New Insight into the Structure-Activity Relationships of the Selective Excitatory Amino Acid Transporter Subtype 1 (EAAT1) Inhibitors UCPH-101 and UCPH-102

DOI: 10.1002/cmdc.201500525

Full Papers

New Insight into the Structure–Activity Relationships of

the Selective Excitatory Amino Acid Transporter Subtype 1

(EAAT1) Inhibitors UCPH-101 and UCPH-102

Stinne W. Hansen, Mette N. Erichsen, Tri H. V. Huynh, Josep A. Ruiz, Isabell Haym,

Walden E. Bjørn-Yoshimoto, Bjarke Abrahamsen, Jeanette Hansen, Morten Storgaard,

Anette L. Eriksen, Anders A. Jensen, and Lennart Bunch*^[a]

In the present study, we made further investigations on the

structure–activity requirements of the selective excitatory

amino acid transporter 1 (EAAT1) inhibitor, 2-amino-4-(4-me-

thoxyphenyl)-7-(naphthalen-1-yl)-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (UCPH-101), by exploring 15 different

substituents (R¹) at the 7-position in combination with eight

different substituents (R²) at the 4-position. Among the 63 new

analogues synthesized, we identified a number of compounds

that unexpectedly displayed inhibitory activities at EAAT1 in

light of understanding the structure–activity relationship (SAR)

of this inhibitor class extracted from previous studies. More-

over, the nature of the R¹and R²substituents were observed

to contribute to the functional properties of the various ana-

logues in additive and non-additive ways. Finally, separation of

the four stereoisomers of analogue 14g (2-amino-4-([1,1’-bi-

phenyl]-4-yl)-3-cyano-7-isopropyl-5-oxo-5,6,7,8-tetrahydro-4H-

chromene) was carried out, and in agreement with a study of

a related scaffold, the R configuration at C4 was found to be

mandatory for inhibitory activity, while both the C7 diastereo-

mers were found to be active as EAAT1 inhibitors. A study of

the stereochemical stability of the four pure stereoisomers

14g-A–D showed that epimerization takes places at C7 via

a ring-opening, CÀC bond rotation, ring-closing mechanism.

Introduction

The five excitatory amino acid transporter (EAAT) subtypes

EAAT1–5 mediate the synaptic reuptake of the major excitatory

neurotransmitter glutamate in the central nervous system.^[1–4]

In the course of our search for subtype-selective inhibitors,

a series of highly selective EAAT1 inhibitors with the general

formula I (Figure 1) was discovered, with the most potent ana-

logues being UCPH-101 and UCPH-102.^[5–9]UCPH-101 has sub-

sequently been applied as a pharmacological tool in several

studies exploring the physiological functions and therapeutic

potential in EAAT1 and other EAATs.^[10–13]

ality leads to complete loss of EAAT1 inhibitory activity.^[5,9,14]

4) The stereochemical requirement at C4 was shown to be the

R configuration.^[9]5) An in silico study concluded that a C7 sub-

stituent occupies roughly the same region in space regardless

of stereochemistry, which, together with the fact that inhibito-

ry activity is present in both pairs of diastereomers (Figure 1),

led to the conclusion that the absolute stereochemistry at C7

is less important.^[5]6) This class of inhibitors bind to the EAAT1

protein in a reversible manner.^[15]

The present study arose from an unexpected observation

made for the four analogues 2–5b (see Table 3 below), which

were synthesized in an attempt to optimize the pharmacoki-

netic properties of this class of inhibitors. Interestingly, the ob-

served inhibitory properties of these four analogues contrasted

with SAR conclusions previously drawn for this inhibitor class.

This finding prompted us to explore the SAR of this compound

series by a systematic approach. This was realized through an

elaborate matrix SAR study, in which 15 and 8 substituents (R¹

and R², respectively) were varied systematically. Amongst the

74 compounds within this matrix of a theoretical total of 120

compounds that were characterized in the present study, sev-

eral analogues displayed unexpected pharmacological proper-

ties and a pattern of cooperativity between the bulkiness and

identities of the R¹and R²substituents with respect to EAAT1

activity that would not necessarily have emerged from a con-

ventional stepwise SAR exploration.

Six key conclusions from our previously published structure–

activity relationship (SAR) studies on this class of inhibitors are:

1) An aryl or heteroaryl group is required as the R¹substituent.

2) A wide range of substituents is allowed at the 4-position

(R²), but not hydrogen. 3) Any chemical change to the 5-oxo

group, the 3-CN group, the 2-NH₂group, or the 1-oxy function-

[a] Dr. S. W. Hansen, Dr. M. N. Erichsen, Dr. T. H. V. Huynh, Dr. J. A. Ruiz,

Dr. I. Haym, Dr. W. E. Bjørn-Yoshimoto, Dr. B. Abrahamsen, J. Hansen,

Dr. M. Storgaard, A. L. Eriksen, Prof. A. A. Jensen, Prof. L. Bunch

Department of Drug Design and Pharmacology

Faculty of Health and Medical Sciences

University of Copenhagen, 2100 Copenhagen Ø (Denmark)

E-mail: lebu@sund.ku.dk

Supporting information for this article is available on the WWW under

http://dx.doi.org/10.1002/cmdc.201500525: Overview of post-purification

diastereomeric ratios for all analogues; HPLC chromatograms of 14g

1

and the four stereoisomers 14g-A–D; H NMR spectra of 14g-A in

[D₆]ethanol/D₂O (1:1) at t=0, 24, 48, and 120 h.

ChemMedChem 2016, 11, 382 – 402

382

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Scheme 1. General pathways for the synthesis of target compounds

(series 1–15) by either a A) one-pot or B) two-step reaction. See Table 3 for

definitions of R¹and R²groups.

Separation and pharmacological characterization of the four

stereoisomers of 14g

Exhaustive attempts to separate the four stereoisomers of

UCPH-101 and UCPH-102 were carried out, all of which failed

due to insufficient chromatographic separation of stereoiso-

mers or problems of solubility in the respective solvent sys-

tems. Several analogues were investigated, and finally ana-

logue 14g proved suitable both in terms of separation and

solubility. The resolution was performed on a Daicel Chiralpac

IF column to afford the four stereoisomers 14g-A, 14g-B, 14g-

C, and 14g-D in high stereochemical purity (>98%). The four

stereoisomers were subsequently characterized by NMR

(Table 1). From these data we concluded that 14g-A,D and

Table 1. HPLC peak identity, ¹H NMR chemical shift of H4, and inhibitory

potencies at EAAT1 of the four stereoisomers of 14g (14g-A–D).^[a]

Figure 1. Parental skeleton of the EAAT1 inhibitor series (general formula I)

and SAR conclusions based on previous studies, chemical structures of ana-

logues UCPH-101 and UCPH-102 (compound 8b) and their respective two

pharmacologically active diastereomers.

