Scholarly article on Benzaldehyde, 2-chloro-3,4-dimethoxy-5-nitro- 870962-50-8 from Journal of Medicinal Chemistry p. 7560

7560

J. Med. Chem. 2005, 48, 7560-7581

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones as Covalent-Binding, Irreversible

Inhibitors of the Kinase Domain of Vascular Endothelial Growth Factor

Receptor-2

Allan Wissner,* M. Brawner Floyd, Bernard D. Johnson, Heidi Fraser, Charles Ingalls, Thomas Nittoli,

Russell G. Dushin, Carolyn Discafani, Ramaswamy Nilakantan, Joseph Marini, Malini Ravi, Kinwang Cheung,

Xingzhi Tan, Sylvia Musto, Tami Annable, Marshall M. Siegel, and Frank Loganzo

Chemical and Screening Sciences and Oncology Research, Wyeth Research, 401 N. Middletown Road,

Pearl River, New York 10965

Received June 13, 2005

A series of 2-(quinazolin-4-ylamino)-[1,4] benzoquinone derivatives that function as potent

covalent-binding, irreversible inhibitors of the kinase domain of vascular endothelial growth

factor receptor-2 (VEGFR-2) has been prepared by ceric ammonium nitrate oxidation of

substituted (2,5-dimethoxyphenyl)(6,7-disubstituted-quinazolin-4-yl)amines and by displace-

ment of the chlorine atom of substituted 2-chloro-5-(6,7-disubstituted-quinazolin-4-ylamino)-

[1,4]benzoquinones with various amines, anilines, phenols, and alcohols. Enzyme studies were

conducted in the absence and presence of glutathione and plasma. Several of the compounds

inhibit VEGF-stimulated autophosphorylation in intact cells. Kinetic experiments were

performed to study the reactivity of selected inhibitors toward glutathione. Reactivities

correlated with LUMO energies calculated as averages of those of individual conformers

weighted by the Boltzmann distribution. These results and molecular modeling were used to

rationalize the biological observations. The compounds behave as non-ATP-competitive

inhibitors. Unequivocal evidence, from mass spectral studies, indicates that these inhibitors

form a covalent interaction with Cys-1045. One member of this series displays antitumor activity

in an in vivo model.

Angiogenesis is the process of new blood vessel

growth. Several biological effectors mediate angiogen-

esis, including various growth factors and their recep-

tors.¹Vascular endothelial growth factor receptor-2

(VEGFR2), also known as KDR, is a receptor tyrosine

kinase expressed on the endothelial cells that comprise

blood vessels.²Solid tumors require vascularization in

order to grow beyond a small size and to metastasize;

therefore, many tumors express and secrete growth

factors, including VEGF, to promote blood vessel devel-

opment.^3,4Animal studies support a critical role for the

paracrine activation of VEGFR-2 by its ligand, VEGF,

in tumor neovascularization, growth, and metastasis.

In patient tumor samples, VEGF or VEGFR-2 has been

shown to be elevated and negatively correlates with

survival.⁵

The proposed clinical benefit of therapeutics that

inhibit angiogenesis has been widely reviewed. For

example, such therapy may produce less drug resistance

by targeting untransformed endothelial cells rather

than genetically unstable cancer cells and also may be

effective when drug resistance has already developed.^6,7

In addition, slowly growing tumors are often refractory

to standard cytotoxic cancer therapy but responsive to

continuous, low-dose antiangiogenic drugs.⁸Further-

more, by primarily targeting the vasculature rather

than the tumor, expression of the target by the tumor

itself is not required.

In animal models, inhibition of VEGFR-2 and/or

VEGF effectively inhibits the growth of solid tumors

derived from lung, colorectal, breast, prostate, and other

tissues. Neutralizing antibodies to VEGF (e.g., bevaci-

zumab) or to VEGFR-2 (e.g., DC101) inhibit primary

tumor growth and/or the incidence of metastases in

vivo.^9,10Combinations with standard cytotoxic drugs,

such as paclitaxel, can improve efficacy. Ribozymes or

RNA interference (RNAi) constructs that specifically

decrease VEGF or VEGFR-2 levels have demonstrated

activity in cellular, angiogenesis, or animal models.^11-13

Effort has also focused on the design and development

of small molecule inhibitors of VEGFR-2 kinase, several

of which are in clinical development.^14,15Lead molecules

inhibit purified recombinant VEGFR-2 kinase with IC₅₀

<100 nM. Many of these small molecule inhibitors are

also nonselective against related kinases, which may be

of benefit, since some of these kinases have been

implicated in tumor growth and angiogenesis. The

reported VEGFR-2 kinase inhibitors decrease VEGFR-2

phosphorylation in cells and/or inhibit VEGF-stimulated

growth of cultured endothelial cells. Activity is fre-

quently demonstrated in alternative angiogenesis mod-

els. VEGFR-2 kinase inhibitors also slow the growth of

human tumor xenografts in rodents.

Phase I clinical trials of single-agent antibody or small

molecule inhibitors of VEGFR-2 or VEGF have shown

these agents to be generally well tolerated. In phase II/

III trials in patients with solid tumors, stable disease

and partial regressions have been observed with small

* Corresponding author. Tel: (845) 602-3580. Fax: (845) 602-5561.

E-mail: wissnea@wyeth.com.

Published on Web 11/03/2005

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7561

molecule VEGFR-2 inhibitors, either as monotherapy

or in combinations.^16-18In combination with chemo-

therapeutics, the anti-VEGF antibody bevacizumab

produced complete regressions and increased survival

time in colorectal cancer patients¹⁹and increased

response rate in some lung cancers.²⁰These data in

patients help to validate inhibition of the VEGF path-

way as a potentially useful method to control some

forms of cancer.

In earlier studies, using molecular modeling, we had

some success in developing covalent binding inhibitors

of two growth factor receptor kinases, EGFR and HER-

2. Two of these structurally related inhibitors are

currently in clinical trial.^21-24These inhibitors function

as irreversible binding inhibitors by virtue of the fact

that they form a covalent interaction with a conserved

Cys residue (Cys-773 in EGFR and Cys-805 in HER-2)

located near the ATP binding pocket of these enzymes.

With the availability of the X-ray crystal structure of

the catalytic domain of VEGFR-2,²⁵we thought that it

might be possible to design covalent binding inhibitors

of this enzyme, provided that there was present, at the

ATP binding site, a Cys residue that would be accessible

to a bound inhibitor.

It has been reported that the quinazoline-based

kinase inhibitors are effective inhibitors of VEGFR-2.²⁶

Additionally, there is extensive knowledge, derived from

molecular modeling studies and experimental data, that

shows how this class of inhibitors binds at the active

site of kinases.^27-29We therefore decided to use the

quinazoline core structure as the foundation for the

construction of these covalently binding VEGFR-2 in-

hibitors. Accordingly, we docked the known VEGFR-2

inhibitor 1 (ZD-4190), reported by the Zeneca group,³⁰

into the active site of VEGFR-2, as shown in Figure 1.

The molecule was oriented at the site in a manner

consistent with what is known about how these types

of inhibitors bind to kinases. While it was immediately

obvious that the Cys residue that we targeted in EGFR

and HER-2 was not conserved in VEGFR-2, we did note

that there appears to be an accessible Cys residue, Cys-

1045, located near the 4-anilino group of the inhibitor.

This suggested that it might be possible to design a

quinazoline-based inhibitor that could function as a

covalent-binding inhibitor of VEGFR-2, by replacing the

4-anilino substituent with an appropriately reactive

group approximately the same size as the anilino

substituent. We also noted that there are certain other

kinases of interest that are also involved in angiogen-

esis, such as PDGFR and VEGFR1, where a Cys residue

homologous to Cys-1045 is conserved. The inhibitors we

describe herein potentially could also function as cova-

lent binding inhibitors of these targets.

Figure 1. Binding model of 1³⁰at the active site of VEGFR-

2. Note the proximity of Cys-1045 to the 4-anilino substituent.

potentially high reactivity of these quinone-based in-

hibitors would be a major issue in designing an entity

with druglike properties. Nevertheless, we felt that we

might be able to moderate the reactivity of this type of

inhibitor, to the required degree, by choosing appropri-

ate substituents on the quinone ring. One assumption

inherent in any such approach involving the use of a

highly electrophilic species as a potential drug is that

the drug needs to be more reactive toward the target

enzyme than it is toward other nucleophilic species

[such as glutathione (GSH)] that may be present in high

concentrations in the circulation or in the cytosol.

Assuming that these quinone-based inhibitors bind as

proposed with the quinone ring located nearby Cys-

1045, we can rely on an anticipated entropic effect

wherein the reactivity of such an inhibitor, when bound

at the active site, will be enhanced relative to its

reactivity in solution due to the two reactive centers

being held in close proximity. In addition, one could

speculate that a drug based on a moderately reactive

irreversible inhibitor might not need to exhibit a half-

life in circulation comparable to that of a conventional

inhibitor, since once the target is deactivated by the

covalent interaction, the drug need not be present in

circulation to sustain the desired inhibitory activity. In

We prepared a number of quinazoline derivatives

where we surveyed different types of reactive groups

attached at the 4-position of the quinazoline core.

Among the most biologically active compounds were a

series, based on structure 2, where the phenyl ring of

the 4-anilino substituent of a conventional quinazoline

inhibitor had been replaced with a quinone ring, result-

ing in a series of 2-(quinazolin-4-ylamino)-[1,4]benzo-

quinones.

We realized from the onset that a quinone moiety can

be a highly reactive type of substituent and that the

7562 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

effect, the duration of action of such a drug will be a

function of the rate of receptor turnover.

were then oxidized with ceric ammonium nitrate (CAN)

in aqueous CH₃CN to give the respective quinone

derivatives 9a,b,e-h,k-m.

In this paper, we will describe our work on this class

of compounds. We will discuss the syntheses of members

of this series, describe the results of reactivity studies,

and show how they relate to the biological properties of

these inhibitors. The results of molecular orbital calcu-

lations and molecular modeling were used to rationalize

the experimental observations. Additionally, we will

present our evidence suggesting that these inhibitors

do indeed bind in a covalent manner. Finally, we will

show that one member of this series displays antitumor

activity in an in vivo model.

Some of the quinone inhibitors thus prepared also

served as intermediates to prepare other members of

the series. For example, 9b can be converted to 9c by

the reaction with sodium phenoxide in methanol. Ad-

ditionally, treatment of 9c with methylmercaptan in

CH₂Cl₂resulted in an initial reductive addition product

that was reoxidized in situ with DDQ to furnish the

6-chloro-5-methoxy-3-thiomethyl derivative 9d. In a

similar manner, involving an initial reductive addition

followed by a reoxidation, inhibitors 9i and 9j were

obtained. For reactions involving 9h with nucleophiles,

the reductive additions occurred at the 3-position of the

quinone ring. For the derivative 9j, the regiochemistry

was confirmed by a NOESY NMR experiment, where

we observed an interaction between the methoxy group

and the hydrogen atom at the 5-position. The regio-

chemistry of the product 9i resulting from hydrogen

chloride addition to 9h followed by reoxidation was

confirmed by a NOESY experiment and, later, by an

independent synthesis of the other regioisomer 9c.

Chemistry

The numbering of the carbon atoms of the quinone

ring used in the discussion that follows is as indicated

in structure 2. While this numbering system may not

The quinone derivatives 9a and 9b, having a chlorine

atom at the 5-position, proved to be very versatile

intermediates, since the chlorine atom could be dis-

placed with a variety of nucleophiles, resulting in the

inhibitors shown in Scheme 2. Depending on the type

of nucleophile, different methods were used to effect

these displacements. The phenoxy-substituted deriva-

tive 13 was prepared by treating 9a with phenol and

K₂CO₃in acetone at room temperature (method D). The

other phenoxy-substituted inhibitors, 18-20, were pre-

pared using phase-transfer conditions, where the phenol

and 9a were stirred in a two-phase mixture of 1 N

NaOH and CH₂Cl₂in the presence of the phase-transfer

catalyst Aliquat-336 (tricaprylylmethylammonium chlo-

ride) at room temperature (method F).

The inhibitor 12 was prepared simply by stirring a

solution of 9a and an excess of the N-methylaniline in

THF or glyme (method B). Inhibitors containing a

dialkylamino group at the 5-position of the quinone were

prepared by one of two methods. Simply stirring a

solution of 9a and an excess of the dialkylamine in THF

or dioxane using pyridine hydrochloride as a catalyst

produced the inhibitors 10 and 21-25 (method A). The

exact conditions used with respect to time and temper-

atures are discussed in detail in the Experimental

Section. The other 5-amino substituted derivatives, such

as 11 and 17, were prepared from 9a in a similar

manner without using a catalyst (method B).

