3-diamine (30, 0.225g, 1.07 mmol), using General Procedure 1, Step 1. After workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 50% EtOAc in CH₂Cl₂, intermediate 35 (0.057 g, 48%) was deprotected with K₂CO₃ (0.047 g, 0.34 mmol), using General Procedure 1, Step 2. Following workup, the free aminoquinoline was converted to the trihydrochloride salt, following General Procedure 1, Step 3, to give the title compound as a white solid (0.043 g, 63% from 35): mp 216-218 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.16 (s, 1 H), 9.10 (s, 1 H), 8.34 (d, J = 9.0 Hz, 1 H), 8.16 (s, 1 H), 7.89 (d, J = 8.0 Hz, 1 H), 7.65 (s, 1 H), 7.49 (d, J = 8.0 Hz, 1 H), 7.17 (dd, J = 8.0, 8.0 Hz, 1 H), 7.06 (d, J = 9.0 Hz, 1 H), 6.90 (br s, 1 H), 6.81 (br s, 1 H), 6.57 (br s, 1 H), 4.50 (s, 2 H), 3.02 (s, 6 H); the anilinium protons are broadened into residual water and appear as a broad hump at 3.48 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.5, 149.5, 145.4, 143.3, 136.3, 130.9, 129.3, 124.9, 120.4, 115.5, 113.7, 113.0, 108.0, 104.6, 46.8, 45.7; one of the aminoquinoline carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 293 (MH⁺, 100); HRMS calcd for C₁₈H₂₁N₄, 293.1766; found, 293.1759.

N-((2-Aminoquinolin-7-yl)methyl)-N,N-dimethylbenzene-1,4-diamine Trihydrochloride (11)

Prepared from 26 (0.056 g, 0.20 mmol) and N,N-dimethylbenzene-1,4-diamine (31, 0.127g, 0.60 mmol), using General Procedure 1, Step 1. After workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 50% EtOAc in CH₂Cl₂, intermediate 36 (0.052 g, 77%) was deprotected with K₂CO₃ (0.044 g, 0.32 mmol), using General Procedure 1, Step 2. Following workup and purification by flash column chromatography, eluting with a gradient of EtOAc to 15% MeOH in EtOAc, the free aminoquinoline was converted to the trihydrochloride salt, following General Procedure 1, Step 3, to give the title compound as an off-white solid (0.027 g, 41% from 36): mp 242-244 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.02 (s, 1 H), 12.36 (s, 1 H), 9.09 (s, 1 H), 8.33 (d, J = 9.0 Hz, 1 H), 8.13 (s, 1 H), 7.88 (d, J = 8.0 Hz, 1 H), 7.60 (s, 1 H), 7.47 (d, J = 8.0 Hz, 1 H), 7.43 (d, J = 6.0 Hz, 2 H), 7.03 (d, J = 9.0 Hz, 1 H), 6.64 (d, J = 6.0 Hz, 2 H), 4.50 (s, 2 H), 3.01 (s, 6 H); the anilinium protons are broadened into residual water and appear as a broad hump at 4.09 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.9, 149.3, 146.1, 143.3, 136.3, 129.4, 124.6, 122.1, 120.4, 115.1, 113.7, 112.8, 46.3; the methyl carbon signal overlaps with the solvent signal, and one of the aminoquinoline carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 293 (MH⁺, 100); HRMS calcd for C₁₈H₂₁N₄, 293.1766; found, 293.1764.

7-(((3-(Dimethylamino)benzyl)amino)methyl)quinolin-2-amine Trihydrochloride (12)²⁵

3-(Aminomethyl)-N,N-dimethylaniline (37, 0.044 g, 0.29 mmol) and Cs₂CO₃ (0.094 g, 0.29 mmol) were diluted with anhydrous DMF (3 mL) and stirred for 30 min at room temperature. A solution of 26 (0.070 g, 0.25 mmol) in anhydrous DMF (1 mL) was added dropwise over 5 min while stirring and the resulting yellow solution was stirred at room temperature for 16 h and concentrated. The residue was diluted with EtOAc (50 mL), and washed with H₂O (2 × 50 mL) and sat. aq. NaCl (50 mL). The organics were dried with anhydrous sodium sulfate, concentrated, and diluted with anhydrous THF (5 mL). A solution of Boc₂O (0.073 g, 0.34 mmol) in anhydrous THF (2 mL) was added dropwise and the mixture was stirred at room temperature for 18 h. The mixture was concentrated, diluted with CH₂Cl₂ (25 mL), and washed with sat. aq. NaHCO₃ (30 mL), H₂O (30 mL), and sat. aq. NaCl (30 mL). The organic layer was dried over anhydrous sodium sulfate, concentrated, and purified by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 40% EtOAc in CH₂Cl₂ to afford 39 as a clear oil (0.050 g, 45%). This was deprotected using K₂CO₃ (0.031 g, 0.22 mmol), using General Procedure 1, Step 2. After workup, the resulting clear oil was treated with methanolic HCl (3 mL), the mixture was stirred at room temperature for 16 h, and ether (15 mL) was added, affording an off-white solid (0.030 g, 65% from 39) after filtration and washing with ether: mp 294-295 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.61 (s, 1 H), 10.08 (s, 2 H), 9.39 (br s, 1 H), 8.39 (d, J = 9.5 Hz, 1 H), 8.36 (br s, 1 H), 7.97 (d, J = 8.5 Hz, 1 H), 7.87 (s, 1 H), 7.71 (d, J = 8.5 Hz, 1 H), 7.49-7.31 (m, 2 H), 7.18 (d, J = 9.5 Hz, 1 H), 7.11 (br s, 2 H), 4.32 (t, J = 5.0 Hz, 2 H), 4.16 (t, J = 5.0 Hz, 2 H), 2.99 (s, 6 H); the anilinium proton is broadened into residual water and appears as a broad hump at 4.90 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 155.2, 143.1, 137.0, 135.8, 133.4, 130.0, 129.4, 127.2, 121.3, 119.3, 117.2, 115.0, 50.6, 49.8, 42.7; one of the aminoquinoline carbons and two of the aryl carbons are not visible due to baseline broadening; ESIMS m/z (rel. intensity) 307 (MH⁺, 100); HRMS calcd for C₁₉H₂₃N₄, 307.1923; found, 307.1915.