Compd

HPLC ID

d_H4[ppm]^[b]

EAAT1 IC₅₀[mm] (pIC₅₀ÆSEM)

14g-A

14g-B

14g-C

14g-D

Peak 1

Peak 2

Peak 3

Peak 4

4.36

4.38

4.36

>100 (<4.0)

1.6 (5.88Æ0.15)

1.9 (5.76Æ0.09)

Results and Discussion

[a] All four analogues were inactive at EAAT2 and EAAT3 (IC₅₀>100 mm).

[b] Determined in [D₆]ethanol/D₂O (1:1). [c] Values are the meanÆSEM of

n=3 replicates.

Synthesis

All new analogues reported herein were synthesized by a two-

step protocol (series a–d) or a one-pot three-component reac-

tion (series e–h) as previously described for this compound

class (Scheme 1).^[5,14]The 1,3-diketone components were ob-

tained from commercial suppliers (compound series 1–4, 6, 7,

and 12–15), synthesized in due course (16 for use in com-

pound series 5), or synthesized in accordance with published

procedures (compound series 8 and 9–11). In general, the

crude product was a ~1:1 ratio of diastereomers as determined

14g-B,C are pairs of enantiomers. The four stereoisomers were

characterized pharmacologically as inhibitors of EAAT1, and

the results are summarized in Table 1. Whereas 14g-A,B are in-

active, stereoisomers 14g-C,D were found to inhibit EAAT1-

mediated transport with respective IC₅₀values of 1.6 and

1.9 mm.

The stereochemical stability was investigated for diastereo-

mers 14g-A and 14g-C by dissolving the pure stereoisomer in

a 1:1 mixture of ethanol/water with a few drops of dimethyl

sulfoxide to aid solubility at room temperature. The solution

was monitored by HPLC at 3, 24, and 48 h (see Figure S3, Sup-

porting Information). For both 14g-A and 14g-C a time-de-

pendent epimerization was observed to form 14g-B and 14g-

D, respectively. While negligible epimerization was observed

up to 3 h, as much as 18/21% was observed after 48 h. With

respect to the method of pharmacological characterization,

1

by H NMR spectroscopy. However, this ratio would then vary

according to the method of purification: while being retained

by column chromatography, precipitation and/or crystallization

gave diastereomeric ratios up to 1:9 (Table S1, Supporting In-

formation). Analogues 12a and 13a proved unstable in aque-

ous media and were therefore not included in the study.

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

383

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

stereochemical stability up to 3 h at room temperature is con-

sidered sufficient (see Experimental Section below).

concentrations up to 300 mm (Table 3). This finding was quite

surprising, as the corresponding 4-phenyl analogue 2e was

previously found to be a fairly potent EAAT1 inhibitor (IC₅₀=

7.3 mm).^[14]To investigate this observation further, we synthe-

sized the respective 4-phenyl analogues 3e, 4e, and 5e,

which, analogously to 2e, were found to inhibit EAAT1 uptake,

exhibiting IC₅₀values ranging from 3.8 to 24 mm at the trans-

porter (Table 3). The trend was further supported by the fact

that increasing the size of the R²substituent by the introduc-

tion of a para-biphenyl group in these furyl and thiophenyl an-

alogues resulted in even more potent EAAT1 inhibitors (2g,

3g, 4g, and 5g; Table 3, and light-grey boxes, Table 4). To in-

vestigate whether this correlation between the relative sizes of

R¹and R²groups and EAAT1 activity of the analogue is also

present in the inverted situation—that is, smaller R¹groups

versus larger R²groups—we synthesized 13g, which has

a methyl group at the R¹position and a para-biphenyl moiety

at the R²position. According to our previous SAR observations,

this analogue would be predicted to be inactive, but interest-

ingly, 13g potently inhibited EAAT1-mediated uptake (IC₅₀=

1.9 mm; Table 3, and white box in Table 4). Notably, the intro-

duction of groups smaller than the para-biphenyl at the R²po-

sition led to completely inactive analogues (13c–f).

1

To determine the mechanism of epimerization, an H NMR

experiment was performed: stereoisomer 14g-A was dissolved

1

in [D₆]ethanol/D₂O (1:1), and H NMR spectra were collected at

regular time intervals. The formation of 14g-B was clearly evi-

dent with no incorporation of deuterium [at time T with per-

centage of 14g-D in parentheses: t=0 min (0%), 24 h (5%),

48 h (9%), and 120 h (18%)]. These findings, together with the

fact that the stereochemistry at C4 is essential for inhibitory ac-

tivity,^[15]led us to conclude that epimerization takes place at

C7 by a stepwise ring-opening, C4ÀC5 bond rotation, ring-clos-

ing mechanism, as outlined in Table 2. Unfortunately, we were

unable to determine the respective cis–trans relationships of

the four 14g stereoisomers by X-ray crystallography. However,

importantly, the two 7-epimers 14g-C and 14g-D, which are

comprised in each of their respective diastereomer pairs, are

equipotent EAAT1 inhibitors. Therefore, we conclude that the

obtained variation in percent diastereomeric excess due to

method of purification (Table S1, Supporting Information) is

unlikely to influence the pharmacological properties of the dif-

ferent compounds.

We found these observations highly interesting and there-

fore decided to expand this two-dimensional SAR study to

a matrix study of the R¹and R²substituents covering lipophilic

groups of various size, which would neither engage in hydro-

gen bonding nor salt-bridge formation with the transporter

protein. The substituents chosen were: R¹=H (series 12), Me

(series 13), iPr (series 14), iBu (series 15), and 11 different aryl

or heteroaryl groups (series 1–11); and R²=H (series a), Me

(series b), iPr (series c), iBu (series d), and five different aryl

groups (series e–h). The pharmacological activities of 74 ana-

logues (63 new and 11 previously published) at EAAT1 in the

[³H]d-Asp uptake assay are summarized in Table 3 (with

a color-coded representation of the data in Table 4). A number

of interesting observations were made from this matrix study.