The remaining compounds shown in Scheme 2 are

those having a 5-alkoxy group attached to the quinone

ring. These inhibitors, 26-40, were prepared by stirring

a CH₂Cl₂solution of 9a,b with an excess of the alcohol

in the presence of at least 1 equiv of sodium phenoxide-

(H₂O)₃(method E). By observing the mass spectrum at

different time intervals, we monitored the course of

several of these reactions. We observed the initial

formation of the phenoxy derivative 13 followed by a

subsequent buildup of the 5-alkoxy-substituted product

and an associated decrease in 13. Thus, under these

conditions, the substitution reaction proceeds through

the intermediacy of 13. Indeed, treatment of 13 directly

always conform to the proper numbering, keeping the

ring numbers constant at the same positions between

compounds will help clarify the discussion. The proper

ring numbering was used in naming the compounds in

the Experimental Section.

In our preliminary work, we confined our efforts to

optimize the quinone substituents with respect to the

reactivity and in vitro biological activity of the inhibi-

tors. We therefore decided to keep the substituents at

the 6 and 7 positions of the quinazoline constant during

this optimization phase. In this initial effort, we made

quinazolines with 6-methoxy-7-(2-methoxyethoxy) sub-

stituents. Later, only after we had a better understand-

ing of the SAR of the quinone moiety, did we alter the

side chain at the 6 and 7 positions of the quinazoline

ring in an attempt to improve the bioavailability of the

molecules. Additionally, since we knew that the high

reactivity of these inhibitors would be an issue with

respect to them functioning in cellular or in vivo assays,

we concentrated our efforts toward the synthesis of less

reactive quinone derivatives, that is, quinones with

electron-donating substituents. As will be discussed in

detail later, electron-donating substituents on the quino-

ne ring usually will raise the LUMO energy of these

molecules and thereby decrease their reactivity.

The key formamidine intermediate 6 (Scheme 1) was

simply prepared by alkylation of 4-hydroxy-3-methoxy-

benzonitrile with 1-bromo-2-methoxyethane using NaH

in DMF. The resulting compound 3 could be nitrated

selectively, giving 4, which after catalytic reduction

furnished the aniline derivative 5. Condensation of 5

with dimethylforamide-dimethylacetal (DMF-DMA)

furnished 6. Cyclization to the quinazoline and incor-

poration of the 4-anilino group was accomplished in a

single step by refluxing 6 with the substituted anilines

7a,b,e-h,k-m in acetic acid to give the respective

products 8a,b,e-h,k-m. These intermediates, in turn,

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7563

Scheme 1^a

^a(i) NaH, CH₃OCH₂CH₂Br, DMF, 70 °C, 5.5 h. (ii) TFAA/CHCl₃, NH₄NO₂, reflux, 2 h. (iii) Cyclohexene, Pd/C, EtOH, reflux, 24 h. (iv)

DMF-DMA, 100 °C, 5 h. (v) AcOH, reflux. (vi) Ce(NH₄)₂(NO₃)₆, H₂O, MeCN, rt. (vii) ROH, NaOPh(3H₂O), CH₂Cl₂, rt. (viii) (1) MeSH,

CH₂Cl₂; (2) DDQ, CH₂Cl₂. (ix) (1) HCl, CHCl₃, rt, overnight; (2) DDQ, NaOAc, MeCN, H₂O, 1 h, rt. (x) (1) 2-mercaptopyridine, CHCl₃, 0.5

h; (2) DDQ, MeCN.

with alcohols and a base catalyst resulted in the

formation of 14-16.

and 50, respectively, resulting from displacement of the

chlorine atom of 9e, while the reaction with 2-propanol

gave 52, the result of displacement of the methoxy group

of 9e. The reaction of 9e with cyclopropylmethanol gave

a mixture of displacement products 48 and 51, although

only 51 was obtained in a pure state. The dependence

of the regiochemistry of the displacement products on

the nature of the alcohol used is curious and, frankly,

we cannot offer a good rationale for these results.

Having explored the reactivity issues and the SAR

resulting from alterations of the substituents on the

quinone ring, we next turned our attention to modifica-

tions of the 6- and 7-positions of the quinazoline ring.

It is well-known that modifications at these two posi-

tions have an important influence on the water solubil-

ity and bioavailability of these quinazoline-based kinase

inhibitors.^22,24,26,32We prepared inhibitors that retain

the 6-methoxy group of the quinazoline while having a

basic side chain at the 7-position. The preparation of

the chloroquinone intermediates 59a,b needed to pre-

pare these compounds is shown in Scheme 4. We

developed a flexible synthesis that allows the 7-sub-

stituent to be altered by utilizing the relatively ad-

vanced intermediate 56. Reductive amination utilizing

aldehyde 53 and aniline 7a gave the required interme-

diate 54. This was condensed with the known³³4-chloro-

Treatment of the 5-(N-aziridinyl)-substituted deriva-

tive 21 with hydrochloric acid in THF resulted in

hydrolytic ring opening, giving the 2-hydroxypropy-

lamino-substituted inhibitor 41. This, in turn, was

converted to the oxazolidin-2-one derivative 42 by

reaction with carbonyl diimidazole.

The substituted aniline 7e needed to prepare 9e was

itself prepared starting with the known substituted

benzaldehyde 43³¹(Scheme 3). The acetate group of 43

was converted to a methoxy group by adding a KOH

solution to a solution of the compound in EtOH in the

presence of dimethyl sulfate, resulting in 44. This

compound was then selectively demethylated with LiCl

in hot DMF. The resulting intermediate 45 was then

oxidized to the hydroquinone derivative 46 using basic

H₂O₂. The free hydroxyl groups were then re-methylated

with dimethyl sulfate and K₂CO₃in DMF and the nitro

group was reduced with iron and acetic acid in refluxing

MeOH, giving the desired intermediate 7e. This was

used to prepare 9e as described above.

The reaction of 9e with an excess of some alcohols

using cesium carbonate in methylene chloride gave

unanticipated results. The reaction with methanol or

1,3-difluoropropan-3-ol gave the expected products, 49

7564 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

Scheme 2^a

^a(i) Method A: R₂NH or RNH₂, pyridine hydrochloride, THF or dioxane. Method B: R₂NH or RNH₂, THF or ethylene glycol dimethyl

ether. Method C: ROH, Et₃N, CH₂Cl₂, 60 °C. Method D: ROH, K₂CO₃, acetone, rt. Method E: ROH, NaOPh(3H₂O), CH₂Cl₂, rt. Method

F: ROH, 1 M NaOH, CH₂Cl₂, rt, Aliquat 336. (ii) HCl, THF, H₂O, 18 h, rt. (iii) Carbonyl diimidazole, 1-methyl-2-pyrrolidinone, 80 °C, 27

h.

7-fluoro-6-methoxyquinazoline, 55, in a mixture of

refluxing pyridine-tert-butyl alcohol to give 56. The

fluorine atom of 56 could be readily displaced by

utilizing sodium alkoxides derived from alcohols 57a,b.

The products of these displacements were not purified

but were directly deprotected using TFA to give the

intermediates 58a,b; these in turn, were oxidized with

CAN to the quinones 59a,b as described previously.

Displacement of the 5-chloro substituent in 59a,b with

various alcohols utilizing the methods already discussed

gave the inhibitors 60-75 shown in Scheme 5.

Reactivity Studies. Since we are proposing that

these quinone derivatives function as covalent-binding

inhibitors, information about the reactivity of these

compounds toward sulfur-containing nucleophiles would

be of obvious interest. Nucleophiles of major importance

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7565

Scheme 3^a

^a(i) EtOH, (Me)₂SO₄, 40% KOH, 55 °C, 1 h. (ii) LiCl, DMF, H₂O, 110 °C, 3 h. (iii) 1 N NaOH, H₂O, H₂O₂, MeOH, 50 °C, 3.5 h. (iv)

(Me)₂SO₄, K₂CO₃, DMF, 80 °C, 1 h. (v) MeOH, Fe, AcOH, reflux, 2 h. (vi) ROH, CsCO₃, CH₂Cl₂, rt.

Scheme 4^a

^a(i) AcOH, NaBH(OAc)₃, 1,2-dichloroethane, 2.5 h, rt. (ii) pyridine, t-BuOH, 24 h, reflux. (iii) (1) NaHMDS, THF; (2) TFA. (iv) CAN,

MeCN, H₂O, rt.

would be Cys-1045 of the target protein and, perhaps,

other sulfhydryl-containing species, such as glutathione,

which is expected to be present at high concentrations

in the circulation or the cytosol. According to frontier

molecular orbital theory, a major component of the

reactivity of such a system would be determined by the

difference in energy between the HOMO of the nucleo-

phile and the LUMO of the electrophile. The smaller

this HOMO-LUMO gap the more reactive the system

is expected to be. Since in the discussions to follow,

7566 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

Scheme 5^a

in our binding models of these inhibitors at the active

site in VEGFR-2. Accordingly, the conformation we used

to calculate the frontier orbitals of 9g shown in Figure

2 is of this type.

Conceivably, under neutral or physiological condi-

tions, the reaction of a substituted quinone-based

inhibitor such as 76 with a sulfhydryl-containing spe-

cies, either in solution or in the VEGFR-2 enzyme-bound

state, can involve a redox process consisting of a one-

electron reduction, wherein discrete semiquinone-like

intermediates³⁶such as 77a-c are formed (Scheme 6).

We are assuming that the formation of these intermedi-

ates occurs in the rate-determining step of the reaction.

These intermediates can be interrelated by tautomer-

ization. As stated above, the propensity to form these

intermediates will be a function of the LUMO energy

of 76. These semiquinone intermediates, in turn, can

combine with the sulfur radical formed in the reaction

or the neutral sulfhydryl species, followed by a subse-

quent tautomerization, to produce a reductive addition

product A. Alternatively, when a quinone substituent

has the ability to function as a leaving group and the

position of attack of the sulfur radical is at the site of

the quinone substituent, an ipso substitution product,

B, can result. Since the ipso substitution product is itself

a quinone, further reaction, when in solution, with the

sulfhydryl species is possible, resulting in the formation

of a bis-adduct F. Additionally, in the presence of an

excess of reducing species, the semiquinone intermedi-

ates 77a-c or the ipso substitution product B can be

further reduced to give the hydroquinone products E

and C, respectively. Finally, since the reactions studied

were done without protection from atmospheric oxygen

and since any quinoid species present can act as an

oxidizing agent, the hydroquinone products A, C, E, and

F can be reoxidized and subsequent addition of a

sulfhydryl species could result in multiple addition

products such as G. Under physiological conditions, it

is expected that products of type C-G can form only

when the inhibitor is unbound, in the circulation or

cytosol, but not when the inhibitor is in the enzyme-

bound state, since a second reducing group is absent at

the active site and there is no oxidizing source to

reoxidize a hydroquinone adduct. Therefore, it is ex-

pected that the drug-VEGFR-2 conjugate will be a

product of either type A or B, where the SR′′ group

represents the protein.

^a(i) Method C: ROH, Et₃N, CH₂Cl₂, 60 °C. Method E: ROH,

NaOPh(3H₂O), CH₂Cl₂, rt. Method G: ROH, CsCO₃, CHCl₃, rt.

where the nucleophile is kept constant, the reactivity

will be a function of the LUMO energy of the inhibitor,

the lower the energy, the more reactive the molecule.

Furthermore, with reference to the Arrhenius equation,

it is expected that the relative reactivity of these

inhibitors with a given nucleophile will depend expo-

nentially on the LUMO energy. Earlier studies of simple

substituted quinones do indicate a correlation between

reactivity and LUMO energy.^34,35Figure 2 shows the

frontier orbitals for the unsubstituted quinone deriva-

tive 9g. It can be seen that the LUMO of 9g has largely

π-character and that the most significant part of the

orbital density is centered on the quinone ring. Most of

our molecular orbital calculations were performed using

the semiempirical AM1 method. Since we are comparing

relative LUMO values, we felt that high accuracy in

these calculations would not be crucial for our purposes.

In performing these calculations, we observed a signifi-

cant dependence of the LUMO energies on the confor-

mations of the inhibitors with respect to rotations of

bonds located nearby the quinone ring and with respect

to the orientation of some types of quinone substituents.

Molecular mechanics calculations indicate that confor-

mations where the NH group joining the quinazoline

and quinone portions of the molecule forms an intramo-

lecular H-bond with one of the quinone carbonyl groups

are more favored than alternate conformations lacking

this H-bond. Also, this type of conformation is favored

The types of products formed and their regiochemistry

will be determined in steps after the rate-determining

step and will depend on a number of factors, including

the relative populations of the semiquinone tautomers

and their conformations, the steric influence of the

quinone substituents, the ability of a substituent to

serve as a leaving group, and the distribution of the

orbital density in the singly occupied molecular orbital

(SOMO) of the semiquinone intermediates. In the case

of the enzyme-bound inhibitor, the relative orientation

of the two reacting centers may also be a factor in

product formation. To further understand the regio-

chemistry of this type of reaction, we determined the

products resulting from the reaction of some selected

inhibitors with simple sulfhydryl-containing nucleo-

philes (Scheme 6 and Table 1). The mass spectra of the

reaction mixtures were monitored over a 24 h period.