7-(((4-(Dimethylamino)benzyl)amino)methyl)quinolin-2-amine Trihydrochloride (13)²⁵

4-(Aminomethyl)-N,N-dimethylaniline dihydrochloride (38, 0.041 g, 0.41 mmol) and Cs₂CO₃ (0.370 g, 1.13 mmol) were diluted with anhydrous DMF (3.5 mL) and stirred for 30 min at room temperature. A solution of 26 (0.100 g, 0.36 mmol) in anhydrous DMF (1 mL) was added dropwise over 5 min while stirring. The resulting pale-yellow solution was stirred at room temperature for 16 h and concentrated. The residue was diluted with EtOAc (50 mL) and washed with H₂O (2 × 50 mL) and sat. aq. NaCl (50 mL). The organic layer was dried over anhydrous sodium sulfate, concentrated, and diluted with anhydrous THF (5 mL). A solution of Boc₂O (0.103 g, 0.47 mmol) in anhydrous THF (3 mL) was added dropwise and the mixture was stirred at room temperature for 18 h. The mixture was concentrated, diluted with CH₂Cl₂ (25 mL), and washed with sat. aq. NaHCO₃ (30 mL), H₂O (30 mL), and sat. aq. NaCl (30 mL). The organic layer was dried over anhydrous sodium sulfate, concentrated, and purified by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 40% EtOAc in CH₂Cl₂ to afford 40 as a clear oil (0.071 g, 37%). This was deprotected using K₂CO₃ (0.031 g, 0.22 mmol), using General Procedure 1, Step 2. After workup, the resulting clear oil was treated with methanolic HCl (3 mL), the mixture was stirred at room temperature for 16 h, and ether (15 mL) was added, affording a white solid (0.042 g, 64% from 40) after filtration and washing with ether: mp 366-368 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.65 (s, 1 H), 10.01 (s, 2 H), 9.43 (s, 1 H), 8.38 (d, J = 9.0 Hz, 1 H), 8.36 (br s, 1 H), 7.95 (d, J = 8.5 Hz, 1 H), 7.85 (s, 1 H), 7.70 (d, J = 8.5 Hz, 1 H), 7.57 (br s, 2 H), 7.24 (br s, 2 H), 7.18 (d, J = 9.0 Hz, 1 H), 4.28 (t, J = 5.5 Hz, 2 H), 4.13 (t, J = 5.5 Hz, 2 H), 3.00 (s, 6 H); the anilinium proton is broadened into residual water and appears as a broad hump at 4.62 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 155.2, 148.0, 143.1, 137.1, 135.7, 132.0, 129.4, 127.1, 121.3, 119.2, 116.5, 114.9, 50.0, 49.6, 42.9; one of the aminoquinoline carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 307 (MH⁺, 100); HRMS calcd for C₁₉H₂₃N₄, 307.1923; found, 307.1914.

General Procedure 2: Synthesis of 2-Aminoquinolines Containing a Phenethylamine or Phenylpropylamine-Derived Tail Portion

Step 1. The requisite phenethylamine or propylamine (1.1-1.2 eq.) was diluted with anhydrous CHCl₃ (6-9 mL) or CHCl₃/MeOH (10:1-5:1). If the amine was a hydrochloride salt, Et₃N (1.5-2 eq.) was added and the mixture was stirred until clear. Aldehyde 54 or 90 (1 eq., as a solution in minimal CHCl₃) and anhydrous sodium sulfate (∼1 g) were added to the reaction mixture and the resulting suspension was stirred at room temperature for 1 h. Acetic acid (10 μL/10 mg amine) was added and the mixture was stirred at room temperature for 16 h. The resulting solution was filtered and concentrated to give the crude imine, which was diluted with MeOH (4-7 mL) and cooled to 0 °C. NaBH₄ (1.5 eq.) was added while stirring, and the mixture was warmed to room temperature, stirred for 20 min, and concentrated. The residue was diluted with EtOAc (30 mL) and washed with sat. aq. NaHCO₃ (25 mL), H₂O (25 mL), and sat. aq. NaCl (25 mL). The organic layer was dried over anhydrous sodium sulfate and concentrated. Step 2. The crude amine was diluted with anhydrous THF (5-7 mL) and Boc₂O (1.1-1.2 eq., as a solution in minimal anhydrous THF) was added. The mixture was stirred at room temperature for 4-18 h, concentrated, and the residue was purified by flash column chromatography (SiO₂; the gradient is described below for individual compounds) to yield the Boc-protected amine. Step 3. This intermediate was not characterized but was instead deprotected as in General Procedure 1, Step 2. Step 4. After workup, the resulting unprotected aminoquinoline was treated with methanolic HCl (1.5 mL) in ether (5-10 mL), and the mixture was stirred at room temperature for 16 h. Ether (15-20 mL) was then added, which afforded the desired compound after filtration and washing with ether. If one ether wash was insufficient to remove impurities, compounds were precipitated from methanol (1 mL) with ether (15 mL), washed with ether, and dried in vacuo.