Comparison of the pharmacological profile of the four ana-

logues 8b, 8g, 14b, and 14g (Table 3, and dark-grey boxes in

Table 4) shows that if both R¹and R²are either small or large,

inhibitory activity at EAAT1 is completely lost (compounds 14b

and 8g), whereas both small–large size combinations lead to

the most potent analogues of the series (8b and 14g). The

three analogues 12b, 12e, and 12g, which hold no substitu-

ent at the R¹position (R¹=H), but have methyl, phenyl, or

para-biphenyl as their respective R²substituent, are completely

inactive. On the other hand, analogues with small (methyl) or

bulky (isopropyl) aliphatic R¹substituents (series 13 and 14)

displayed everything from complete inactivity to potent inhibi-

tion of EAAT1-mediated transport. In fact, there is a highly pro-

nounced trend between R²substituent size and transporter ac-

tivity (13b–h, 14b–h). A general comparison of the two latter

series revealed that analogues with a methyl group as the R¹

substituent require considerably larger substituents at the R²

position in order to inhibit EAAT1 activity (13b–h) than ana-

logues with isopropyl as the R¹substituent (14b–h). Analogues

with large aromatic R¹substituents generally display potent in-

Table 2. Stereochemical stability of 14g-A,C in EtOH/H₂O/DMSO at RT

(HPLC) and mechanism for epimerization at C7 shown for the 14g-

A$14g-B pair based on ¹H NMR spectroscopic investigations in

[D₆]ethanol/D₂O (1:1).

Compd

HPLC

3 h

24 h

48 h

14g-A

14g-B

14g-C

14g-D

Peak 1

Peak 2

Peak 3

Peak 4

<2% 14g-B

8% 14g-B

18% 14g-B

–

<1% 14g-D

10% 14g-D

21% 14g-D

–

2D SAR matrix study

In our ongoing work to optimize the pharmacokinetic proper-

ties of this series of EAAT1 inhibitors to facilitate their use as

pharmacological tools for in vivo studies, we attempted to de-

crease the lipophilicity of UCPH-102 by substituting the 1-

naphthyl group with a heterocycle. Thus, we designed and

synthesized furyl analogues 2b and 3b and thiophenyl ana-

logues 4b and 5b. Their functional properties were character-

ized at EAAT1 in a conventional [³H]d-Asp uptake assay, none

of the four analogues displayed significant inhibitory activity at

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

384

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Table 3. Inhibitory potencies at EAAT1 in the [³H]d-Asp uptake assay.^[a]

R²

Series

R¹

H

Me

iPr

iBu

Ph

2-Naph

Ph-pPh

Ph-pNaph

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

[b]

12

13

14

15

H

–

>300

[<3.5]

>300

[<3.5]

>300

[<3.5]

–

>300

[<3.5]

>100

[<4.0]

11

–

>100

[<4.0]

1.9

[5.73Æ0.07]

0.64

–

[b]

Me

iPr

iBu

–

>300

[<3.5]

~30

[~4.5]

–

~300

[~3.5]

~100

[~4.0]

–

>100

[<4.0]

1.6

1.3

[5.88Æ0.03]

4.2

–

[4.95Æ0.03]

[5.80Æ0.14]

[6.20Æ0.08]

[5.37Æ0.07]

11

–

3.2

–

[4.96Æ0.14]

[5.48Æ0.07]

6.3

~100

[~4.0]

~30

[~4.5]

1

2

3

>300^[16]

9.2^[16]

3.9^[16]

3.6^[14]

2.1^[14]

[5.19Æ0.04]

>300

[<3.5]

>300

[<3.5]

15

~100

[~4.0]

>100

[<4.0]

1.2

5.3

7.3^[14]

[4.83Æ0.09]

[5.96Æ0.11]

[5.28Æ0.12]

>300

[<3.5]

>300

[<3.5]

24

1.9

–

[4.34Æ0.05]

[5.73Æ0.07]

>300

[<3.5]

>300

[<3.5]

10

2.0

4

5

[5.04Æ0.09]

[5.69Æ0.11]

>300

[<3.5]

~300

[~3.5]

3.8

2.1

[5.42Æ0.06]

[5.75Æ0.18]

21

2.6

5.6

>100

[<4.0]

6

7

–

[4.73Æ0.15]

[5.68Æ0.22]

[6.33Æ0.19]

>300

[<3.5]

>300

[<3.5]

2.3

>30

[<4.5]

[5.71Æ0.17]

13

0.6

1.0

0.93

[6.11Æ0.20]

0.62

[6.21Æ0.12]

>300

[<3.5]

8

9

0.42^[16]

0.87^[13]

~10^[13]

–

[4.91Æ0.1]

[6.82Æ0.08]

[6.00Æ0.13]

12

>300

[<3.5]

–

1.2^[13]

–

[5.00Æ0.23]

>300

[<3.5]

>300

[<3.5]

>300

[<3.5]

10

11

>300

[<3.5]

>300

[<3.5]

>100

[<4.0]

>300^[13]

[a] All analogues were characterized as mixtures of cis/trans diastereomers. IC₅₀values are given in mm, with pIC₅₀ÆSEM values in square brackets (n=3–5).

None of the analogues displayed significant inhibitory activity at EAAT2 and EAAT3 when tested at concentrations up to 30 mm (IC₅₀>30 mm), 100 mm

(IC₅₀>100 mm), or 300 mm (IC₅₀>300 mm) (depending on the solubility of the compounds in the assay buffer). [b] Analogues 12a and 13a proved unstable

in aqueous buffer solution and are therefore not suited for pharmacological characterization.

hibition, except those bearing either H or the largest substitu-

ent (para-biphenyl) at the R²position, for which EAAT1 inhibi-

tory activity is detrimentally decreased (6a,b,e,g, 7a,b,e,g, 8a–

g, and 9a,b,e,g), thus defining the upper and lower limits for

R²substituent size in analogues with bulky aromatic R¹sub-

stituents. This overall trend was also true for the R¹phenyl ana-

logues 1a–h, but the presence of a smaller aromatic substitu-

ent at the R¹position made it possible for the molecule to ac-

commodate the large para-biphenyl group as R²substituent

(1g) and retain EAAT1 inhibitory activity, whereas introduction

of an even larger aromatic group (para-naphthylphenyl, 1h)

dramatically decreased transporter activity (Table 3). The inac-

tivity or weak EAAT1 inhibitory activity displayed by analogues

bearing meta- or para-biphenyl as R¹substituents and H,

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

385

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Table 4. Sorting of R¹groups according to bulk, and color representation of the inhibitory potencies of the analogues at EAAT1.^[a]

[a] *: IC₅₀<1 mm, *: IC₅₀1–5 mm, *: IC₅₀5–25 mm, *: IC₅₀>30, >100, or >300 mm; the color scheme is based on the specific values listed in Table 1.

methyl, phenyl, or para-biphenyl as R²substituents (10a,b,e,g

and 11 a,b,e,g) deviate from this general pattern, which is also

shared by the R¹ortho-biphenyl analogues. This can most

likely be ascribed to a steric clash between these biphenyl

groups with the transporter protein, and it may suggest that

meta- or para-biphenyl groups in these analogues project into

regions of the EAAT1 binding pocket that an ortho-biphenyl R¹

substituent does not.