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7567

Figure 2. Frontier molecular orbitals of a low-energy conformation of 9g by the AM1 method using the MOPAC module in

Sybyl. Note that the LUMO is largely of π-character and is localized predominately on the quinone ring.

If a mass corresponding to any of the potential products

A-G was observed in the mass spectrum during this

time period, it is so indicated in the table. In some cases

the products were isolated and their regiochemistry was

determined by NMR. Product ratios varied over time

with multiple addition/reoxidation products increasing

with time.

Knowledge of the rates of reaction of these quinone-

based inhibitors with a sulfhydryl-containing species

could be important information that might help in the

interpretation of some of our biological observations.

Accordingly, we measured the half-life (t_1/2) of selected

5-substituted inhibitors in the presence of reduced

glutathione in H₂O-acetonitrile (1:1). A 1 or 10 µM

solution of the inhibitor was incubated in the presence

of 100 µM of glutathione, and the amount of inhibitor

remaining over time was monitored by HPLC-MS. It

is clear from the data shown in Table 2 that the rates

of reaction of these molecules are highly dependent on

the quinone substituents with t_1/2values varying be-

tween less than 1 min to weeks long.

nally, a single LUMO energy value was obtained that

was derived from a weighted average of the individual

LUMO values on the basis of the Boltzmann distribu-

tion of the conformers. In general, we find that electron-

withdrawing groups attached to the quinone ring lower

the LUMO energy (more reactive) while electron-donat-

ing groups raise it (less reactive).

It is evident from the data shown in Table 2 that the

reactivities of these inhibitors are correlated with the

averaged LUMO energies. The minor discrepancies in

the correlation can be attributed to other effects such

as those of steric origin and the fact that these calcula-

tions are for neutral molecules in the gas phase and

certain derivatives such as 23 are expected to be

protonated under the reaction conditions. In general,

the lower the LUMO energy, the more reactive the

molecule appears to be. Moreover, these data suggest

that the correlation is not linear but is approximately

exponential. We therefore feel that these LUMO values

can be useful predictors of the reactivities of these

inhibitors and they will be used as such in the discus-

sions to follow.

Also shown in Table 2 are the calculated LUMO

energies for the inhibitors. During preliminary calcula-

tions it was evident that, in some cases, the LUMO

energies varied significantly between different confor-

mations of the inhibitors. While the LUMO values were

usually insensitive to conformational changes at posi-

tions remote from the quinone ring (such as the 6- and

7-positions of the quinazoline), the energies were sensi-

tive to rotations of bonds located nearby the quinone

ring and to the orientation of certain quinone substit-

uents. Since the kinetic studies were done under pseudo-

first-order conditions, it is expected that, for a given

compound, the overall rate of reaction will be a combi-

nation of the rates of reaction of the individual conform-

ers weighted by their concentration. Similarly, it seems

appropriate to express the LUMO energies as weighted

averages of the energies of the individual conformers.

Accordingly, for the compounds listed in Table 2, we did

a full conformational grid search with respect to bond

rotations located nearby the quinone ring. For each of

these conformers we calculated a LUMO energy. Fi-

Molecular Modeling. A binding model for one of our

inhibitors, 26, in the active site of VEGFR-2 is presented

in Figure 3. The initial protein coordinates used in these

studies are those reported for the crystal structure of

the catalytic domain of VEGFR-2.²⁵Water molecules

were removed and the inhibitor was docked using the

docking algorithm FRED.³⁷Low-energy poses were

rescored using the Poisson Boltzmann molecular me-

chanics scoring function (PBMM-SA).³⁸The resulting

model shows the inhibitor oriented in a manner consis-

tent with the way other quinazoline derivatives have

been shown to bind to different kinases.^28,29

A favorable interaction between the quinazoline N1

of 26 and the hinge region backbone NH of Cys-919 is

found to be significant in our model. The polar aromatic

hydrogen atom between N1 and N3 of the quinazoline

ring is favorably oriented at approximately 3 Å from the

backbone carbonyl of Glu-917. In the most favored

bound conformation, the oxygen atom of a quinone

carbonyl group makes an intramolecular H-bond to the

7568 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

Scheme 6

NH substituent on the quinazoline. This is similar to

the low-energy conformations found for the unbound

inhibitors. The other quinone carbonyl is oriented

toward the side chain of Lys-868, within approximately

3.4 Å. The center of the quinone ring is perpendicularly

oriented toward the sulfhydryl group of Cys-1045, being

located about 4.5 Å away, and is within a distance that

should allow covalent bond formation considering ther-

mal motions of the bound inhibitor and the protein.

Interactions of quinone substituents with the VEGFR-2

binding pocket will vary, depending on the nature of

the substituent and its location on the quinone ring. For

26, the 1,3-difluoroisopropoxy substituent occupies a

region of the binding pocket proximal to the region

where Lys-868 and Asp-1046 are found. This region

extends into a hydrophobic pocket over 10 Å up toward

the N-terminus of the kinase where larger quinone

substituents, on the 5-position of the quinone ring, can

fit. Thus, lipophilic interactions in this region could

result in enhanced protein-ligand binding energetics.

Biology

The results of our biological assays are shown in Table

3. Our primary assay is a DELFIA kinase assay,³⁹

where recombinant full cytoplasmic domain of human

VEGFR-2 is utilized and where the activity of an

inhibitor is expressed as an IC₅₀. Usually, this assay is

conducted using a 10 µM concentration of ATP. For

selected compounds, to explore the ATP-competitive

nature of the inhibition, the assay was also conducted

at a 1 mM ATP concentration.

As stated above, since the reactivity of our inhibitors

toward sulfhydryl-containing nucleophiles is an impor-

tant issue, we decided, for selected compounds, to also

determine IC₅₀s in the presence of 100 µM concentration

of glutathione and in the presence of fresh mouse

plasma. It is unlikely that a compound that loses

inhibitory activity when tested in the presence of these

agents could ever become a useful drug.

While the IC₅₀of a conventional ATP competitive

inhibitor is a direct measure of the compound’s ability

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7569

Table 1. Observed Products (see Scheme 6) of the Reaction of

Selective Inhibitors with a Sulfhydryl-Containing Species^a

product type (Scheme 6)

compd

R

A

B

C

D

E

F

G

16^b

15^b

13^c

26^d

9h^e

9a^f

9l

5-OMe

5-OCH₂Ph

5-OPh

5-OCH(CH₂F)₂

6-OMe

5-Cl

5-SMe

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

+

-

^aReactions were conducted under air in CH₃CN/H₂O or THF/

MeOH/H₂O solutions with the sulfhydryl species in excess (4-10-

fold). Reactions were analyzed at various time points by HPLC-

MS or by measuring the MS of the mixture directly. A + sign

indicates observation of the designated product, while a - sign

indicates that it was not observed at the time of the measurement.

Minor products have a + sign. Unless indicated otherwise the

sulfhydryl nucleophile was ethylmercaptan. ^bThe regiochemistry

of A was 5-OR-3-SR′′ as shown by NOE. ^cThe sulfhydryl nucleo-

phile was glutathione. The regiochemistry of the products were

not determined. ^dThe regiochemistry of D was 5-OCH(CH₂F)₂-3-

SR′′ as shown by NOE. The oxidized forms of F and G were

observed. ^eThe regiochemistry of the products is suggested to be

6-OMe-3-SR′′ in analogy to 9j. ^fThe regiochemistry of A and F

was not determined.

Figure 3. Binding model for 26 at the active site of VEGFR-

2. Protein coordinates are those from the reported crystal

structure. The distance of the sulfur atom of Cys-1045 from

the center of the quinone ring is about 4.5 Å.

the rate of reaction of this type of inhibitor with the

target enzyme, the IC₅₀value could simultaneously

consist of components reflecting both reversible and

irreversible binding.

For comparison, enzyme data for the known revers-

ible-binding ATP-competitive inhibitors, the quinazoline

78 (ZD4190)⁴⁰and the phthalazine 79 (PTK787),⁴¹are

included in Table 3.

Table 2. Half-Life (t_1/2) of Selected Compounds at 1 and 10 µM

Concentrations in the Presence of 100 µM of Glutathione in

H₂O-Acetonitrile (1:1)^aand the Calculated LUMO Energies^b

t_1/2(h)

compd

R₁

R₂10 µM 1 µM LUMO^b(eV)

9a

13

18

9c

9l

26

15

36

16

23

29

12

10

Cl

H

<0.02 <0.02

-1.92

-1.82

-1.78

-1.89

-1.81

-1.80

-1.68

-1.52

-1.54

-1.59

-1.48

-1.62

-1.51

OPh

OPh-4-N-imidazolyl

OCH₃

SCH₃

OCH(CH₂F)₂

OCH₂Ph

OCH₂CH₃

OCH₃

Cl 0.5

<0.5

2

H

3

15

54

72

80

11

55

62

In addition to the aforementioned enzyme assays, we

evaluated these inhibitors in KDR15 cells (human

embryonic kidney 293 cells transfected with full length

VEGFR-2),⁴²where we measured the inhibition of

VEGF-stimulated VEGFR-2 autophosphorylation. Again,

the measurement is expressed as an IC₅₀value.

Also included in Table 3, as predictors for reactivity,

are the calculated LUMO energy values. While the

LUMO values for those compounds given in Table 2,

calculated from the Boltzmann distribution of the

conformers, are repeated in Table 3, the majority of the

LUMO values in Table 3 were not derived from the

Boltzmann distribution but, instead, are calculated for

single minimized conformations. The conformations

used in the calculation are low in energy and each has

the intramolecular H-bond, between the quinone car-

bonyl group and the NH group with the quinazoline-

68

N-piperazinyl-4-CH₂Ph H 85

73

OCH(CH₃)₂

N(CH₃)Ph

N(CH₃)₂

H

227

295

601

213

341

160

^aKinetic studies were conducted at room temperature in air.

^bLUMO energies are averages of those of individual conformers

weighted by the Boltzmann distribution.

to fit at the active site and compete with ATP, the

situation is expected to be more complicated for an

irreversible binding inhibitor where the IC₅₀will be a

measure of both the fit at the active site as well as the

reactivity of the molecule. In this situation, one expects

that for two molecules that fit the site equally well, the

more reactive molecule will have a lower IC₅₀. The IC₅₀

value that we measure for a covalent-binding inhibitor

could be time-dependent and additionally, depending on

7570 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

Table 3. Inhibition of VEGFR-2 Kinase, Inhibition of VEGF-Stimulated VEGFR-2 Autophosphorylation in KDR15 Cells, and LUMO

Energies

IC₅₀(nM)

5%

IC₅₀(nM)

5%

10 µM 100 µM

compd ATP^a

1 mM KDR15

10 µM 100 µM

LUMO (eV)^fcompd ATP^a

1 mM KDR15

cells^e

GSH^bplasma^cATP^dcells^e

GSH^bplasma^cATP^d

LUMO (eV)^f

9a

9b

9c

9d

9e

9f

9g

9h

9i

4.4

13.5

1

67.1

1.7

109.4

6

3.8

111.9

>1000 569.1

>1000 -1.90 (-1.92)

31

32

33

34

35

36

37

38

39

40

41

42

49

50

51

52

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

78

79

<1

6.8

21.3

70

41.9

32.7

-1.96

-2.00

<1

>1000 -1.87

-1.62

20.9 5.1

>1000 >1000

>1000 375.9

<1

466

-1.70 (-1.89)

-1.90

52.4

31.2

23.6

>1000

256.3

239.2

>1000

57.4

495.2

32.5

16.5

>1000

59.7

28.7

41

-1.60

-1.91

>1000 -1.65

>1000 -1.56 (-1.52)

>1000 -1.69

-1.71

607.8 <15

-1.65

>1000 >1000

>1000 -1.64

4.6

5.8

4.5

3.4

365.9

18.6

25.2

8.1

-1.64

-1.90

870

-1.74

-1.61

-1.43

9j

9k

9l

>1000 >1000 >1000

-1.69

15.8

30.1

9

881.6

803.1

-1.91

>1000 >1000

>1000 114.7

>1000 >1000

43.3

456

31

-1.80 (-1.81)

-1.92

10.3

16.1

5.7

2.3

1.9

>1000

7.3

>1000 -1.71

-1.77

9m

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

1144.1 5000.3 >20000

101.2

-1.38 (-1.51)