3-(Dimethylamino)phenethyl)amino)methyl)quinolin-2-amine Trihydrochloride (14)

Prepared from aldehyde 54 (0.080 g, 0.37 mmol) and phenethylamine 44 (0.073 g, 0.44 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.021 g, 0.55 mmol), and workup, the secondary amine was protected with Boc₂O (0.089 g, 0.41 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 15% EtOAc in CH₂Cl₂, afforded the protected intermediate 56 (0.153 g, 90%). This was immediately deprotected with K₂CO₃ (0.091 g, 0.66 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed, following General Procedure 2, Step 4. An analytically pure sample for assay was prepared by preparative LC-MS, using the instrument and column detailed in the General Procedures section, eluting with a gradient of H₂O + 0.1% formic acid to 10% MeOH + 0.1% formic acid/90% H₂O + 0.1% formic acid. The purified compound was converted to the trihydrochloride salt following General Procedure 1, Step 3, to give the title compound as an off-white solid (0.016 g, 10% from 56): mp 276-278 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.50 (s, 1 H), 9.68 (s, 2 H), 9.31 (br s, 1 H), 8.39 (d, J = 9.0 Hz, 1 H), 8.28 (br s, 1 H), 7.99 (d, J = 8.0 Hz, 1 H), 7.88 (s, 1 H), 7.69 (d, J = 8.0 Hz, 1 H), 7.27 (br s, 1 H), 7.16 (d, J = 9.0 Hz, 1 H), 6.86 (br s, 3 H), 4.36 (t, J = 6.0 Hz, 2 H), 3.22-3.16 (m, 2 H), 3.01-2.90 (m, 8 H); the anilinium proton is broadened into residual water and appears as a broad hump at 4.41 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 155.4, 143.2, 139.0, 137.1, 135.9, 130.2, 129.6, 127.0, 121.4, 119.2, 115.0, 49.9, 48.0, 32.1; four of the aryl carbons are not visible due to baseline broadening and the methyl carbon signal overlaps with the solvent signal; ESIMS m/z (rel. intensity) 321 (MH⁺, 100); HRMS calcd for C₂₀H₂₅N₄, 321.2079; found, 321.2072.

7-(((4-(Dimethylamino)phenethyl)amino)methyl)quinolin-2-amine Trihydrochloride (15)

Prepared from aldehyde 54 (0.080 g, 0.37 mmol) and phenethylamine 47 (0.073 g, 0.44 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.021 g, 0.55 mmol), and workup, the secondary amine was protected with Boc₂O (0.089 g, 0.41 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 15% EtOAc in CH₂Cl₂, afforded the protected intermediate 57 (0.069 g, 40%). This was immediately deprotected with K₂CO₃ (0.041 g, 0.30 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4. An analytically pure sample for assay was prepared by preparative LC-MS, using the instrument and column detailed in the General Procedures section, eluting with a gradient of H₂O + 0.1% formic acid to 20% MeOH + 0.1% formic acid/80% H₂O + 0.1% formic acid. The purified compound was converted to the trihydrochloride salt following General Procedure 1, Step 3, to give the title compound as a white solid (0.008 g, 13% from 57): mp 243-245 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.37 (s, 1 H), 9.48 (s, 2 H), 9.24 (s, 1 H), 8.39 (d, J = 9.0 Hz, 1 H), 8.25 (s, 1 H), 7.99 (d, J = 8.0 Hz, 1 H), 7.85 (s, 1 H), 7.65 (d, J = 8.0 Hz, 1 H), 7.54 (d, J = 9.0 Hz, 1 H), 6.84 (br s, 4 H), 4.35 (t, J = 5.5 Hz, 2 H), 3.12 (br s, 2 H), 2.92 (br s, 8 H); the anilinium proton is broadened into residual water and appears as a broad hump at 3.81 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 155.1, 143.2, 137.0, 135.9, 129.8, 129.6, 127.0, 126.9, 121.4, 119.2, 115.0, 49.9, 48.5, 31.1, 29.0; two of the aminoquinoline carbons are not visible due to baseline broadening; ESIMS m/z (rel. intensity) 321 (MH⁺, 100); HRMS calcd for C₂₀H₂₅N₄, 321.2079; found, 321.2071.