To address the influence of lipophilicity of the analogues on

the inhibitory potency at EAAT1, cLogP_(o/w)values were calcu-

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

386

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Table 5. Lipophilicity (cLogP_(o/w)) of series 1–15 in numbers and grouped by background color.^[a]

[a] &: ꢀ1, &: 1ꢀ3, &: 3<5, &:ꢁ5.

lated (Tables 5 and 6). Comparison of the lipophilicities of the

analogues with their respective EAAT1 activities did not reveal

a uniform and conventionally assumed correlation. For exam-

ple, analogues of series 14 (R¹=iPr, R²=Me to 2-Naph, 14b–g)

displayed a substantial correlation between lipophilicity and in-

hibitory potency. However, the analogues of series 8 (R²=Me

to 2-Naph, 8b–g) all exhibited high inhibitory potencies at

EAAT1 regardless of their lipophilicities, and analogues from

series 13 (R¹=Me, R²=Me to 2-Naph, 13b–g), characterized by

the same lipophilicity range, were all either inactive or low-po-

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

387

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Table 6. Lipophilicity (cLogP_(o/w)) as background color,^[a]and inhibition at EAAT1 as dot color.^[b]

[a] &: ꢀ1, &: 1ꢀ3, &: 3<5, &:ꢁ5. [b] *: IC₅₀<1 mm, *: IC₅₀1–5 mm, *: IC₅₀5–25 mm, *: IC₅₀>30, >100, or >300 mm. The color schemes are based

on the specific values listed in Table 1.

tency inhibitors. Moreover, whereas the similar logP and IC₅₀

values of analogues 9b and 14g could suggest that lipophilici-

ty plays a major role in EAAT1 inhibitory activity, analogues 8b

and 13 f are examples of inversely correlated lipophilicity and

potency. Overall, this analysis leads to the conclusion that lipo-

philicity alone does not determine the EAAT1 inhibitory activity

of this inhibitor class.

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

388

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

Physical form of inhibitors when bound to EAAT1

chemistry at C7 is not vital for EAAT1 inhibitory activity. More-

over, an NMR spectroscopic study revealed time-dependent

epimerization at C7 via a ring-opening, CÀC bond rotation,

ring-closing mechanism. As discussed above, this raises the

question as to whether these inhibitors bind EAAT1 in their

ring-closed or ring-open forms.

As demonstrated for analogue 14g (Table 2), this class of inhib-

itors seems to equilibrate between closed-ring and open-ring

forms in aqueous solution. Combining the cooperative nature

of the R¹and R²substituents as a key determinant for EAAT1

activity with the previously reported observation that chemical

changes to the 1-oxy, 2-NH₂, 3-CN, and 5-oxo functionalities re-

sults in loss of inhibitory activity, we speculate that the ring-

open form of the inhibitor could play a key role in its inhibition

of EAAT1-mediated glutamate uptake. Because the open-ring

form is far more flexible than the closed-ring form, the corpo-

rative nature of the R¹and R²groups in determining the

EAAT1 inhibitory activity might be rooted in this mechanism.

Furthermore, the stereochemistry at C7 is abolished in the

open form due to the symmetry plane of the diketone ring

(Figure 2), which is in agreement with the observation that the

Experimental Section

Chemistry

All reactions involving dry solvents or sensitive agents were per-

formed under a nitrogen or argon atmosphere and glassware was

dried prior to use. Commercially available chemicals were used

without further purification. THF and CH₂Cl₂were dried using an

SG water solvent purification system. Reactions were monitored by

analytical thin-layer chromatography (TLC, Merck silica gel 60 F₂₅₄

aluminum sheets). Flash chromatography was carried out using

1

Merck silica gel 60A (35–70 mm). H NMR spectra were recorded on

a 300, 400, or 600 MHz Bruker Avance instrument, and ¹³C NMR

spectra on a 75, 100, or 150 MHz Bruker Avance. HPLC was per-

formed using a Dionex UltiMate 3000 pump and photodiode array

detector (l 200 and 210 nm) installed with an XTerra MS C₁₈

column (3.5 mm, 4.6 mm150 mm), using a 5–95% MeCN gradient

in H₂O containing 0.1% TFA.

Preparative HPLC was carried out on an Agilent Prep HPLC system

with an Agilent 1100 series pump, an Agilent 1200 series diode

array, a multiple wavelength detector (G1365B), and an Agilent

Prep HT High-Performance Preparative Cartridge Column (Zorbax,

300SB-C₁₈Prep HT, 21.2250 mm, 7 mm). A Daicel Chiralpac IF

column (10 mm Ø, 25 cm) was used for preparative separation of

the four stereoisomers of 14g (14g-A–D). LC–MS spectra were re-

corded using an Agilent 1200 series solvent-delivery system

equipped with an auto-injector coupled to an Agilent 6400 series

triple quadrupole mass spectrometer equipped with an electro-

spray ionization source. Gradients of 5% aqueous MeCN+0.1%

HCO₂H (solvent A), and 95% aqueous MeCN+0.05% HCO₂H (sol-

vent B) were used. Melting points were measured with an MPA 100

Optimelt automatic melting point system and are uncorrected.

Compounds were dried under high vacuum using a Holm & Halby

Heto LyoPro 6000 freeze drier. The purity of compounds submitted

for pharmacological characterization was determined by elemental

analysis and/or HPLC to be >95%.

Figure 2. Suggested open form of EAAT1 inhibitors (see Table 2 for mecha-

nism of epimerization at C7).

absolute stereochemistry at C7 has little influence on EAAT1 in-

hibitory activity (Table 1). On the other hand, it is also a distinct

possibility that EAAT1 binds the closed form of the inhibitor

and that the highly diverse EAAT1 activities exhibited by the

74 analogues with different R¹/R²combinations thus arise from

their respective capacities to fit into the binding pocket in the

transporter protein. In the absence of solid experimental evi-

dence for either of the two scenarios, we will refrain from

drawing conclusions as to which physical form of the inhibitor

(Figure 2) actually binds to the EAAT1 protein.

2-Amino-7-(furan-2-yl)-5-oxo-5,6,7,8-tetrahydro-4H-chromene-3-

carbonitrile (2a). 5-(Furan-2-yl)cyclohexane-1,3-dione (0.135 g,

0.757 mmol) was suspended in CH₂Cl₂(10 mL). Malononitrile

(2.50 g, 37.8 mmol) was added followed by N-methylmorpholine

(84 mL, 0.757 mmol). Then formaldehyde (37 wt% in H₂O; 57 mL,

0.757 mmol) was added and the mixture was stirred at RT for 4.5 h.