-1.96

>1000 -1.49

193.5

97.3

129.9

3.7

-1.91

16

2.7

>1000 >1000

>1000 560.1

>1000 -1.45 (-1.62)

-1.87

184

-1.75 (-1.82)

-1.71

239.3 239.7

3020.7

11.2

5

-1.53

-1.62

-1.53

-1.55

-1.58

-1.81

-1.73

-1.68

-1.54

-1.71

-1.67

-1.71

-1.66

-1.69

-1.53

-1.8

236.5 474.4

5.1

6.9

15.3

29.5

189

329

536

3

43

50.3

366.1

>1000 436.3 25

-1.59 (-1.68)

-1.59 (-1.54)

-1.41

3.9

5.8

240.9 558.6

13.6

26.1

158.6

7.4

359.4 >1000

310

271.8 406.3

1274

1.7

7.6

>1000 >1000 5.8

310

-1.91 (-1.78)

-1.75

6.4

12.8

1.8

15.1

31.8

6.3

8.8

9

83.1

42.1

5.1

18.3

31.3

0.9

1.5

1.9

2.2

13.7

9.9

5.2

4.8

2.6

78

206

-1.55

9.6

540.2 >1000

298.6 231.8

-1.71

16.9

4.5

>1000

>1000 75.5

>1000

-1.35

84.8

133

125.4

270.6

>1000 -1.58 (-1.59)

>1000 -1.42

-1.33

3.6

5.1

70.7

35.9

4.2

3.2

41.8

138

10.4

371.9 >1000 >1000

9.1

88.7

1.7

491.3 >1000 >1000 542.9 481

0.4 70.9 792

53.5

29.6

13.9

2.4

21.3

110.5

4.8

24.8

>1000 107.8 193

117.8

18.5

212

-1.89 (-1.80)

-1.66

>1000 -1.64

-1.54 (-1.48)

-2.07

149.8

5980.5 <15

6308.1 <15

3

^aThe assay was conducted using an ATP concentration of 10 µM. The IC₅₀is the concentration in nM needed to inhibit the

phosphorylation of the poly(Glu₄-Tyr) peptide by 50% as determined from the dose-response curve. Where there were multiple

determinations, the IC₅₀values were averaged, and the standard deviations, on average, were 43% of the average IC₅₀value. ^bThe assay

was conducted in the presence of 100 µM of glutathione using an ATP concentration of 10 µM. Where there were multiple determinations,

the IC₅₀values were averaged, and the standard deviations, on average, were 39% of the average IC₅₀value. ^cThe assay was conducted

in the presence of 5% (v/v) fresh mouse plasma using an ATP concentration of 10 µM. Where there were multiple determinations, the IC₅₀

values were averaged, and the standard deviations, on average, were 41% of the average IC₅₀value. ^dThe assay was conducted using an

ATP concentration of 1 mM. Where there were multiple determinations, the IC₅₀values were averaged, and the standard deviations, on

average, were 29% of the average IC₅₀value. ^eConcentration in nM needed to inhibit the VEGF-stimulated autophosphorylation VEGFR-2

in intact KDR15 cells by 50% as determined from the dose-response curve. Where there were multiple determinations, the IC₅₀values

were averaged, and the standard deviations, on average, were 42% of the average IC₅₀value. ^fLUMO values calculated by the AM1

method using MOPAC as implemented in Sybyl for a single low-energy conformation similar to that as shown in Figure 2. LUMO energies

in parentheses are averages of those of the individual conformations weighted by the Boltzmann distribution.

NH-quinone torsional angles similar to those shown

in the structure in Figure 2. Conformations of this type

are energetically favored and favored by our binding

models. However, an arbitrary, but minimized, orienta-

tion was used for the quinone substituents. With the

possible exception of some of the compounds with

conformationally mobile quinone substituents having

multiple heteroatoms or those compounds where rota-

tions along the quinazoline-NH-quinone torsional

angles result in large changes in the location of polar

or conjugated substituents, these LUMO values should

approximate those obtained by utilizing a full confor-

mational analysis. However, due to the uncertainties

resulting from not evaluating these LUMO energies

using a complete conformational analysis, these LUMO

energies should only be viewed as rough estimates for

the reactivities.

inhibitor in our VEGFR-2 enzyme assay. The quinazo-

line-based inhibitor 78 was studied in detail and it is

evident that it retains full activity when tested in the

presence of glutathione. It loses some (2-3-fold) activity

when tested in the presence of plasma, possibility due

to protein binding. Both 78 and 79 show good potency

in our cell-based autophosphorylation assay. These

results will serve as a basis with which to compare

reversible-binding ATP-competitive inhibitors with our

quinone-based irreversible binding inhibitors.

Among the more potent enzyme inhibitors are those

that have an electron-withdrawing halogen substituent

at the 5-position of the quinone ring, such as 9a and

9m. As indicated by the short half-life observed in the

kinetic experiment with 9a and suggested by the very

low LUMO energies calculated for these derivatives,

these 5-halo substituted compounds have high reactiv-

ity. It is therefore not surprising that they lose a

significant part of their inhibitory activity when tested

With respect to the conventional reversible binding

inhibitors 78 and 79, each is a moderately potent

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7571

in the presence of glutathione or plasma. One of these

derivatives, 9a, was evaluated in the cell-based phos-

phorylation assay and was found to be inactive. Pre-

sumably, as a result of its high reactivity, 9a is

scavenged by endogenous nucleophiles within the cell

before it gets a chance to reach the target enzyme. As

one might expect, the addition of a second Cl atom, to

give the 5,6-dicloro derivative 9b, decreases the LUMO

energy further with retention of good potency in the

enzyme assay. Placement of an electron-releasing alkoxy

group at the 6-position, as in compounds 9e, 51, and

52 does not significantly overcome the influence of the

5-chloro substituent, both with respect to the calculated

LUMO energies and the biological properties of the

molecules.

found in our kinetic experiments (Table 2). Perhaps

reflecting this low reactivity is the low potency exhibited

by this and related compounds in our enzyme assay.

While several of these amino-substituted derivatives

such as 22-24 appear to retain much of their activity

when tested in the presence of glutathione, it is not clear

whether these compounds are functioning entirely as

covalent-binding inhibitors, and in fact, they may partly

function as conventional reversible binding inhibitors

over the short time course of the enzyme experiment.

The relatively high IC₅₀values would then simply reflect

a less than optimal fit at the active site. Among the

alkylamino-substituted inhibitors that have been evalu-

ated in our cell-based autophosphorylation assay, some,

such as 10, 17, and 21, show inhibitory activity com-

parable to that observed in the enzyme assay.

Like the 6-alkoxy-5-chloro analogues, the 5-alkoxy-

6-chloro derivatives such as 9c and 32 are very potent

inhibitors of VEGFR-2. However, in contrast to the

former series, these inhibitors tend to retain good

activity when tested in the presence of glutathione or

plasma, despite their low LUMO energies (compare the

isomers 9c and 9e). The 6-alkoxy-3-chloro derivative 9i,

which is isomeric with both 9e and 9c, is a significantly

less potent inhibitor (64-110 fold) than either of the

other isomers, despite the fact that the calculated

LUMO energies for 9e, 9c, and 9i are about the same.

In this case, it is likely that the relative potency of these

isomers is significantly influenced by other factors in

addition to reactivity. It is possible that the presence of

a 3-sustituent prevents the inhibitor from adopting an

appropriate binding conformation by forcing the quinazo-

line-NH-quinone torsional angles into a less than ideal

arrangement. Indeed, none of the inhibitors we prepared

that have 3-substituents, such as 9d and 9j (and others

not shown), have high potency. Another example of the

influence of a 3-substituent can be seen in comparing

the activities and LUMO energies of 9d and 9c.

Another group of inhibitors that generally showed

high potency in the enzyme assay and high reactivity

are the derivatives with 5-phenoxy-substituted quinones

such as 13 and 18-20. The kinetic experiment (Table

2) with 18 clearly illustrates its high reactivity. While

the calculated LUMO energies are not as low as those

obtained for the halogen-substituted quinone deriva-

tives, the energies are still well below those calculated

for simple alkoxy-substituted compounds. For those

cases (13 and 18) where we performed the enzyme assay

in the presence of glutathione or plasma, the compounds

lost most of their activity. Again, we attribute this to

the instability of the compounds in the presence of

reducing species. Nevertheless, some of these com-

pounds do have activity in the cell autophosphorylation

assay with IC₅₀values in the sub-µM range, suggesting

that the relative reactivity of these analogues, in cells,

with the VEGFR-2 target is competitive with that of any

endogenous reducing species.

The derivative 42, having a cyclic urethane moiety

at the 5-postion of the quinone ring, showed a significant

increase in potency relative to the alkylamino-substi-

tuted inhibitors in the enzyme assay. One could antici-

pate that the electron-donating ability of the urethane

N-atom would be greatly attenuated relative to that of

an ordinary dialkylamino group and this appears to be

reflected in the calculated LUMO energy of 42, which

has decreased significantly compared to that found for

the alkylamino-substituted derivatives. It is likely that

the increase in potency observed for 42 is due, in part,

to its increased reactivity. Additionally, this increased

reactivity of 42 could be a controlling factor resulting

in a loss of its activity in the enzyme assay when tested

in the presence of glutathione, and it may be respon-

sible, in part, for the lack of activity in the cell assay.

The inhibitor incorporating an N-methylaniline moi-

ety at the 5-position of the quinone ring, 12, proved to

be a good inhibitor of VEGFR-2 with an IC₅₀) 16 nM.

However, the compound lost activity when tested in the

presence of glutathione, despite the fact that the kinetic

experiment (Table 2) suggested that this compound is

only a moderately reactive substance.

An example of an inhibitor with a sulfur atom

attached to the 5-position of the quinone ring was

prepared, the methylthio derivative 9l. On the basis of

the LUMO energy, it is predicted to be fairly reactive,

and this was confirmed by the kinetics exhibited by the

compound (Table 2). The compound showed good po-

tency in the enzyme assay, but it did not retain its

activity when tested in the presence of glutathione.

One group of compounds that we extensively studied

comprises inhibitors having an alkoxy substituent at the

5-position of the quinone ring. The LUMO energies that

we calculate for inhibitors with simple alkyl groups fall

within the midrange of the LUMO scale. Additionally,

we found that we could modulate the LUMO energies

by altering the substituents attached to the alkoxy

group. We observed an interesting trend with respect

to the methoxy-substituted derivatives 16, 60, and 64.

While these compounds are only moderately potent

inhibitors in our enzyme assay, they are among the most

potent inhibitors in our cell-based autophosphorylation

assay. In fact, for these compounds, the IC₅₀values for

inhibiting VEGFR-2 autophosphorylation in intact cells

is lower than that observed for inhibiting the enzyme

in solution. A truly satisfying explanation for this

observation is lacking, but one can speculate that these

Among the inhibitors we prepared, those with an

alkylamino substituent on the quinone ring, such as 10,

11, 17, and 21-25, are predicted to be the least reactive.

The alkylamino groups are the strongest electron-donor

groups that we incorporated onto our inhibitors, and the

calculated LUMO energies for these compounds are

among the highest found. In particular, the 5-dimethy-

lamino derivative 10 is the least reactive molecule we

7572 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

compounds are more efficient inhibitors of the full-

length, membrane bound, enzyme present in cells

compared to the truncated cytoplasmic domain used in

the enzyme assay, and in fact, in an earlier study²⁷with

EGFR, we observed such a situation where IC₅₀mea-

surements were higher with a truncated enzyme com-

pared to the full length enzyme. On the basis of the

kinetic experiments (Table 2), the simple alkoxy deriva-

tives 16, 29, and 36 are only moderately reactive.

Nevertheless, as will be discussed below, mass spectral

studies do prove that one of these compounds, 16, does

indeed form a covalent interaction with the enzyme. The

moderate reactivity of these compounds could also be a

factor contributing to the improved cell activities.

Compared to the 5-methoxy analogues, those having

isopropoxy or ethoxy substituents, 29 and 36, respec-

tively, are less potent in the enzyme assay, even though

they have similar LUMO energies.

Intuitively, one might have expected that adding an

additional electron-donating methoxy group to the

6-position of the quinone to give the 5,6-dimethoxy

derivative 49 would have resulted in a molecule having

a higher LUMO energy and decreased reactivity, but

in fact, this additional methoxy group results in a

lowering of the LUMO energy. This could be a contrib-

uting factor in the greater potency of 49 compared to

16 in the enzyme assay.

We prepared a number of compounds where the

5-alkoxyquinone substituent has an attached unsatur-

ated and lipophilic group, for example, the 5-benzyloxy-

substituted derivatives 15, 33-35, 38, 63, 68, and 69.