7-(((3-(Pyridin-3-yl)propyl)amino)methyl)quinolin-2-amine Trihydrochloride (16)

Prepared from aldehyde 54 (0.080 g, 0.37 mmol) and phenpropylamine 50 (0.086 g, 0.41 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.021 g, 0.55 mmol), and workup, the secondary amine was protected with Boc₂O (0.088 g, 0.40 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of EtOAc to 5% MeOH in EtOAc, afforded the protected intermediate 58 (0.110 g, 68%). This was immediately deprotected with K₂CO₃ (0.070 g, 0.51 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.075 g, 74% from 58): mp: 260-262.5 °C. ¹H-NMR (500 MHz; DMSO- d₆): δ 14.57 (br s, 1 H), 9.72 (s, 2 H), 9.33 (br s, 1 H), 8.82 (s, 1 H), 8.74 (d, J = 4.7 Hz, 1 H), 8.40-8.34 (m, 3 H), 7.98 (d, J = 8.2 Hz, 1 H), 7.91-7.88 (m, 2 H), 7.71 (dd, J = 8.2, 1.4 Hz, 1 H), 7.16 (d, J = 9.3 Hz, 1 H), 4.31 (t, J = 5.6 Hz, 2 H), 2.98-2.93 (m, 2 H), 2.89 (t, J = 7.6 Hz, 2 H), 2.09 (quintet, J = 7.5 Hz, 2 H); the pyridinium proton is broadened into residual water and appears as a broad hump at 4.52 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.8, 144.0, 143.2, 142.9, 141.6, 139.9, 136.8, 135.5, 129.2, 126.7, 126.4, 121.0, 118.9, 114.7, 49.5, 45.8, 28.8, 26.3; ESIMS m/z (rel. intensity) 293 (MH⁺, 100); HRMS calcd for C₁₈H₂₁N₄, 293.1766; found, 293.1759.

7-(((3-(4-Methylpyridin-3-yl)propyl)amino)methyl)quinolin-2-amine Trihydrochloride (17)

Prepared from aldehyde 54 (0.080 g, 0.37 mmol) and phenpropylamine 67 (0.092 g, 0.41 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.021 g, 0.55 mmol), and workup, the secondary amine was protected with Boc₂O (0.088 g, 0.40 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of EtOAc to 5% MeOH in EtOAc, afforded the protected intermediate 70 (0.118 g, 71%). This was immediately deprotected with K₂CO₃ (0.072 g, 0.53 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.075 g, 69% from 70) after precipitation from hot MeOH (1 mL) using ether (15 mL): mp: 287-290 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.73 (br s, 1 H), 9.81 (s, 2 H), 9.34 (br s, 1 H), 8.74 (s, 1 H), 8.67 (d, J = 5.9 Hz, 1 H), 8.39 (d, J = 9.3 Hz, 1 H), 8.30 (br s, 1 H), 7.98 (d, J = 8.2 Hz, 1 H), 7.88 (s, 1 H), 7.86 (d, J = 5.8 Hz, 1 H), 7.74 (dd, J = 8.2, 1.4 Hz, 1 H), 7.17 (d, J = 9.3 Hz, 1 H), 4.32 (t, J = 5.4 Hz, 2 H), 3.01-2.98 (m, 2 H), 2.89 (t, J = 7.8 Hz, 2 H), 2.55-2.53 (m, 3 H), 2.08-2.02 (m, 2 H); the pyridinium proton is broadened into residual water and appears as a broad hump at 3.65 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.7, 142.7, 140.5, 139.41, 139.40, 138.9, 136.7, 135.4, 129.1, 127.8, 126.6, 120.9, 118.7, 114.5, 49.4, 45.9, 26.5, 24.6, 19.5; ESIMS m/z (rel. intensity) 307 (MH⁺, 100); HRMS calcd for C₁₉H₂₃N₄, 307.1923; found, 307.1923.

7-(((3-(4-Methoxypyridin-3-yl)propyl)amino)methyl)quinolin-2-amine Trihydrochloride (18)

Prepared from aldehyde 54 (0.080 g, 0.37 mmol) and phenpropylamine 68 (0.107 g, 0.45 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.021 g, 0.55 mmol), and workup, the secondary amine was protected with Boc₂O (0.098 g, 0.45 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of EtOAc to 7% MeOH in EtOAc, afforded the protected intermediate 71 (0.133 g, 77%). This was immediately deprotected with K₂CO₃ (0.072 g, 0.53 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.100 g, 81% from 71) after precipitation from hot MeOH (1 mL) using ether (15 mL): mp: 225-228 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 9.63 (br s, 2 H), 8.77 (d, J = 6.7 Hz, 1 H), 8.68 (s, 1 H), 8.39 (d, J = 9.4 Hz, 1 H), 7.98 (d, J = 8.2 Hz, 1 H), 7.87 (s, 1 H), 7.69 (dd, J = 8.2, 1.4 Hz, 1 H), 7.61 (d, J = 6.8 Hz, 1 H), 7.15 (d, J = 9.3 Hz, 1 H), 4.32-4.30 (m, 2 H), 4.11 (s, 3 H), 2.98-2.93 (m, 2 H), 2.76 (t, J = 7.5 Hz, 2 H), 2.04-1.98 (m, 2 H). The pyridinium, quinolinium, and aminoquinoline N-H protons are broadened into residual water and appear as a broad hump at 3.39 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.9, 143.1, 142.8, 141.3, 136.7, 135.7, 129.2, 127.8, 126.7, 121.1, 118.9, 114.7, 109.3, 58.0, 49.6, 46.0, 24.3, 23.9; one of the aryl carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 323 (MH⁺, 100); HRMS calcd for C₁₉H₂₃N₄O, 323.1872; found, 323.1867.