The mixture was diluted with CH₂Cl₂(40 mL) and then quenched

by addition of 1m NaOH (40 mL). The two phases were separated,

and the organic phase was washed further with 1m NaOH (3

40 mL), dried using anhydrous MgSO₄, filtered, and evaporated to

give a yellow solid (0.066 g). The crude product was triturated with

Et₂O to give the title compound (0.043 g, 22%) as an orange solid.

R_f(EtOAc/MeOH/AcOH, 100:10:1) 0.92; ¹H NMR (400 MHz,

[D₆]DMSO): d=7.59 (dd, 1H, J=0.7 Hz, J=1.7 Hz), 6.93 (bs, 2H),

6.40 (dd, 1H, J=1.9 Hz, J=3.2 Hz), 6.20 (ddd, 1H, J=0.9 Hz, J=

0.9 Hz, J=3.2 Hz), 3.62–3.53 (m, 1H), 2.84–2.72 (m, 4H), 2.70 (dd,

1H, J=4.8 Hz, J=16.3 Hz), 2.59 (dd, 1H, J=10.3 Hz, J=16.3 Hz);

¹³C NMR (100 MHz, [D₆]DMSO): d=195.1, 162.6, 159.0, 155.6, 141.9,

120.3, 110.4, 109.8, 105.1, 50.3, 40.3, 31.1, 30.8, 18.7; LC–MS: two

Conclusions

From the present SAR study we conclude that the inhibitory

activity of this class of chromene-based EAAT1 inhibitors is re-

lated to the relative sizes and identities of R¹and R²substitu-

ents in a cooperative manner. We find it truly intriguing that

this finely tuned interplay between the two substituents is

a key determinant in EAAT1 inhibitory activity of this com-

pound series. The four stereoisomers of the potent analogue

14g (designated 14g-A, 14g-B, 14g-C, and 14g-D) were sepa-

rated, and it was unequivocally confirmed that the stereo-

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

389

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

peaks in UV (l=254 nm): Peak 1: [M+H]⁺calcd for C₁₄H₁₃N₂O₃

257.09, found 257.1, Peak 2: [M+H₂O+H]⁺calcd for C₁₄H₁₅N₂O₄

275.10, found 275.2; Anal. calcd for C₁₄H₁₂N₂O₃: C 65.62, H 4.72, N

solid. R_f(EtOAc/MeOH/AcOH, 100:10:1) 0.92; ¹H NMR (400 MHz,

[D₆]DMSO): d=7.41–7.38 (m, 1H), 7.00–6.97 (m, 2H), 6.93 (bs, 2H),

3.81–3.71 (m, 1H), 2.86–2.61 ppm (m, 6H); ¹³C NMR (100 MHz,

[D₆]DMSO): d=195.1, 162.8, 159.0, 146.2, 127.0, 124.1, 124.0, 120.3,

110.0, 50.3, 43.7, 34.3, 33.0, 18.7 ppm; LC–MS conditions: two

peaks in UV (l=254 nm): Peak 1: [M+H]⁺calcd for C₁₄H₁₃N₂O₂S

273.07, found 273.1, Peak 2: [M+H₂O+H]⁺calcd for C₁₄H₁₅N₂O₃S

291.08, found 291.1; Anal. calcd for C₁₄H₁₂N₂O₂S: C 61.75, H 4.44, N

10.93, found:

C 65.53, H 4.75, N 11.02; mp: 172.0–173.68C

(18Cmin^À1, decomp.).

2-Amino-7-(furan-3-yl)-5-oxo-5,6,7,8-tetrahydro-4H-chromene-3-

carbonitrile (3a). 5-(Furan-3-yl)cyclohexane-1,3-dione (0.136 g,

0.757 mmol) was suspended in CH₂Cl₂(10 mL). Malononitrile

(2.50 g, 37.8 mmol) was added followed by N-methylmorpholine

(84 mL, 0.757 mmol). Then formaldehyde (37 wt% in H₂O; 57 mL,

0.757 mmol) was added and the mixture was stirred at RT for 5 h.

The mixture was diluted with CH₂Cl₂(40 mL) and then quenched

by addition of 1m NaOH (40 mL). The phases were separated and

the organic phase was washed further with 1m NaOH (340 mL),

dried using anhydrous MgSO₄, filtered and evaporated to give

a yellow solid (0.065 g). The crude product was triturated with Et₂O

to give the title compound (0.043 g, 13%) as an orange solid. R_f

10.29, found:

(decomp.).

C 61.65, H 4.42, N 10.31; mp: 175.6–176.58C

2-Amino-7-(benzo[b]thiophen-3-yl)-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (6a). 5-(Benzo[b]thiophen-3-yl)cyclohex-

ane-1,3-dione (50 mg, 0.21 mmol) was suspended in CH₂Cl₂

(3.5 mL) at RT. Malononitrile (676 mg, 10.23 mmol) was added fol-

lowed by N-methylmorpholine (23 mL, 0.21 mmol). Then formalde-

hyde (37 wt% in H₂O (15 mL, 0.21 mmol) was added and the reac-

tion mixture was stirred at RT for 18 h. The mixture was diluted

with CH₂Cl₂(40 mL) and then quenched with 1m NaOH (40 mL).

The two phases were separated and the organic phase was

washed further with 1m NaOH (340 mL) and dried over anhy-

drous MgSO₄. After concentration in vacuo the crude product was

purified by column chromatography on silica gel to afford the title

compound as a white solid (11 mg, 34 mmol, 17%): R_f(EtOAc/hep-

tane 3:2) 0.40; ¹H NMR (400 MHz, CDCl₃): d=7.90 (dd, J=7.2,

1.6 Hz, 1H), 7.75 (dd, J=7.2, 1.6 Hz, 1H), 7.45–7.37 (m, 2H), 7.18 (s,

1H), 4.51 (s, 2H), 3.90–3.82 (m, 1H), 3.03 (s, 2H), 2.98–2.71 ppm (m,

4H); ¹³C NMR (100 MHz, CDCl₃): d=195.9, 162.7, 158.0, 140.8,

137.5, 136.7, 124.8, 124.3, 123.3, 121.4, 121.2, 119.1, 111.2, 56.5,

42.6, 33.3, 32.4, 18.6 ppm; LC–MS: [M+H]⁺, calcd for C₁₈H₁₄N₂O₂S

323.38, found 323.2; HPLC: purity (l=254 nm) >95%; mp: 208–

2108C (decomp.).