On the basis of the weighted average LUMO energies

calculated for the benzyloxy analogue 15 and for the

simple alkoxy derivatives 16, 29, and 36, it appears that

the benzyloxy group contributes to a lowering of the

energy. Thus, compared to these simple alkoxy-substi-

tuted derivatives, 15 is predicted to be more reactive.

The kinetic experiments do confirm this, although the

differences in reactivity are modest. These benzyloxy

derivatives are clearly more potent than the simple

alkoxy analogues. A combination of higher reactivity

and an improved fit at the active site for the benzyloxy-

substituted inhibitors could account for the improved

potency. To be sure, our binding model for these

inhibitors indicates that there is a large lipophilic pocket

that could easily accommodate a benzyloxy group,

resulting in an improved interaction with the enzyme.

Interestingly, each of these benzyloxy-substituted in-

hibitors retains a significant portion of the enzyme

activity when tested in the presence of glutathione and,

in some cases, in the presence of plasma as well. The

related heterocyclic analogues 27, 67, and 74 show an

in vitro profile similar to that of the benzyloxy substi-

tuted inhibitors. The inhibitor 31, which has a 6-chloro

group on the quinone ring in addition to the 5-(m-

fluorobenzyloxy) group, is an extremely potent inhibitor

of VEGFR-2, even in the presence of glutathione,

although the calculated LUMO energy suggests that it

should be a very reactive molecule.

retain their activity when tested in the presence of

glutathione or plasma. The LUMO energies calculated

for these compounds suggest intermediate reactivity.

A most interesting group of inhibitors are those that

have a fluoroalkoxy group attached at the 5-position of

the quinone ring. These inhibitors include 26, 30, 50,

65, and 75. MO calculations indicate that the strong

inductive electron-withdrawing nature of the fluorine

atoms can result in a significant lowering of the LUMO

energies. Calculated LUMO energies for these inhibitors

were very sensitive to the overall orientation of the

fluoroalkyl group and to the relative orientation of the

fluorine atoms within the group. In particular, for the

1,3-difluoroisopropoxy-substituted compounds, we found

that LUMO energies could vary more than 0.3 eV

between low-energy conformers having different relative

orientations of the two fluorine atoms. Because of this,

the LUMO energies calculated from the weighted aver-

ages of the conformations should be considered more

reliable than those calculated from single conformations.

By comparing compounds 26 and 29, having the steri-

cally similar 1,3-difluoroisopropoxy and isopropoxy groups

attached to the quinone ring, respectively, one can see

the striking influence the fluorine atoms have on

reactivity (Table 2). The fluorine atoms lower the LUMO

energy of 26 by over 0.3 eV compared to 29. The result

of this is the very significant decrease in the t_1/2of 26

compared to 29. The enhanced reactivity of 26 likely

accounts for its greater potency in the enzyme assay

compared to 29. In addition, much of the activity is

retained for 26 when the assay is conducted in the

presence of glutathione or plasma. The increased po-

tency for 26 compared to 29 is also evident in the cell

assay, but it is not as impressive.

It is well-established for the quinazoline and the

related 3-cyanoquinoline-based kinase inhibitors that

placement of a water-solublizing substituents at the 6-

or 7-position results in improved solubility and bioavail-

ability. Binding models for these types of compounds

indicate that such water-solublizing groups point out

toward solvent, and therefore, we expected that incor-

poration of a water-solubilizing substituent at the

7-position in our series would have a minimal effect on

the potency of these inhibitors in our enzyme assay. For

the most part, this is what we observed. For example,

compare the triplets of compounds with benzyloxy

substituents (15, 63, and 68), methoxy substituents (16,

60, and 64), and 1,3-difluoroisopropoxy substituents (26,

61, and 65), where in each trio, the side chain at the

7-position of the quinazoline is altered. In each case,

there is a minimal change in the enzyme inhibitory

activity resulting from altering the side chain.

Since the molecules discussed herein are reactive, one

might wonder whether they will lack selectivity by

acting as indiscriminant alkylators of other kinases. In

fact, this appears not to be the case. Two compounds of

moderate reactivity, 60 and 61, were evaluated against

a panel of other kinases consisting of AKT, BTK, CDK4,

EGFR, GSK, IGFR, IKK, ITK, LCK, LYN, Mek, MK2,

mTor, PDK1, PI3K, PKC-θ, and Tpl2. In these assays,

IC₅₀values for 60 varied between 2.2 and >200 µM and

those for 61 varied between 1.8 and >117 µM. For both

compounds, in the majority of the assays (14 out of 17),

the IC₅₀values were g10 µM. While for each compound,

Additional inhibitors that have unsaturated and

lipophilic oxy substituents at the 5-position of the

quinone ring, such as 28, 40, 62, 66, and 70-73, are

also potent VEGFR-2 inhibitors with IC₅₀values in the

range 1.5-70 nM, and almost all of these compounds

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7573

Table 4. MALDI Mass Spectral Results for Incubation of 16

and 18 with VEGFR-2^a

obsd mass

(Da)

D mass

(Da)

no. of

molecules

compd

VEGFR-2

18

16

63 897

65 112

64 462

1215

565

2.3-3.4

1.2

^aVEGFR-2 was incubated with a 10-fold excess of the inhibitor

for 1 h.

Table 5. HPLC-ESI Mass Spectral Results from a Time

Course Study of 16 Incubated with VEGFR-2 over a 20-h

Period^a

incubation

time (h)

obsd mass

(Da)

D mass

(Da)

no. of

Figure 4. The effect of 61 on the growth of DLD1 cells

implanted into athymic mice and staged to ∼100 mg. Mice

were treated with 100 mg/kg of 61 (po/b.i.d.), vehicle (Methocel/

Tween, po/b.i.d.) or 100 µg/dose bevacizumab (anti-VEGF

antibody, ip, days 1, 5, 9, 13). Relative tumor growth was

determined every 7 days for a period of 21 days. Data are the

mean fold increase in tumor volume in each group. *p e 0.05

by Student’s t-test.

molecules

0

64 134 ( 31

64 608 ( 79

64 896 ( 70

64 845 ( 49

2.5

5

475

763

712

1.2 ( 0.22

2.0 ( 0.20

1.9 ( 0.15

20

^aVEGFR-2 was incubated with a 10-fold excess of 16.

was only about 2-fold, and for some compounds, the

increase was negligible. This is exactly the result one

would expect for an irreversible inhibitor of the enzyme.

Additional, more definitive, evidence that our com-

pounds are functioning as covalent-binding inhibitors

of VEGFR-2 kinase was gathered from mass spectral

studies, where the VEGFR-2 protein was incubated with

the inhibitors 18 and 16, the former representing the

very reactive and the latter representing the moderately

reactive types of inhibitors. In addition to VEGFR-2, we

utilized the proteins BSA and HSA as controls. Table 4

summarizes the results of MALDI-MS studies with

these compounds. The MALDI mass spectrum of our

sample of the intracellular domain of human VEGFR-2

showed a broad peak corresponding to a protein with a

mass of 63 897 Da. Incubation of VEGFR-2 with a 10-

fold excess of the reactive inhibitor 18 for 1 h resulted

in the mass increase shown in Table 4. Depending on

whether the alkylations of the protein resulted from

reductive addition and/or ipso-substitution (products A

and/or B in Scheme 6), the observed mass increase

corresponds to 2.3-3.4 inhibitor molecules being added

to the protein mass. We found that 18 also reacted with

the control proteins BSA and HSA, as determined from

the mass spectra. Evidently, 18 was indiscriminate in

its reactions with sulfhydryl groups, as was evident by

its reaction at multiple sites on the target VEGFR-2

protein and its reaction with the proteins BSA and HSA,

which presumably do not have a true binding site for

the inhibitor.

In contrast, the more moderately reactive inhibitor

16 was more discriminating. In the MALDI experiment,

we observed, after 1 h incubation with VEGFR-2, a mass

increase corresponding to the addition of 1.2 molecules,

assuming reductive addition (product A in Scheme 6).

Additionally, we found that 16 did not react with the

control proteins BSA or HSA, as determined from the

mass spectra.

We also studied the reaction between an excess of 16

and VEGFR-2 by HPLC-ESI-MS as a function of time.

While the protein peaks showed poor resolution, we still

were able to extract the data shown in Table 5. These

data suggest that there may be up to two molecules of

16 reacting with VEGFR-2 over a 5 h incubation period.

Extending the incubation time beyond 5 h resulted to

low µM activity was observed in inhibiting LCK and

PI3K, 60 was still 5-6 times more potent and 61 was

120-500 times more potent in inhibiting VEGFR-2

compared to LCK and PI3K, respectively. In many of

the above assays, thiol reagents were used as part of

the buffer mixture. It is therefore not entirely clear what

influence these thiol reagents may have had on these

selectivity measurements.

In Vivo Efficacy. We selected 61 for in vivo evalu-

ation since it is a potent inhibitor of the enzyme that

also retains activity when tested in the presence of

glutathione or plasma and has reasonable potency in

the cell-based assay. Additionally, the kinetic experi-

ments with the related quinone 26 suggest it may have

a useful half-life in circulation. The compound was

evaluated in a nude mouse xenograft model bearing the

human colon carcinoma DLD1. The results are shown

in Figure 4. For comparison, data are also shown for

the VEGF antibody bevacizumab. The inhibitor was

administered, at the indicated dose, orally twice daily

for the first 20 days of the experiment. Bevacizumab

was dosed ip every fourth day. It is evident that oral

administration of 61 results in a statistically significant

reduction in tumor growth. Relative to the control

animals, there was a 54% reduction in the tumor mass

after 20 days of dosing. The efficacy of 61 in this

experiment was comparable to that exhibited by beva-

cizumab.

Evidence for Covalent Binding. We have ac-

cumulated a body of evidence that strongly suggests

that these quinone-based inhibitors bind irreversibly to

VEGFR-2 kinase. For example, these compounds appear

to behave as non-ATP-competitive inhibitors, as shown

by the data given in Table 3, where we compared our

inhibitors with the known conventional ATP-competi-

tive inhibitors 78 and 79. One would expect, for an ATP-

competitive inhibitor, that the IC₅₀value for inhibition

would increase when the concentration of ATP is

increased. The data presented in the table for 78 and

79 clearly indicate that this is so; the IC₅₀values

increased 142- and 45-fold, respectively, when changing

the ATP concentration from 10 µM to 1 mM. In contrast,

when our inhibitors were tested at the two ATP con-

centrations, the increase in the IC₅₀values, on average,

7574 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

anion exchange, GSH-affinity, another HiQ anion exchange,

and G3000 sizing columns. VEGFR-2 domain was cleaved from

N-terminal GST-His with thrombin. VEGFR-2 purity was

∼90% (MALDI-MS).

no further increase in the mass of the protein. While

one might expect that the target Cys residue (Cys-1045)

is the first to react, we have not performed an ad-

ditional, better time-resolved, kinetic experiment to

determine if the two alkylations occur at different rates

(however, see the trypsin digestion experiment below).

It is interesting to note that the rate at which 16 reacts

with VEGFR-2, in these mass spectral experiments,

appears to be significantly faster than the rate at which

it reacts with a large excess of glutathione (Table 2),

suggesting that the anticipated increase of the reaction

rate, due to the entropic effect of binding at the active

site, may be operative.

To identify the reacting Cys residues, we performed

trypsin digestion experiments. Digestion of our VEG-

FR-2 sample with trypsin followed by HPLC-MS

resulted in the identification of the T37 fragment, as

singly and doubly charged ions at 894.5 and 447.7 m/z,

respectively, corresponding to the predicted fragment

containing Cys-1045 with molecular weight of 893.5 Da.

In a parallel experiment, VEGFR-2 was first incubated

overnight with a 10-fold excess of 16. Most of the excess

drug was removed using ultrafiltration and the result-

ing isolate was digested with trypsin. Again, the frag-

ment corresponding to T37 was identified, but this time

two ions were observed at 1277.6 and 639.3 m/z,

corresponding to singly and doubly charged ions with a

molecular mass of 1276.6 Da. The mass difference

between these T37 fragments is consistent with the

reductive addition of a molecule of 16 (product A in

Scheme 6) to VEGFR-2 followed by a reoxidation to the

quinone (product D in Scheme 6) due to the manipula-

tions being carried out in air. The exact mass measure-

ment for the doubly charged ion was consistent with the

proposed elemental composition for the oxidized T37-

16 covalent complex. These results supply definitive

evidence that 16 forms a covalent product with VEG-

FR-2 and that the most significant target residue is Cys-

1045. Furthermore, by analogy to the observed solution

chemistry of 16 (Table 1), we would predict that the

inhibitor is attached to VEGFR-2 via the 3-position of

the quinone ring.