7-(((3-(5-Fluoropyridin-3-yl)propyl)amino)methyl)quinolin-2-amine Trihydrochloride (19)

Prepared from aldehyde 54 (0.065 g, 0.30 mmol) and phenpropylamine 69 (0.082 g, 0.36 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.017 g, 0.45 mmol), and workup, the secondary amine was protected with Boc₂O (0.072 g, 0.33 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with EtOAc, afforded the protected intermediate 72 (0.108 g, 79%). This was immediately deprotected with K₂CO₃ (0.066 g, 0.48 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.039 g, 40% from 72): mp 236-237 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.60 (s, 1 H), 9.80 (s, 2 H), 9.43 (s, 1 H), 8.57 (s, 1 H), 8.46 (s, 1 H), 8.38 (d, J = 9.5 Hz, 1 H), 8.35 (br s, 1 H), 7.96 (d, J = 8.0 Hz, 1 H), 7.87 (s, 1 H), 7.85 (s, 1 H), 7.73 (d, J = 8.0 Hz, 1 H), 7.19 (d, J = 9.5 Hz, 1 H), 4.30 (t, J = 5.5 Hz, 2 H), 2.97-2.87 (m, 2 H), 2.80 (t, J = 7.5 Hz, 2 H), 2.10-2.04 (m, 2 H); the pyridinium proton is broadened into residual water and appears as a broad hump at 5.29 ppm. ¹³C-NMR (126 MHz; DMSO-d₆): δ (160.7 + 158.7, 1 C), 155.2, 144.9, 143.1, 140.0, 137.2, 135.8, (134.8 + 134.6, 1 C), 129.5, 127.0, (125.4 + 125.3, 1 C), 121.3, 119.1, 115.0, 49.8, 46.2, 28.9, 26.7; ESIMS m/z (rel. intensity) 311 (MH⁺, 100); HRMS calcd for C₁₈H₂₀FN₄, 311.1672; found, 311.1669.

3-(2-(((2-Aminoquinolin-7-yl)methyl)amino)ethyl)benzonitrile Dihydrochloride (20)

Prepared from aldehyde 54 (0.065 g, 0.30 mmol) and phenethylamine 53 (0.061 g, 0.33 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.016 g, 0.42 mmol), and workup, the secondary amine was protected with Boc₂O (0.072 g, 0.33 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH₂Cl₂ to 30% EtOAc in CH₂Cl₂, afforded the protected intermediate 59 (0.120 g, 89%). This was immediately deprotected with K₂CO₃ (0.078 g, 0.54 mmol) following General Procedure 2, Step 3. Following workup and purification by flash column chromatography, eluting with a gradient of EtOAc to 5% MeOH in EtOAc, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.076 g, 75% from 59): mp 268-269 °C (softens), 290-293 °C (melts). ¹H-NMR (500 MHz; DMSO- d₆): δ 14.52 (s, 1 H), 9.72 (s, 2 H), 9.31 (br s, 1 H), 8.38 (d, J = 9.3 Hz, 1 H), 8.30 (br s, 1 H), 7.97 (d, J = 8.2 Hz, 1 H), 7.87 (s, 1 H), 7.78 (s, 1 H), 7.74 (dt, J = 7.7, 1.3 Hz, 1 H), 7.68 (dd, J = 8.2, 1.0 Hz, 1 H), 7.64 (d, J = 8.0 Hz, 1 H), 7.56 (t, J = 7.7 Hz, 1 H), 7.15 (d, J = 9.3 Hz, 1 H), 4.34 (s, 2 H), 3.24-3.23 (m, 2 H), 3.11 (t, J = 7.9 Hz, 2 H). ¹³C-NMR (126 MHz; DMSO- d₆): δ 154.7, 142.6, 138.9, 136.4, 133.9, 132.3, 130.7, 129.8, 129.1, 126.5, 120.9, 118.8, 118.7, 114.5, 111.5, 49.5, 47.1, 30.8; one of the quinoline carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 303 (MH⁺, 100); HRMS calcd for C₁₉H₁₉N₄, 303.1610; found, 303.1602.