1

(EtOAc/MeOH/AcOH, 100:10:1) 0.92; H NMR (400 MHz, [D₆]DMSO):

d=7.60 (dd, 1H, J=1.7 Hz, J=1.7 Hz), 7.53 (ddd, 1H, J=0.9 Hz, J=

0.9 Hz, J=1.8 Hz), 6.91 (bs, 2H), 6.54 (dd, 1H, J=0.9 Hz, J=1.7 Hz),

3.37–3.27 (m, 1H), 2.77 (s, 2H), 2.72–2.65 (m, 2H), 2.62 (dd, 1H, J=

4.3 Hz, J=16.3 Hz), 2.51 ppm (dd, 1H, J=11.5 Hz, J=16.3 Hz);

¹³C NMR (100 MHz, [D₆]DMSO): d=195.8, 163.3, 159.0, 143.4, 138.7,

127.1, 120.3, 109.8, 109.7, 50.3, 42.3, 32.7, 28.8, 18.7 ppm; LC–MS:

two peaks in UV (l=254 nm): Peak 1: [M+H]⁺calcd for

C₁₄H₁₃N₂O₃257.09, found 257.1, Peak 2: [M+H₂O+H]⁺calcd for

C₁₄H₁₅N₂O₄275.10, found 275.1; Anal. calcd for C₁₄H₁₂N₂O₃: C 65.62,

H 4.72, N 10.93, found: C 65.58, H 4.75, N 10.90; mp: 171.3–

172.18C (18Cmin^À1, decomp.).

2-Amino-7-(thiophen-3-yl)-5-oxo-5,6,7,8-tetrahydro-4H-chro-

mene-3-carbonitrile (4a). 5-(Thiophen-3-yl)cyclohexane-1,3-dione

(0.056 g, 0.288 mmol) was suspended in CH₂Cl₂(3.8 mL). Malononi-

trile (0.952 g, 14.4 mmol) was added followed by N-methylmorpho-

line (32 mL, 0.288 mmol). Then formaldehyde (37 wt% in H₂O;

22 mL, 0.288 mmol) was added and the mixture was stirred at RT

for 4 h. The mixture was diluted with CH₂Cl₂(40 mL) and then

quenched by addition of 1m NaOH (40 mL). The phases were sepa-

rated and the organic phase was washed further with 1m NaOH

(340 mL), dried using anhydrous MgSO₄, filtered and evaporated

to give the title compound (0.019 g, 24%) as light-yellow solid. R_f

2-Amino-7-(benzo[b]thiophen-2-yl)-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (7a). 5-(Benzo[b]thiophen-2-yl)cyclohex-

ane-1,3-dione (50 mg, 0.21 mmol) was suspended in CH₂Cl₂

(3.5 mL) at RT. Malononitrile (672 mg, 10.3 mmol) was added fol-

lowed by N-methylmorpholine (23 mL, 0.21 mmol). Then formalde-

hyde (37 wt% in H₂O (15 mL, 0.21 mmol) was added and the reac-

tion mixture was stirred at RT for 22 h. The mixture was diluted

with CH₂Cl₂(40 mL) and then quenched with 1m NaOH (40 mL).

The two phases were separated and the organic phase was

washed further with 1m NaOH (340 mL), dried over anhydrous

MgSO₄, filtered and evaporated to give the title compound as

a yellow solid (20 mg, 62 mmol, 29%). ¹H NMR (400 MHz, CDCl₃):

d=7.80 (dd, J=7.2, 1.6 Hz, 1H), 7.72 (dd, J=7.2, 1.6 Hz, 1H), 7.39–

7.31 (m, 2H), 7.11 (s, 1H), 4.49 (s, 2H), 3.83–3.75 (m, 1H), 3.00 (s,

2H), 2.93–2.72 ppm (m, 4H); ¹³C NMR (100 MHz, CDCl₃): d=194.9,

162.0, 157.9, 146.1, 139.5, 138.8, 124.6, 124.5, 123.4, 122.3, 120.5,

119.0, 111.5, 56.6, 43.8, 34.8, 34.7, 18.6 ppm; LC–MS: [M+H]⁺calcd

for C₁₈H₁₄N₂O₂S, 323.38 found 323.21; HPLC: purity (l=254 nm)

>95%; mp: 190–1928C (decomp.).

1

(EtOAc/MeOH/AcOH, 100:10:1) 0.91; H NMR (400 MHz, [D₆]DMSO):

d=7.51 (dd, 1H, J=3.0 Hz, J=5.0 Hz), 7.30 (ddd, 1H, J=1.2 Hz, J=

1.2 Hz, J=3.0 Hz), 7.16 (dd, 1H, J=1.3 Hz, J=5.0 Hz), 6.91 (bs, 2H),

3.55–3.45 (m, 1H), 2.83–2.57 ppm (m, 6H); ¹³C NMR (100 MHz,

[D₆]DMSO): d=195.8, 163.4, 159.0, 143.8, 127.0, 126.4, 120.5, 120.3,

109.8, 50.3, 42.9, 33.2, 33.1, 18.7 ppm; LC–MS: two peaks in UV

(l=254 nm): Peak 1: [M+H]⁺calcd for C₁₄H₁₃N₂O₂S 273.07, found

273.1, Peak 2: [M+H₂O+H]⁺calcd for C₁₄H₁₅N₂O₃S 291.08, found

291.1; Anal. calcd for C₁₄H₁₂N₂O₂S: C 61.75, H 4.44, N 10.29, found:

C 61.65, H 4.45, N 10.26; mp: 180.5–181.98C (decomp.).

2-Amino-5-oxo-7-(thiophen-2-yl)-5,6,7,8-tetrahydro-4H-chro-

mene-3-carbonitrile (5a). 5-(Thiophen-2-yl)cyclohexane-1,3-dione

(16; 0.147 g, 0.757 mmol) was suspended in CH₂Cl₂(10 mL). Malo-

nonitrile (2.50 g, 37.8 mmol) was added followed by N-methylmor-

pholine (84 mL, 0.757 mmol). Then formaldehyde (37 wt% in H₂O;