VEGFR-2 Kinase Assay. The kinase activity of recombi-

nant VEGFR-2 was evaluated using a DELFIA (dissociation-

enhanced lanthanide fluorescent immunoassay) (Perkin-Elmer

Life Sciences) essentially as described previously.³⁹Specifi-

cally, Nunc Maxisorb 96-well plates were coated at room

temperature for 1-2 h with 100 µL per well of 25 µg/mL poly-

(Glu₄-Tyr) peptide (Sigma) in Tris-buffered saline (TBS) (25

mM Tris pH 7.2, 150 mM NaCl). Unbound peptide was washed

three times with TBS. The cytoplasmic domain of VEGFR-2

enzyme was diluted (depending on the specific activity of the

batch, from 10-to 20-fold) in 0.1% BSA/4 mM HEPES. A master

mix of enzyme plus kinase buffer was prepared: (per well) 10

µL of diluted enzyme, 10 µL of 5× kinase buffer (5× ) 20 mM

HEPES, pH 7.4, 5 mM MnCl₂, 100 µM Na₃VO₄), and 9 µL of

water. Master mix (29 µL) was added to each well, and

compounds (1 µl) prepared in 100% dimethyl sulfoxide (DMSO)

were added to appropriate wells. Test compounds were added

as 50× stocks as necessary for single point or dose-response

analyses. Controls were done by adding DMSO alone, i.e., no

test compound, to wells containing the master mix of enzyme

plus kinase buffer. After 15 min at room temperature, ATP/

MgCl₂(20 µL of 25 µM ATP, 25 mM MgCl₂, 10 mM HEPES,

pH 7.4) was added to each well to initiate the reaction. Final

concentrations of the assay components were 10 µM ATP, 10

mM MgCl₂, 1 mM MnCl₂, 4 mM HEPES, pH 7.4, 20 µM Na₃-

VO₄, 20 µg/mL BSA, 2% DMSO. After 40 min at room

temperature, the liquid was removed, and plates were washed

three times with TBST (TBS with 0.05% Tween-20). The wells

were then incubated for 1 h at room temperature with 75 µL

of ∼0.1 µg/mL europium-conjugated anti-phosphotyrosine

antibody (PT66; Perkin-Elmer) prepared in assay buffer (Per-

kin-Elmer). Plates were washed three times with TBST and

then incubated for 15 min in the dark with 100 µL of

enhancement solution (Perkin-Elmer). Plates were read in a

Victor-V multilabel counter (Perkin-Elmer) using the default

europium detection protocol. Percent inhibition or IC₅₀values

of compounds were calculated by comparison with DMSO-

treated control wells. The results are shown in Table 3.

In addition to the conventional assay, the assay was

conducted in the presence of 100 µM of glutathione (Sigma)

or 5% (v/v) fresh mouse plasma using the above conditions.

For the ATP competition experiments, the assay was con-

ducted as above except that a final concentration of 1 mM ATP

was used.

Inhibition of VEGFR-2 Autophosphorylation in Cells.

Human embryonic kidney 293 cells were transfected with full

length VEGFR-2 and designated KDR15 cells.⁴²Cells were

maintained in 10% fetal calf serum (FCS) in DMEM (Life

Technologies), penicillin/streptomycin, plus 0.4 µg/mL puro-

mycin. Cells were plated in 24-well dishes (approximately 4000

cells per well) and allowed to adhere for 1 day. Compounds

prepared in DMSO were diluted into serum-free DMEM media

at appropriate final concentrations. Growth media was aspi-

rated from each well, cells were washed with serum-free

DMEM, and replaced with 0.5 mL of compound-containing

serum-free media. Cells were incubated for 1 h on ice and then

55 µL of 500 ng/mL VEGF (final concentration 50 ng/mL;

VEGF₁₆₅, R&D Systems) was added to each well and the

mixture incubated for 30 min on ice. Cells were resuspended

during VEGF incubation, transferred to 1.5-mL tubes, and

then centrifuged at 12 000 rpm for 10 min. Pellets were lysed

in 50 µL of NP40 lysis buffer (150 mM NaCl, 50 mM Tris, pH

7.5, 2 mM EDTA, 1% NP-40 [Ipegal CA-630], 1 mM Na₃VO₄,

1 mM PMSF, 20 KIU/mL aprotinin, 1 µg/mL pepstatin, 0.5

µg/mL leupeptin). Lysates were centrifuged for 10 min at

12 000 rpm at 4 °C and the supernatants transferred to fresh

tubes and frozen until use. Equal volumes of lysates were

fractionated by SDS-PAGE and transferred to PVDF mem-

branes (BioRad). Blots were blocked in 8% BSA/TBST for 1 h

at room temperature and then incubated overnight at 4°C with

The trypsin digestion experiment also showed weak

signals in the mass spectrum corresponding to alkylated

T45, the trypsin fragment predicted to contain Cys-

1116. No other alklyated trypsin fragments were ob-

served. The X-ray structure of the catalytic domain of

VEGFR-2 shows that Cys-1116 resides in a shallow

pocket located near the surface of the protein with the

sulfhydryl group pointing outward, and it should there-

fore be accessible to an electrophile in solution. The ratio

of the ion counts for the T37 and T45 fragments before

the reaction with 16 was 0.312. After the reaction, the

ratio of the ion counts for the alkylated fragments was

about 46, suggesting that 16 reacts significantly faster

with the active site Cys-1045 than it does with Cys-

1116.

Experimental Section

Expression of Recombinant VEGFR-2 Kinase. The

human VEGFR-2 cytoplasmic domain (from Val-805 to Val-

1356) was cloned by PCR from first-strand human placental

cDNA (Invitrogen), subcloned into pAcGHLT-B (Pharmingen),

and expressed in Sf9 cells. GST-His-VEGFR-2 was purified

from Sf9 cells by sequential chromatography on NiNTA, HiQ

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7575

1:1000 anti-phospho-VEGFR-2-Y996 antibody (specifically

detects phosphorylated tyrosine-996 on VEGFR-2; Cell Signal-

ing) in 4% BSA/TBST. Blots were washed with TBST, incu-

bated with secondary antibody (1:1000 HRP-conjugated goat

anti-rabbit IgG) in 5% milk/TBST, washed, detected with

enhanced chemiluminescent reagents (Amersham), and ex-

posed to film. Autoradiographs were quantified by scanning

on a FluorS imager (BioRad) and data normalized to untreated

controls. To confirm equal loading of protein, blots were

stripped in SDS/Tris at 50 °C, followed by immunoblot analysis

in 1:1000 anti-VEGFR-2 antibody in 5% milk/TBST. The

results are shown in Table 3.

These cell-based phosphorylation studies were conducted on

ice to reduce VEGFR-2 receptor internalization. Cold temper-

atures are commonly used in studies of receptor internalization

to prevent receptor turnover that may interfere with data

interpretation. Moreover, while this assay format was used

for the primary screening of compounds, follow-up experiments

were performed in these cells at 37 °C for lead compounds and

confirmed the observations that were reported at 4 °C.

In Vivo Studies. Inhibitor 61 was evaluated in vivo using

standard pharmacological test procedures, which measures the

ability to inhibit the growth of human tumor xenografts.

Human colon carcinoma DLD-1 cells (American Type Culture

Collection, Manassas, VA) were grown in tissue culture in

DMEM (Gibco/BRL, Gaithersburg, MD) supplemented with

10% FBS (Gemini Bio-Products Inc., Calabasas, CA). Athymic

nu/nu female mice (Charles River, Wilmington, MA) were

injected subcutaneously (sc) in the flank area with 6 × 10⁶

DLD-1 cells. When tumors attained a mass of between 120

and 160 mg, the mice were randomly assigned into treatment

groups containing 5-10 animals per group. Animals were

treated either orally (po) once daily with 100 mg/kg 61

prepared in 0.5% Methocel/0.4% Tween 80 or vehicle alone on

days 1-20 poststaging (day zero) or intraperitoneal (ip) on

days 1, 5, 9, and 13 with 100 or 200 µg per dose bevacizumab

prepared in 0.9% saline. Tumor mass was determined every

7 days [(length × width²)/2] for 21 days poststaging. Relative

tumor growth (mean tumor mass on days 7, 14, and 21 divided

by the mean tumor mass on day zero) was determined for each

treatment group. Statistical analysis (one-tailed Student’s

t-test) of log relative tumor growth was used to compare the

treated verses control group in each experiment. A p-value

e0.05 indicates a statistically significant reduction in relative

tumor growth of treated group compared to the vehicle control.

Kinetic Experiments. Determinations of the Half-Life

(t_1/2) of Inhibitors in the Presence of Glutathione. Solu-

tions containing 1 µM or 10 µM of the inhibitor, 10 µM of an

internal standard, and 100 µM of reduced glutathione in CH₃-

CN-H₂O (1:1) were prepared. The reaction mixtures were

maintained at ambient temperature without protection from

air. The amount of inhibitor remaining as a function of time

was monitored using an Agilent LC-1100-MSD equipped with

a Phenomenex Luna C5 5 µm column (50 × 4.6 mm). Injection

volumes were 5 µL. The mobile phase flow rate was 1.0 mL/

min. The HPLC mobile phase gradient was optimized for each

compound. For a typical example, the separation was initiated

at 25% eluent B, followed by a linear increase to 95% eluent

B over 15 min. Eluent A was 10 mM ammonium acetate in

95% water-5% CH₃CN, and eluent B was 10 mM ammonium

acetate in 5% water 95% CH₃CN. The HPLC column temper-

ature was maintained at 40 °C during the experiment.

Detection was conducted by single-ion-monitoring (SIM) of

electrosprayed positive ions of the drug and internal standard.

From the rate plot of the data and assuming pseudo-first-order

reaction kinetics, the t_1/2was determined from the equation

t_1/2) ln(2)/k, where k is the exponential rate of decay.

sinapinic acid matrix in 0.1% TFA (70:30 H₂O:CH₃CN), spotted

on the MALDI plate, slowly dried, and then analyzed.

HPLC Electrospray Ionization Mass Spectrometry

(HPLC-ESI-MS). Electrospray ionization mass spectra were

obtained in the positive ionization mode with a Waters-

Micromass LCT electrospray time-of-flight mass spectrometer.

The samples were analyzed with a capillary voltage of 3500 V

and cone voltage of 30 V. N₂was used as the nebulizer and

desolvation gas with flow rates of 275 and 675 L/h, respec-

tively. The desolvation temperature was set to 180 °C and the

source temperature to 120 °C. The appropriate scans were

averaged, smoothed, baseline subtracted, and for protein data

transformed using the Micromass transform or maximum

entropy programs. The solvent gradients used were CH₃CN:

H₂O from 5:95 to 95:5 (v/v) containing 0.2% formic acid. For

protein-drug reaction assays, a Waters Symmetry 300 C4

column was used (1 × 150 mm, 5 µm particles, Cat #

186000276) with a solvent flow rate of 50 µL/min for 1 h. For

protein digest assays, a Waters Symmetry 300 C18 column

was used (1 × 150 mm, 5 µm particles, Cat # 186000185) with

a solvent flow rate of 65 µL/min for 90 min. Exact mass

measurements were achieved by use of buffer ions and trypsin

fragment ions found within the retention time window of the

unknown as internal reference ions.

Protein-Drug Reactions for Mass Spectrometry.

Samples of VEGFR-2, BSA, and HSA (3-5 µM) were each

incubated with the inhibitors at a molar ratio of 1:10 at 25

°C. The reaction products were sampled after 1 h for the

MALDI-MS experiments. For the ESI-MS kinetics experiment,

VEGFR-2 samples were analyzed at 0, 2.5, 5, and 20 h

incubation times.

Trypsin Proteolysis Reactions for Mass Spectrometry.

Sequencing grade trypsin was purchased from Promega (Madi-

son, WI, Cat # V511A) and used to cleave the peptide bonds

C-terminally at lysine and arginine residues. For the VEGFR-2

control, 20 µL of VEGFR-2 (3.8 µM, prepared at pH of 7.5 in

25 mM HEPES, 75 mM NaCl and 30% glycerol) was incubated

with 1.25 µL of 0.1 µg/µL trypsin and 2 µL DMSO and the

mixture incubated overnight at 25 °C. The molar ratio of

VEGFR-2 to trypsin was 15:1. For the reaction of VEGFR-2

with 16, 50 µL of VEGFR-2 was mixed with 1.5 µL of 2.28

mM of 16 (in DMSO) and 3.5 µL DMSO, which reacted

overnight at 25 °C. The molar ratio of VEGFR-2 to 16 was

1:10. To remove the excess drug, the reaction products were

washed four times with ∼100 µL of 4 mM HEPES solution

(pH 8) using a Microcon ultrafiltration filter with a 10 kDa

cutoff membrane (Cat. # YM-10). The filtrate was discarded,

and the retentate was treated with 3 µL of 0.1 µg/µL trypsin,

incubated overnight at 25 °C, and then analyzed by HPLC-

ESI-MS.