4-(2-(((2-Aminoquinolin-7-yl)methyl)amino)ethyl)benzonitrile Dihydrochloride (21)

Prepared from aldehyde 54 (0.070 g, 0.33 mmol) and 4-cyano-phenethylamine hydrochloride (55, 0.071 g, 0.39 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.019 g, 0.50 mmol), and workup, the secondary amine was protected with Boc₂O (0.078 g, 0.36 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 10% EtOAc in CH₂Cl₂, afforded the protected intermediate 60 (0.118 g, 79%). This was immediately deprotected with K₂CO₃ (0.072 g, 0.52 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.057 g, 57% from 60): mp 292-294 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.57 (s, 1 H), 9.85 (s, 2 H), 9.36 (s, 1 H), 8.38 (d, J = 9.5 Hz, 1 H), 8.33 (br s, 1 H), 7.97 (d, J = 8.0 Hz, 1 H), 7.87 (s, 1 H), 7.81 (d, J = 8.0 Hz, 2 H), 7.71 (d, J = 9.5 Hz, 1 H), 7.49 (d, J = 8.0 Hz, 2 H), 7.17 (d, J = 8.0 Hz, 1 H), 4.34 (s, 2 H), 3.21 (t, J = 5.0 Hz, 2 H), 3.18-3.12 (m, 2 H). ¹³C-NMR (126 MHz, DMSO-d₆): δ 159.9, 148.5, 147.9, 147.8, 141.7, 140.7, 137.8, 135.0, 134.3, 131.8, 126.1, 124.0, 119.7, 114.9, 54.7, 52.2, 36.6; ESIMS m/z (rel. intensity) 303 (MH⁺, 100); HRMS calcd for C₁₉H₁₉N₄, 303.1610; found, 303.1603.

4-(2-(((2-Aminoquinolin-7-yl)methyl)amino)ethyl)-2-methylbenzonitrile Dihydrochloride (22)

Prepared from aldehyde 54 (0.037 g, 0.17 mmol) and phenethylamine 81 (0.040 g, 0.20 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.010 g, 0.26 mmol), and workup, the secondary amine was protected with Boc₂O (0.041 g, 0.19 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 15% EtOAc in CH₂Cl₂, afforded the protected intermediate 83 (0.043 g, 55%). This was immediately deprotected with K₂CO₃ (0.026 g, 0.19 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.019 g, 54% from 83): mp 316-317 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.43 (s, 1H), 9.65 (s, 2H), 9.24 (s, 1H), 8.38 (d, J = 9.0 Hz, 1H), 8.28 (br s, 1H), 7.98 (d, J = 8.0 Hz, 1H), 7.86 (s, 1H), 7.74 (d, J = 8.0 Hz, 1H), 7.65 (d, J = 8.0 Hz, 1H), 7.37 (s, 1H), 7.29 (d, J = 8.0 Hz, 1H), 7.14 (d, J = 9.0 Hz, 1H), 4.35 (s, 2H), 3.26-3.17 (m, 2H), 3.12-3.08 (m, 2H), 2.47 (s, 3H). ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.6, 152.5, 143.7, 137.7, 135.8, 133.0, 130.3, 126.8, 121.7, 119.6, 119.3, 113.7, 110.2, 49.8, 47.4, 31.9, 19.5; three of the aminoquinoline carbons are not visible due to baseline broadening; ESIMS m/z (rel. intensity) 317 (MH⁺, 100); HRMS calcd for C₂₀H₂₁N₄, 317.1766; found, 317.1759.

4-(2-(((2-Aminoquinolin-7-yl)methyl)amino)ethyl)-2-chlorobenzonitrile Dihydrochloride (23)

Prepared from aldehyde 54 (0.029 g, 0.13 mmol) and phenethylamine 82 (0.035 g, 0.16 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.008 g, 0.20 mmol), and workup, the secondary amine was protected with Boc₂O (0.031 g, 0.14 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of 5% EtOAc in CH2Cl2 to 35% EtOAc in CH₂Cl₂, afforded the protected intermediate 84 (0.046 g, 72%). This was immediately deprotected with K₂CO₃ (0.027 g, 0.19 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a cream-colored solid (0.022 g, 56% from 84): mp 309-311 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.37 (s, 1 H), 9.58 (s, 2 H), 9.24 (br s, 1 H), 8.38 (d, J = 8.0 Hz, 1 H), 8.24 (br s, 1 H), 8.02-7.94 (m, 2 H), 7.85 (s, 1 H), 7.77 (s, 1 H), 7.64 (d, J = 8.0 Hz, 1 H), 7.49 (dd, J = 9.0 Hz, 1.5 Hz, 1 H), 7.14 (d, J = 9.0 Hz, 1 H), 4.39-4.32 (m, 2 H), 3.28 (s, 2 H), 3.17-3.12 (m, 2 H). ¹³C-NMR (126 MHz; DMSO-d₆): δ 155.1, 145.9, 143.2, 135.9, 135.2, 130.8, 129.6, 129.2, 126.9, 121.5, 119.3, 116.5, 115.0, 110.8, 60.1, 47.1, 31.6; two of the aminoquinoline carbons are not visible due to baseline broadening; ESIMS m/z (rel. intensity) 337/339 (MH⁺, 100/35); HRMS calcd for C₁₉H₁₈ClN₄, 337.1220; found, 337.1218.