57 mL, 0.757 mmol) was added and the mixture was stirred at RT

for 4 h. The mixture was diluted with CH₂Cl₂(40 mL) and then

quenched by addition of 1m NaOH (40 mL). The phases were sepa-

rated and the organic phase was washed further with 1m NaOH

(340 mL), dried using anhydrous MgSO₄, filtered and evaporated

to give a yellow solid (0.066 g). The crude product was triturated

with Et₂O to give the title compound (0.044 g, 21%) as yellow

2-Amino-7-(naphthalen-1-yl)-5-oxo-5,6,7,8-tetrahydro-4H-chro-

mene-3-carbonitrile (8a). 5-(naphthalen-1-yl)cyclohexane-1,3-

dione (0.110 g, 0.462 mmol) was suspended in CH₂Cl₂(6 mL). Malo-

nonitrile (1.525 g, 23.0 mmol) was added followed by N-methyl-

morpholine (51 mL, 0.462 mmol). Then formaldehyde (37 wt% in

H₂O; 35 mL, 0.462 mmol) was added and the mixture was stirred at

RT for 5 h. The mixture was diluted with CH₂Cl₂(10 mL) and then

quenched by addition of 1m NaOH (10 mL). The two phases were

separated and the organic phase was washed further with 1m

NaOH (210 mL), dried using anhydrous MgSO₄, filtered and

evaporated to give a yellow solid (0.087 g). The crude product was

recrystallized from absolute EtOH to give the title compound

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

390

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers

(0.033 g, 23%) as a light-yellow solid. R_f(EtOAc/heptane 1:1) 0.32;

¹H NMR (400 MHz, [D₆]DMSO): d=8.24 (d, 1H, J=7.6 Hz), 7.96 (dd,

1H, J=2.1 Hz, J=7.5 Hz), 7.85 (dd, 1H, J=2.5 Hz, J=6.6 Hz), 7.60–

7.48 (m, 4H), 6.93 (bs, 2H), 4.39–4.30 (m, 1H), 2.98 (dd, 1H, J=

10.6 Hz, J=17.0 Hz), 2.88 (dd, 1H, J=12.3 Hz, J=16.2 Hz), 2.86 (s,

2H), 2.72 (dd, 1H, J=4.2 Hz, J=17.0 Hz), 2.63 ppm (dd, 1H, J=

3.7 Hz, 16.2 Hz); ¹³C NMR (100 MHz, [D₆]DMSO): d=196.6, 164.2,

159.6, 139.0, 134.0, 131.1, 129.3, 127.8, 126.9, 126.2, 126.1, 123.7,

123.5, 120.9, 110.1, 50.9, 43.4, 33.9, 33.4, 19.4 ppm; LC–MS: two

peaks in UV (l=254 nm): Peak 1: [M+H]⁺calcd for C₂₀H₁₇N₂O₂

317.13, found 317.1, Peak 2: [M+H₃O]⁺calcd for C₂₀H₁₉N₂O₃

335.14, found 335.1; Anal. calcd for C₂₀H₁₆N₂O₂·0.1C₂H₆O: C 75.59,

H 5.21, N 8.73, found: C 75.52, H 5.13, N 8.67; mp: 194.5–196.98C

(decomp.).

7-([1,1’-Biphenyl]-3-yl)-2-amino-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (11a). To a suspension of dione (100 mg,

0.378 mmol) in CH₂Cl₂(4.9 mL) was added first malononitrile

(1.24 g, 18.9 mmol) followed by N-methylmorpholine (42 mL,

0.38 mmol). To the resulting suspension was finally added formal-

dehyde (37% aqueous solution, 31 mL, 0.38 mmol) and the reaction

mixture was allowed to stir at RT for 16 h. The solvent was then

evaporated and the resulting residue was dissolved in the mini-

mum amount of CH₂Cl₂and purified directly by gradient column

chromatography (0–5% EtOAc in CH₂Cl₂). Further purification was

performed by recrystallization in CH₂Cl₂/MeOH (1:1) giving the title

compound as yellow crystals (41 mg, 32%). R_f(EtOAc/CH₂Cl₂8:92)

0.34; ¹H NMR (400 MHz, [D₆]DMSO): d=7.70–7.60 (m, 4H), 7.45

(ddq, J=2.23, 6.50, 8.71 Hz, 4H), 7.39–7.32 (m, 1H), 6.94 (s, 2H),

3.49 (ddt, J=4.41, 10.65, 12.43 Hz, 1H), 2.90 (ddt, J=2.15, 10.84,

17.06 Hz, 1H), 2.84–2.70 (m, 3H), 2.58 ppm (ddd, J=3.32, 16.53,

29.08 Hz, 2H); ¹³C NMR (101 MHz, [D₆]DMSO): d=195.9, 163.6,

159.1, 142.0, 139.8, 138.7, 128.9, 128.8, 127.6, 127.5, 127.3, 127.0,

126.8, 126.6, 126.6, 120.3, 109.8, 50.4, 43.0, 37.3, 33.5, 18.8 ppm;

HPLC: >96%, t_R=1.22 min, reversed phase, H₂O/MeCN/TFA, gradi-

ent run, 5 mLmin^À1; LC–MS: [M+H]⁺(decomp.); mp: 200.5–

204.38C (decomp.).

7-([1,1’-Biphenyl]-2-yl)-2-amino-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (9a). 5-(1,1’-Biphenyl)cyclohexane-1,3-

dione (0.200 g, 0.757 mmol) was suspended in CH₂Cl₂(10 mL). Ma-

lononitrile (2.498 g, 37.8 mmol) was added followed by N-methyl-

morpholine (84 mL, 0.757 mmol). Then formaldehyde (37 wt% in

H₂O; 57 mL, 0.757 mmol) was added and the mixture was stirred at

RT for 4 h. The mixture was diluted with CH₂Cl₂(40 mL) and then

quenched by addition of 1m NaOH (40 mL). The two phases were

separated and the organic phase was washed further with 1m

NaOH (240 mL), dried using anhydrous MgSO₄, filtered and

evaporated to give a yellow solid (0.157 g). The crude product was

recrystallized from absolute EtOH to give the title compound

(0.061 g, 24%) as a white solid. R_f(EtOAc/heptane, 1:1) 0.42;

¹H NMR (400 MHz, [D₆]DMSO): d=7.61 (d, 1H, J=7.8 Hz), 7.47–7.28

(m, 7H), 7.17 (dd, 1H, J=1.5 Hz, J=7.8 Hz), 6.87 (bs, 2H), 3.48–3.37

(m, 1H), 2.93 (dd, 1H, J=11.5 Hz, J=17.3 Hz), 2.76 (dd, 1H, J=

13.6 Hz, J=16.3 Hz), 2.71 (s, 2H), 2.39–2.29 ppm (m, 2H); ¹³C NMR

(100 MHz, [D₆]DMSO): d=195.6, 163.4, 158.9, 141.3, 140.7, 139.7,

130.1, 128.9, 128.4, 127.9, 127.2, 126.6, 126.4, 120.2, 109.5, 50.4,

43.2, 34.2, 33.4, 18.6 ppm; LC–MS: two peaks in UV spectrum (l=

254 nm): Peak 1: [M+H]⁺calcd for C₂₂H₁₉N₂O₂343.14, found 343.2,

Peak 2: [M+H₂O+H]⁺calcd for C₂₂H₂₁N₂O₃361.16, found 361.2;

Anal. calcd for C₂₂H₁₈N₂O₂: C 77.17, H 5.30, N 8.18, found: C 77.22,

H 5.30, N 8.31; mp: 194.3–197.88C (decomp.).