Molecular Modeling. The inhibitor structures were mini-

mized using the MMFF94 force field in Sybyl⁴³and conforma-

tionally expanded using Omega⁴⁴with a 1.0 Å rmsd cutoff to

filter out degenerate conformers. A public domain crystal

structure of VEGFR-2²⁵was used as the 3-D representation

for molecular docking studies. FRED³⁷was used for the

docking studies. Briefly, FRED uses a shape-based matching

of all possible orientations of each conformer and optimizes

in accordance with the selected scoring function. The PLP, or

piece-wise linear potential function, was used to score the poses

docked using FRED, consisting of a steric matching function

with a less stringent charge matching function of protein-

ligand interactions. The top 10 poses obtained from the FRED

docking procedure were rescored with a Poisson Boltzmann

molecular mechanics scoring function (PBMM-SA),³⁸where the

free energy of protein-ligand binding is described as

PL

∆G_bind^solv) ∆G_bind^gas- ∆G_solv^P- ∆G_solv^L+ ∆G_solv

Matrix-Assisted Laser Desorption/Ionizaton Mass Spec-

trometry (MALDI-MS). The MALDI experiments were per-

formed using an Applied Biosystems Voyager Elite STR

MALDI mass spectrometry system equipped with a nitrogen

laser and delayed extraction. About 0.25-0.6 µL of the reaction

products was mixed with 1 µL of freshly prepared saturated

where the first term to the right of the equal sign is the gas-

phase protein-ligand binding energy and is obtained via

molecular mechanics calculations. The remainder of the terms

represent the solvated free energy terms for the protein and

ligand individually as well as the protein-ligand complex. The

7576 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

solvated free energy terms are evaluated explicitly via Poisson

Boltzmann/surface area calculations. Solvation and hydropho-

bic effects are aptly handled by this scoring function, which is

used to rescore or fine-tune the docked views.

nm) and a flow rate of 1 mL/min was employed. For system

E, a gradient elution of increasing concentrations of CH₃CN

in H₂O containing 0.02% TFA (40-90% over 20 min, wave-

length 230 nm) and a flow rate of 1 mL/min was employed.

For system F, a gradient elution of increasing concentrations

of CH₃CN in H₂O containing 0.02% TFA (10-90% over 20 min,

wavelength 254 nm) and a flow rate of 1 mL/min was

employed. For system G, a gradient elution of increasing

concentrations of CH₃CN in H₂O containing 0.02% TFA (20-

60% over 20 min, wavelength 230 nm) and a flow rate of 1

mL/min was employed. Preparative HPLC was conducted on

a Gilson reverse phase HPLC system over a 30 mm × 150 mm

Luna C₁₈(2) column (10 µm, 100 Å) using UV wavelength

detection at 254 or 215 nm. A gradient elution of increasing

concentrations of CH₃CN in H₂O containing 0.02% TFA (20-

80% over 20 min) and a flow rate of 40 mL/min was employed.

3-Methoxy-4-(2-methoxyethoxy)benzonitrile (3). A so-

lution of 4-cyano-2-methoxyphenol (522 g, 3.50 mol) in anhy-

drous DMF (2.0 L) was stirred and cooled with a H₂O bath.

K₂CO₃(725 g, 5.25 mol) was added in portions and the mixture

was stirred at 20 °C for 1 h. Bromoethyl methyl ether (427.5

mL, 632.3 g, 4.55mol) was added dropwise over 1 h. The

resulting mixture was stirred at room temperature overnight

and then heated at 34-40 °C for 6 h. TLC showed no more

starting material remaining. The mixture was poured into a

mixture of ice/H₂O (5 kg) and the product precipitated. After

stirring for 10 min, the mixture was filtered. The solid was

washed with H₂O (3 × 1.5 L) and air-dried to give 3 (690.0 g,

95%) as a beige solid: MS (ESI, m/z) 208.1 (M + H)⁺; ¹H NMR

(400 MHz, CDCl₃) δ 3.45 (s, 3 H), 3.75-3.84 (m, 2 H), 3.88 (s,

3 H), 4.09-4.29 (m, 2 H), 6.93 (d, J ) 8.31 Hz, 1 H), 7.08 (d,

J ) 1.76 Hz, 1 H). Anal. (C₁₁H₁₃NO₃) C, H, N.

Semiempirical Molecular Orbital Calculations. Inhibi-

tor structures optimized as described above were put into a

Sybyl database. It should be noted that the structures were

inspected to determine the quinazoline-NH-quinone rotomer.

In accordance with the binding mode determined from the

structure-based molecular modeling studies, rotomers that

favor intramolecular interactions between the quinone carbo-

nyl and the linker NH were chosen. Using the MMFF94 force

field within Sybyl, a Powell minimization carried out to an

energy gradient of 0.005 kcal/mol Å was performed on each

structure in the database. The orbital energies of the lowest

unoccupied molecular orbitals (LUMO) were computed for all

of the quinones that were prepared using the MOPAC software

package interfaced through Sybyl under the AM1 paradigm.

No geometry optimization was done at this stage. For a select

set of 13 inhibitors (Table 2), the conformational dependence

of molecular orbital calculations was evaluated. Using the

Sybyl software package, a systematic grid search of the

dihedral angles pertaining to the quinazoline-NH-quinone

and those on the quinone substituents was carried out at 30°

and/or 60° intervals with Powell energy minimization using

the MMFF94 force field (energy gradient e 0.005 kcal/mol Å)

at each step of the grid search. The set of resulting conformers

and energies was imported into a spreadsheet and sorted in

order of ascending energies, and Boltzmann averaging of the

computed LUMO energies was carried out.⁴⁵The Boltzmann

weight, W_i, for each of the N conformers was computed as

follows:

5-Methoxy-4-(2-methoxyethoxy)-2-nitrobenzonitrile (4).

To a stirred solution of 3 (16.7 g, 80.6 mmol) in 100 mL of

trifluoroacetic anhydride and 70 mL of chloroform was added

solid ammonium nitrate (9.7 g,120.9 mmol) portionwise over

10 min. The mixture warmed to a gentle boil during the

addition. After 2 h, the mixture was diluted with hexanes and

the solid was collected. The solid was washed with hexanes,

H₂O, dilute NaHCO₃solution, and then H₂O again. This solid

was air-dried to yield 18.4 g of 4 as a light yellow solid: MS

e^{-(E -E )/KT}

i

0

W_i)

N

e^{-(E -E )/KT}

i

0

∑

i)1

The computed LUMO energy for each conformation, L_i, was

then weighted and summed over all N conformers to give the

Boltzmann averaged LUMO energy as follows:

1

(ESI, m/z) 251.97 (M + H)⁺; H NMR (400 MHz, DMSO-d₆) δ

3.20-3.39 (s, 3 H), 3.58-3.78 (m, 2 H), 3.98 (s, 3 H), 4.24-

4.43 (m, 2 H), 7.70 (s, 1 H), 7.91 (s, 1 H). Anal. (C₁₁H₁₂N₂O₅)

C, H, N.

N

LUMO )

W_iL_i

∑

i)1

2-Amino-5-methoxy-4-(2-methoxyethoxy)benzoni-

trile (5). A mixture of 4 (717 g, 2.84 mol) and Pd/C (10%, 45

g) in anhydrous ethanol (6 L) was stirred and heated under

reflux. Cyclohexene (1440 mL, 1168 g, 14.2 mol) was added

dropwise over 3.5 h. The mixture was refluxed overnight until

no more starting material was present by TLC. The mixture

was cooled to 40 °C, filtered through a Magnesol pad, and

washed with ethanol (4 × 500 mL). The filtrate was concen-

trated to give a solid. The solid was collected by filtration,

washed with ethanol (2 × 500 mL), and dried under nitrogen

to give a crude product (520 g, 93.7% pure on HPLC). The

crude product was suspended in ethanol (750 mL), stirred at

40 °C for 30 min, cooled to room temperature, filtered, washed

with ethanol (300 mL), and dried under nitrogen to produced

pure 5 (498.2 g, 79%, 97.5% on HPLC) as a yellow solid: mp

Chemistry. ¹H NMR spectra were determined with a

Bruker DRX400 spectrometer at 400 MHz. Chemical shifts,

δ, are in parts per million relative to the internal standard

tetramethylsilane. Electrospray (ES) mass spectra were re-

corded in positive mode on a Micromass platform spectrometer.

Electron impact (EI) and high-resolution mass spectra (HRMS)

were obtained on a Finnigan MAT-90 spectrometer. Some

high-resolution electrospray mass spectra with higher preci-

sion were obtained on a Bruker 9.4T FTMS spectrometer.

Chromatographic purifications were by flash chromatography

using Baker 40 µm silica gel. Melting points were determined

in open capillary tubes on a Meltemp melting point apparatus

and are uncorrected. Analytical HPLC was conducted on an

HP 1090 liquid chromatography system over a 4.6 mm × 150

mm YMC ODS-A column (5 um, 120 Å) using multiple

wavelength UV detection (typically 230 and 254 nm). For

system A, a gradient elution of increasing concentrations of

CH₃CN in H₂O containing 0.02% TFA (30-90% over 20 min,

wavelength 230 nm) and a flow rate of 1 mL/min was

employed. For system B, a gradient elution of increasing

concentrations of methanol in H₂O containing 0.02% TFA (70-

90% over 20 min, wavelength 230 nm) and a flow rate of 1

mL/min was employed. For system C, a gradient elution of

increasing concentrations of CH₃CN in H₂O containing 0.02%

TFA (40-90% over 20 min, wavelength 254 nm) and a flow

rate of 1 mL/min was employed. For system D, a gradient

elution of increasing concentrations of methanol in H₂O

containing 0.02% TFA (70-90% over 20 min, wavelength 254

1

102-103 °C; MS (ESI, m/z) 223.15 (M + H)⁺; H NMR (400

MHz, DMSO-d₆) δ 3.20-3.41 (m, 3 H), 3.57-3.73 (m, 5 H),

3.92-4.14 (m, 2 H), 5.59 (s, 2 H), 6.33-6.47 (m, 1 H), 6.83-

6.95 (m, 1 H).

N′-[2-Cyano-4-methoxy-5-(2-methoxyethoxy)phenyl]-

N,N-dimethylformamidine (6). A suspension of 5 (495.4 g,

2.229 mol) in DMF-DMA (1.1 L, 8.28 mol) was stirred, heated

under reflux for 7 h, and cooled to room temperature. The solid

was collected, washed with ether (2 × 500 mL), and dried

under nitrogen at room temperature to give 6 (593.1 g, 96%)

as a yellow solid: mp 149-150 °C; MS (ESI, m/z) 278.16 (M +

1

H)⁺; H NMR (400 MHz, CDCl₃) δ 3.06 (s, 6 H), 3.44 (s, 3 H),

3.76-3.81 (m, 2 H), 3.82 (s, 3 H), 4.14-4.22 (m, 2 H), 6.50 (s,

1 H), 6.93 (s, 1 H), 7.55 (s, 1 H).

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7577

N-(4-Chloro-2,5-dimethoxyphenyl)-6-methoxy-7-(2-

methoxyethoxy)quinazolin-4-amine (8a). A solution of 6

(54.8 g, 0.198 mol) and 7a (40.78 g, 0.218 mol) in AcOH (246

mL) was heated to 110 °C for 1 h. The reaction was cooled to

room temperature and diluted with ether. The resulting solid

was collected by filtration to give 82 g (99%) of 8a: mp 214-

217 °C; MS (ESI) m/z 420.1 (M + H)⁺; ¹H NMR (400 MHz,

CDCl₃) δ 3.49 (s, 3 H), 3.84-3.92 (m, 2 H), 3.95 (s, 3 H, 3.97

(s, 3 H), 4.04 (s, 3 H), 4.23-4.37 (m, 2 H), 6.98 (s, 1 H), 7.02

(s, 1 H), 7.27 (s, 1 H), 7.88 (s, 1 H), 8.71 (d, J ) 4.03 Hz, 2 H).

2-Chloro-5-{[6-methoxy-7-(2-methoxyethoxy)quinazo-

lin-4-yl]amino}benzo-1,4-quinone (9a). Compound 8a (8.57

g, 20.4 mmol) was dissolved in a hot CHCl₃(100 mL), diluted

with CH₃CN (256 mL) and H₂O (38 mL), and heated to boiling.