4-(2-(((2-Amino-4-methylquinolin-7-yl)methyl)amino)ethyl)benzonitrile Dihydrochloride (24)

Prepared from aldehyde 90 (0.060 g, 0.26 mmol) and 4-cyano-phenethylamine hydrochloride (55, 0.058 g, 0.32 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.015 g, 0.39 mmol), and workup, the secondary amine was protected with Boc₂O (0.063 g, 0.29 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 12% EtOAc in CH₂Cl₂, afforded the protected intermediate 91 (0.071 g, 62%). This was immediately deprotected with K₂CO₃ (0.044 g, 0.32 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.029 g, 47% from 91): mp 304-306 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.28 (s, 1 H), 9.73 (s, 2 H), 9.11 (br s, 1 H), 8.18 (br s, 1 H), 8.05 (d, J = 8.0 Hz, 1 H), 7.85 (s, 1 H), 7.82 (d, J = 8.0 Hz, 2 H), 7.70 (d, J = 8.0 Hz, 1 H), 7.49 (d, J = 8.0 Hz, 2 H), 6.98 (s, 1 H), 4.35 (s, 2 H), 3.27-3.20 (m, 2 H), 3.16-3.12 (m, 2 H), 2.64 (s, 3 H). ¹³C-NMR (126 MHz; DMSO-d₆): δ 154.6, 152.5, 143.7, 136.7, 133.0, 130.3, 126.8, 126.4, 121.7, 119.5, 119.3, 113.7, 110.2, 49.8, 47.4, 31.9, 19.5; one of the aminoquinoline carbons is not visible due to baseline broadening; ESIMS m/z (rel. intensity) 317 (MH⁺, 100); HRMS calcd for C₂₀H₂₁N₄, 317.1766; found, 317.1761.

4-(2-(((2-Amino-4-methylquinolin-7-yl)methyl)amino)ethyl)-2-methylbenzonitrile Dihydrochloride (25)

Prepared from aldehyde 90 (0.072 g, 0.315 mmol) and phenethylamine 81 (0.065 g, 0.33 mmol), using General Procedure 2, Step 1. After concentration, reduction with NaBH₄ (0.017 g, 0.45 mmol), and workup, the secondary amine was protected with Boc₂O (0.072 g, 0.33 mmol), following General Procedure 2, Step 2. Workup and purification by flash column chromatography, eluting with a gradient of CH₂Cl₂ to 35% EtOAc in CH₂Cl₂, afforded the protected intermediate 92 (0.141 g, 95%). This was immediately deprotected with K₂CO₃ (0.082 g, 0.60 mmol) following General Procedure 2, Step 3. Following workup, the Boc group was removed using General Procedure 2, Step 4, to give the title compound as a white solid (0.096 g, 80%): mp 300-301 °C. ¹H-NMR (500 MHz; DMSO-d₆): δ 14.15 (s, 1 H), 9.58 (s, 2 H), 9.02 (br s, 1 H), 8.20 (br s, 1 H), 8.05 (d, J = 7.5 Hz, 1 H), 7.82 (s, 1 H), 7.75 (d, J = 8.0 Hz, 1 H), 7.65 (br s, 1 H), 7.37 (s, 1 H), 7.29 (d, J = 8.0 Hz, 1 H), 6.95 (s, 1 H), 4.35 (s, 2 H), 3.26-3.15 (m, 2 H), 3.07 (t, J = 9.0 Hz, 2 H), 2.63 (s, 3 H), 2.47 (s, 3 H). ¹³C-NMR (126 MHz; DMSO-d₆): δ 143.4, 142.3, 133.3, 131.2, 127.5, 126.3, 121.9, 118.4, 113.7, 110.6, 49.9, 47.4, 31.8, 20.4, 19.4; four of the quinoline carbons and two of the aryl carbons are not visible due to baseline broadening; ESIMS m/z (rel. intensity) 331 (MH⁺, 100); HRMS calcd for C₂₁H₂₃N₄, 331.1922; found, 331.1924.

Purified NOS Enzyme Assays. Purified NOS Enzyme Assays

Rat and human nNOS, murine macrophage iNOS, and human and bovine eNOS were recombinant enzymes (expressed in E. coli and purified as reported previously).^44,45,46 The hemoglobin capture assay was used to measure nitric oxide production (to test for NOS inhibition). The assay was performed at 37 °C in HEPES buffer (100 mM with 10% glycerol, pH 7.4) in the presence of 10 μM L-arginine. NADPH (100 μM), CaCl₂ (0.83 mM), calmodulin (approximately 320 units/mL), H₄B (10 μM), and human oxyhemoglobin (3 μM) were also included in the assay mixture. For iNOS, the CaCl₂ and calmodulin were omitted and replaced with HEPES buffer (as neither are required for iNOS function). The assay was performed in 96-well plates using a Synergy 4 BioTek hybrid reader. Each well contained approximately 100 nM enzyme. The dispensing of NOS enzyme and hemoglobin were automated, and after 30 sec (maximum delay), NO production was read by monitoring the absorbance at 401 nm (resulting from conversion of oxyhemoglobin to methemoglobin). Kinetic readouts were performed for 5 min. Each compound was assayed at least in duplicate, and seven to nine concentrations (500 μM-50 nM or 100 μM-10 nM for eNOS and iNOS; 50 μM to 5 nM for rat and human nNOS) were used to construct dose-response curves. IC₅₀ values were calculated by nonlinear regression (variable slope, four parameters, bottom constraint set to 0 and top to 100) using GraphPad Prism software (reported standard error is reported for the LogIC₅₀), and K_i values were obtained from IC₅₀ values, using the Cheng-Prusoff⁴⁷ equation [K_i = IC₅₀/(1+[S]/K_m)] with the following K_m values: 1.3 μM (rat nNOS), 1.6 μM (human nNOS), 8.2 μM (murine macrophage iNOS), 1.7 μM (bovine eNOS) and 3.9 μM (human eNOS).⁴⁸