2-Amino-7-(furan-2-yl)-4-methyl-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (2b). Synthesized in accordance with

procedure published for 9b. Stirring at RT for four days gave the

title compound (0.091 g, 60%) as a yellow solid in a 2:1 mixture of

diastereomers. ¹H NMR (400 MHz, [D₆]DMSO): d=7.58 (dd, 0.7H,

J=0.8 Hz, J=1.8 Hz), 7.56 (d, 0.3H, J=0.8 Hz, J=1.7 Hz), 6.88 (bs,

1.3H), 6.87 (bs, 0.7H), 6.39 (dd, 0.7H, J=1.9 Hz, J=3.2 Hz), 6.38

(dd, 0.3H, J=1.9 Hz, J=3.2 Hz), 6.22–6.19 (m, 0.7H), 6.15–6.13 (m,

0.3H), 3.62–3.43 (m, 1H), 3.14–3.07 (m, 1H), 2.89–2.51 (m, 4H), 1.10

(d, 2H, J=6.5 Hz), 0.99 ppm (d, 1H, J=6.5 Hz); ¹³C NMR (100 MHz,

[D₆]DMSO): d=194.8, 162.4, 161.6, 158.7, 155.6, 141.93, 141.86,

119.8, 115.0, 110.4, 105.2, 105.1, 57.7, 57.6, 40.8, 40.5, 31.1, 31.0,

30.8, 24.7, 22.8, 22.4 ppm; LC–MS: [M+H]⁺calcd for C₁₅H₁₅N₂O₃,

271.10; found 271.2; HPLC: >97%, t_R=1.50 min, reversed phase,

H₂O/MeCN/TFA, gradient run, 5 mLmin^À1

(18Cmin^À1).

; mp: 179.9–180.38C

2-Amino-7-(furan-3-yl)-4-methyl-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (3b). 5-(Furan-3-yl)cyclohexane-1,3-dione

(0.178 g, 1.00 mmol) was dissolved in absolute EtOH (25 mL). The

mixture was cooled to 08C and malononitrile (0.066 g, 1.00 mmol)

was added followed by acetaldehyde (176 mL, 3.00 mmol). After

15 min, N-methylmorpholine (10 mL, 0.09 mmol) was added and

the mixture was stirred at RT for 2.5 h. The mixture was evaporated

to dryness and the residue was recrystallized from absolute EtOH

to give the title compound (0.149 g, 54%) as an off-white crystal-

line solid in a 3:1 mixture of diastereomers. ¹H NMR (300 MHz,

[D₆]DMSO): d=7.60–7.56 (m, 1H), 7.53–7.51 (m, 0.75H), 7.49–7.47

(m, 0.25H), 6.89 (bs, 2H), 6.55–6.52 (m, 0.75H), 6.51–6.49 (m,

0.25H), 3.30–3.18 (m, 1H), 3.10 (q, 1H, J=6.6 Hz), 2.73–2.45 (m,

4H), 1.09 (d, 2.25H, J=6.6 Hz), 1.03 ppm (d, 0.75H, J=6.6 Hz);

¹³C NMR (75 MHz, [D₆]DMSO): d=199.8, 196.1, 163.7, 162.9, 159.3,

144.0, 139.3, 139.2, 127.6, 120.5, 115.6, 115.5, 110.3, 58.4, 58.2, 43.5,

33.6, 33.5, 29.6, 29.4, 25.5, 23.7, 23.2 ppm; LC–MS: [M+H]⁺calcd

for C₁₅H₁₅N₂O₃271.11, found 271.2; Anal. calcd for

C₁₅H₁₄N₂O₃·0.2H₂O: C 66.78, H 5.30, N 10.23, found: C 66.02, H 5.07,

N 9.92; mp: 179.0–181.48C (decomp.).

7-([1,1’-Biphenyl]-3-yl)-2-amino-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (10a). To a suspension of dione (53 mg,

0.20 mmol) in CH₂Cl₂(2.6 mL) was added first malononitrile

(569 mg, 10.0 mmol) followed by N-methylmorpholine (22 mL,

0.20 mmol). To the resulting solution was finally added formalde-

hyde (37% aqueous solution, 17 mL, 0.2 mmol) and the reaction

mixture was allowed to stir at RT for 16 h. The resulting mixture

was diluted with CH₂Cl₂(30 mL) and the reaction was then

quenched with 1m aqueous NaOH (10 mL). The two phases were

separated and the organic phase was further washed with 1m

aqueous NaOH (310 mL), dried over MgSO₄, filtered and evapo-

rated to give a yellow crude. Purification by column chromatogra-

phy (5% EtOAc in CH₂Cl₂) gave the title compound as a pale-

yellow solid (24 mg, 35%). R_f(EtOAc/CH₂Cl₂, 5:95) 0.27; ¹H NMR

(400 MHz, CDCl₃): d=7.57 (d, J=7.24 Hz, 2H), 7.52 (d, J=7.82 Hz,

1H), 7.45 (q, J=7.40 Hz, 4H), 7.37 (t, J=7.31 Hz, 1H), 7.22 (d, J=

7.68 Hz, 1H), 4.51 (s, 2H), 3.54–3.43 (m, 1H), 3.04–2.96 (m, 2H),

2.86–2.65 ppm (m, 4H); ¹³C NMR (101 MHz, CDCl₃): d=196.2, 163.1,

158.2, 142.5, 142.2, 140.9, 129.5, 129.0, 127.7, 127.3, 126.4, 125.7,

125.5, 119.4, 111.3, 56.4, 43.7, 38.8, 34.6, 18.8 ppm; HPLC: >98%,

t_R=1.20 min, reversed phase, H₂O/MeCN/TFA, gradient run,

2-Amino-7-(furan-3-yl)-4-methyl-5-oxo-5,6,7,8-tetrahydro-4H-

chromene-3-carbonitrile (4b). 5-(Furan-3-yl)cyclohexane-1,3-dione

(0.178 g, 1.00 mmol) was dissolved in absolute EtOH (25 mL). The

mixture was cooled to 08C and malononitrile (0.066 g, 1.00 mmol)

5 mLmin^À1

(decomp.).

;

LC–MS: [M+H]⁺(decomp.); mp: 163.3–165.28C

ChemMedChem 2016, 11, 382 – 402

www.chemmedchem.org

391