Inhibitor Complex Crystal Preparation

The sitting drop vapor diffusion methods were used to grow crystals at 4 °C for the heme domains of rat nNOS (8 mg/mL containing 20 mM histidine), the human nNOS K301R/R354A/G357D mutant (10 mg/mL), and human eNOS (7 mg/mL). The crystal growth conditions are as described previously.¹⁸ Fresh crystals were first passed stepwise through cryoprotectant solutions and then soaked with 5–10 mM inhibitor for 3–4 h at 4 °C before being flash cooled with liquid nitrogen and stored until data collection. The presence of an acetate ion near the heme active site in bovine eNOS had caused interference in the binding mode of some phenyl ether-linked aminoquinoline compounds.^20,15 The high concentration of magnesium acetate in the heNOS growth conditions may also introduce an acetate near the active site that may influence the binding mode of inhibitors. To avoid having this acetate in the structure, the magnesium acetate in the cryoprotectant solution was replaced with MgCl₂.

X-ray Diffraction Data Collection, Data Processing, and Structural Refinement

The cryogenic (100 K) X-ray diffraction data were collected remotely at the Stanford Synchrotron Radiation Lightsource (SSRL) or Advanced Light Source (ALS) through the data collection control software Blu-Ice⁴⁹ and a crystal-mounting robot. When a Q315r CCD detector was used, 100–125° of data were typically collected with 0.5° per frame. If a Pilatus pixel array detector was used, 140–160° of fine-sliced data were collected with a 0.2° per frame. Raw CCD data frames were indexed, integrated, and scaled using iMOSFLM⁵⁰, but the pixel array data were processed with XDS⁵¹ and scaled with Aimless.⁵² The binding of inhibitors was detected by initial difference Fourier maps calculated with REFMAC.⁵³ The inhibitor molecules were then modeled in Coot⁵⁴ and refined using REFMAC or PHENIX.⁵⁵ The crystal packing of the MgCl₂-soaked heNOS crystals was changed slightly, resulting in a symmetry change from the orthorhombic P2₁2₁2₁ reported previously¹⁹ to monoclinic P2₁, with a β angle only 0.6–0.7° off compared to the original 90°. Therefore, a molecular replacement calculation with PHASER-MR⁵⁶ was needed to solve the structure. In the P2₁ space group, there are two heNOS dimers in the asymmetric unit. Disordering in portions of inhibitors bound in the NOS active sites was often observed, sometimes resulting in poor density quality. However, partial structural features were usually still visible if the contour level of the sigmaA weighted 2m|Fo| – D|Fc| map was dropped to 0.5 σ, which afforded the building of reasonable models into the disordered regions. Water molecules were added in PHENIX and checked by Coot. The TLS⁵⁷ protocol was implemented in the final stage of refinements with each subunit as one TLS group. The omit Fo – Fc density maps were calculated by removing inhibitor coordinates from the input PDB file before running one more round of TLS refinement in PHENIX (simulated annealing protocol with a 2000 K initial temperature). The resulting map coefficients DELFWT and PHDELWT were used to generate maps. The refined structures were validated in Coot before deposition in the Protein Data Bank.

Caco-2 Permeability Assay

Caco-2 monolayer assays were performed by Cyprotex US, LLC (Watertown, MA), using standard procedures, as previously reported for aminoquinolines.¹⁵

Scheme 4.

Reagents and conditions: (a) i. Et₃N first, or AcOH, Na₂SO₄, CHCI₃, r.t., ii. NaBH₄, MeOH, 0 °C-r.t., iii. Boc₂O, THF, r.t.; (b) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

Scheme 5.

Reagents and conditions: (a) N-Boc-propargylamine, Cul, PPh₃, Pd(PPh₃)₂CI₂, Et₃N, 90 °C; (b) i. H₂, Pd/C, MeOH, r.t., ii. MeOH/HCI, r.t. (after isolation); (c) i. AcOH, Na₂SO₄, CHCI₃, r.t., ii. NaBH₄, MeOH, 0 °C-r.t., iii. Boc₂O, THF, r.t.; (d) i. K₂CO₃, MeOH, reflux, ii. MeOH/HCI, r.t. (after isolation).

Abbreviations

NO

nitric oxide

nNOS

neuronal nitric oxide synthase

eNOS

endothelial nitric oxide synthase

iNOS

inducible nitric oxide synthase

rnNOS

rat nNOS

hnNOS

human nNOS

beNOS

bovine eNOS

heNOS

human eNOS

miNOS

murine macrophage iNOS

P_app

apparent permeability

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

PDSP

psychoactive drug screening program

FMN

flavin mononucleotide

H₄B

(6R)-5,6,7,8-tetrahydrobiopterin

PSA

polar surface area