While still hot the solution was treated with Ce(NH₄)₂(NO₃)₄

(CAN) (27.9 g, 51.0 mmol) in portions over 20 min. The reaction

was then stirred for 1 h, diluted with H₂O (300 mL), and

extracted with CHCl₃(5 × 800 mL). The organic solution was

dried (MgSO₄) and filtered through a pad of Magnesol (eluted

with 10% CH₃OH in EtOAc). The solvent was removed at

reduced pressure. The resulting solid was dissolved in boiling

CH₃CN (200 mL) and diluted with ether (200 mL). A red solid

formed upon cooling and was collected by filtration (5.49 g,

69%): mp 224-226 °C; MS (ESI) m/z 390.1 (M + H)⁺; ¹H NMR

(400 MHz, CDCl₃), δ 3.49 (s, 3 H), 3.81-3.97 (m, 2 H), 4.07 (s,

3 H), 4.26-4.41 (m, 2 H), 7.03 (s, 1 H), 7.10 (s, 1 H), 7.33 (s, 1

H), 8.29 (s, 1 H), 8.49 (s, 1 H), 8.83 (s, 1 H). Anal. (C₁₈H₁₆-

ClN₃O₅) C, H, N.

added. The suspension was stirred at room temperature

overnight. The solvent was removed to give a yellow brown

solid, which was dissolved in 50 mL of CH₃CN and 10 mL of

H₂O. DDQ (1.19 g, 5.24 mmol) was added over 10 min followed

by 823 mg (10 mmol) of sodium acetate. The reaction mixture

was stirred at room temperature for 1 h, poured into saturated

NaHCO_3,and extracted (3×) with CH₂Cl₂. The organic layer

was dried over MgSO₄and concentrated. The residue was

purified on a silica gel column, eluting with CH₂Cl₂, EtOAc,

and 2-propanol-EtOAc. The solvent was removed from prod-

uct fractions. The resulting solid was heated in CH₃CN and

diluted with ether. After cooling, the solid was collected to yield

1.2 g (57%) of 9i as a yellow powder: MS (ESI) m/z 420 (M +

H)⁺; ¹H NMR (400 MHz, DMSO-d₆) δ 3.37 (s, 3 H), 3.80-3.82

(m, 2 H), 3.92 (s, 3 H), 4.04 (s, 3 H), 4.38-4.40 (m, 2 H), 6.51

(s, 1 H), 7.38 (s, 1 H), 8.19 (s, 1 H), 8.92 (s, 1 H). Anal. (C₁₉H₁₈-

ClN₃O₆) C, H, N. The regiochemistry of 9i was confirmed by a

NOESY NMR experiment.

5-Methoxy-3-{[6-methoxy-7-(2-methoxyethoxy)quinazo-

lin-4-yl]amino}-2-(pyridin-2-ylthio)benzo-1,4-quinone (9j).

To a solution of 9h (359 mg, 0.93 mmol) in 12 mL of CHCl₃

was added 2-mercaptopyridine (113 mg, 1.02 mmol). After 30

min, DDQ (0.21 g, 1.0 mmol) and 6 mL of CH₃CN were added.

After stirring for 30 min, the mixture was poured into dilute

Na₂CO₃and extracted with CH₂Cl₂. The solution was dried

(MgSO₄) and poured unto a column of Magnesol. The product

was eluted with CH₂Cl₂-2-propanol mixtures. Solvent was

removed and the residue was recrystallized from CH₃CN-

ether to yield 259 mg (56.3%) of 9j as an orange powder: MS

(ESI) m/z 495 (M + H)⁺, 248 (M + 2H)²⁺; ¹H NMR (400 MHz,

CDCl₃) δ 3.48 (s, 3 H), 3.89-3.90 (m, 5 H), 4.00 (s, 3 H), 4.31-

4.33 (m, 2 H), 6.08 (s, 1 H), 7.01 (dd, J ) 6.55, 5.04 Hz, 1 H),

7.07 (bs, 1 H), 7.20-7.24 (m, 1 H), 7.30-7.33 (m, 1 H), 7.47-

7.52 (m, 1 H), 8.33-8.40 (m, 2 H), 8.69 (bs, 1 H) Anal.

(C₂₄H₂₂N₄O₆S) C, H, N.

N-[2,5-Dimethoxy-4-(methylthio)phenyl]-6-methoxy-7-

(2-methoxyethoxy)quinazolin-4-amine (8l). Compounds 6

(885 mg, 3.19 mmol) and 7l⁴⁶(700 mg, 3.51 mmol) were heated

to 110 °C in AcOH (4 mL) for 3 h. The reaction was partitioned

in H₂O-EtOAc. The solid precipitates were collected and

washed with H₂O and EtOAc. The product was purified on a

silica gel column, eluting with 2.5% CH₃OH-CH₂Cl₂to yield

485 mg (35%) of 8l as a pink solid: MS (ESI) m/z 432.1 (M +

H)⁺; HRMS (ESI-FTMS (M + H)⁺) calcd for C₂₁H₂₅N₃O₅S

432.15877, found 432.15853; ¹H NMR (400 MHz, CDCl₃) δ

2.44-2.47 (m, 3 H), 3.47-3.50 (m, 3 H), 3.84-3.92 (m, 2 H),

3.97 (s, 6 H), 4.01-4.10 (m, 3 H), 4.28-4.38 (m, 2 H), 6.93 (s,

1 H), 7.03 (s, 1 H), 7.28-7.32 (m, 1 H), 8.55 (s, 1 H), 8.71 (s, 1

H).

2-(Dimethylamino)-5-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}benzo-1,4-quinone (10) (Method

A). A mixture of 974 mg (2.5 mmol) of 9a and 5 mL (10 mmol)

of dimethylamine solution (2 M in THF) was added to 289 mg

(2.5 mmol) of pyridine hydrochloride in 15 mL of THF. The

reaction was stirred at room temperature for 3 h. The resulting

solid was filtered, washed with H₂O, and dried to give 936 mg

(94%) of 10 as a light brown solid: mp 221-224 °C; MS (ESI)

m/z 399.2 (M + H)⁺; HRMS (ESI-FTMS, (M + H)⁺) calcd for

C₂₀H₂₂N₄O₅399.16630, found 399.16664; ¹H NMR (400 MHz,

CDCl₃) δ 1.57 (s, 3 H), 3.30 (s, 3 H), 3.48 (s, 3 H), 3.80-3.95

(m, 2 H), 4.06 (s, 3 H), 4.30-4.37 (m, 2 H), 5.63 (s, 1 H), 7.13

(s, 1 H), 7.31 (s, 1 H), 7.79 (s, 1 H), 8.82 (s, 1 H), 9.15 (s, 1 H).

Anal. (C₂₀H₂₂N₄O₅) C, H, N.

3-Chloro-2-methoxy-5-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}benzo-1,4-quinone (9c). A solution

of 9b (479 mg, 1.13 mmol) in 60 mL of CH₂Cl₂was stirred as

a solution of NaOPh(3H₂O) (211.49 mg, 1.24 mmol) in 40 mL

of CH₃OH was added. The reaction was stirred at room

temperature for 1.5 h. It was diluted with CH₂Cl₂and washed

with dilute K₂CO₃. The organic layer was separated and dried

over MgSO₄. The solvent was removed in vacuo. The residue

was chromatographed on Magnesol, eluting with CHCl₃-CH₃-

OH mixtures. The solvent was removed. The residue was

boiled in CH₃CN and cooled. The resulting solid was collected

to yield 170 mg (35.8%) of 9c as a dark red crystalline solid:

1

MS (ESI) m/z 420.1 (M + H)⁺; H NMR (400 MHz, CDCl₃) δ

3.49 (s, 3 H), 3.88-3.90 (m, 2 H), 4.06 (s, 3 H), 4.33-4.34 (m,

2 H), 4.39 (s, 3 H), 7.05 (s, 1 H), 7.32 (s, 1 H), 7.99 (s, 1 H),

8.75 (s, 1 H), 8.83 (s, 1 H). Anal. (C₁₉H₁₈ClN₃O₆·0.33H₂O) C,

H, N.

2-Chloro-3-methoxy-6-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}-5-(methylthio)benzo-1,4-quinone

(9d). CH₃SH was bubbled into 50 mL of CH₂Cl₂for 10 min.

Solid 9c (500 mg, 1.19 mmol) was added and the reaction

mixture was stirred at room temperature for 1.5 h. The solvent

was removed in vacuo. The residue was dissolved in 50 mL of

hot CH₃CN. DDQ (297.4 mg, 1.31 mmol) was added while the

solution was still hot. The reaction mixture was stirred for 10

min and then diluted with 400 mL of CH₂Cl₂. The mixture

was washed with dilute K₂CO₃. The organic layer was sepa-

rated, dried over MgSO₄, and passed through a column of

Magnesol. The solvent was removed in vacuo. The residue was

chromagraphed on silica gel, eluting with CH₂Cl₂, CH₂Cl₂-

EtOAc, and CH₂Cl₂-EtOAc-CH₃OH. Solvent was removed

from product fractions. The residue was treated with ether.

The resulting solid was collected and washed with ether to

yield 356 mg (64.2%) of 9d as a blue-black powder: MS (ESI)

m/z 466.1 (M + H)⁺; MS (ESI) m/z 931.1 (2M + H)¹; HRMS

(ESI-FTMS (M + H)⁺calcd for C₂₀H₂₀ClN₃O₆S 466.08341,

found 466.08433; the purity of 9d was evaluated by two HPLC

systems and found to be 100% (system C, t_R) 2.74 min) and

2-{[6-Methoxy-7-(2-methoxyethoxy)quinazolin-4-yl]ami-

no}-5-morpholin-4-ylbenzo-1,4-quinone (11) (Method B).

To the solution of 1.13 g (2.5 mmol) of 9a in 30 mL of THF

was added 1 mL (11.5 mmol) of morpholine. The reaction was

stirred at room temperature for 3 h. The resulting solid was

filtered, washed with THF and H₂O, and dried to give 1.07 g

(97.1%) of 11 as a light red solid: mp 239-243 °C; MS (ESI)

m/z 441.2 (M + H)⁺; HRMS (ESI-FTMS (M + H)⁺) calcd for

C₂₂H₂₄N₄O₆441.17686, found 441.17668; the purity of 11 was

evaluated by two HPLC systems and found to be 93% (system

1

96% (system D, t_R) 8.63 min); H NMR (400 MHz, CDCl₃) δ

2.32 (s, 3 H), 3.48 (s, 3 H), 3.88-3.90 (m, 2 H), 4.04 (s, 3 H),

4.32-4.35 (m, 5 H), 7.13 (s, 1 H), 7.32 (s, 1 H), 8.01 (bs, 1 H),

8.74 (s, 1 H).

2-Chloro-5-methoxy-3-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}benzo-1,4-quinone (9i). A solution

of 9h (2.0 g, 5.19 mmol) was prepared by heating in 50 mL of

CHCl₃. A solution of 25 mL of CHCl₃saturated with HCl was

1

C, t_R) 2.94 min) and 100% (system D, t_R) 12.47 min); H

7578 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

NMR (400 MHz, DMSO-d₆) δ 3.06-3.15 (m, 4 H), 3.64 (s, 3

H), 3.68-3.79 (m, 6 H), 4.02 (s, 3 H), 4.27-4.37 (m, 2 H), 7.35

(s, 1 H), 7.46 (s, 1 H), 7.64 (s, 1 H), 8.65-8.78 (bs, 1 H), 8.79

(s, 1 H), 9.27 (s, 1 H).

(s, 1 H), 8.08 (s, 1 H), 8.67 (s, 1 H), 8.82 (s, 1 H). Anal.

(C₂₁H₂₁F₂N₃O₆‚0.33 H₂O) C, N, H: calcd, 4.38; found, 4.8.

2-[(2-Hydroxypropyl)amino]-5-{[6-methoxy-7-(2-meth-

oxyethoxy)quinazolin-4-yl]amino}benzo-1,4-quinone (41).

A solution of 21 (250 mg, 0.64 mmol) in 250 mL of THF and

250 mL of H₂O was adjusted to pH 4 with concentrated HCl

and stirred for 18 h. The aqueous solution was extracted three

times with 150 mL of CH₂Cl₂. The CH₂Cl₂layer was separated,

dried over Na₂SO₄, and concentrated. The residue was chro-

matographed on a short silica gel pad, eluting with 7%

2-propanol-CH₂Cl_2.The solvent of the product fraction was

evaporated to give 190 mg (69%) of 41: MS (ESI) m/z 429.2

(M + H)⁺; HRMS (ESI-FTMS (M + H)⁺) calcd for C₂₁H₂₄N₄O₆

429.17686, found 429.17515; ¹H NMR (400 MHz, DMSO-d₆) δ

1.10 (d, J ) 8 Hz, 3 H), 3.04-3.13 (m, 1 H), 3.14-3.22 (m, 1

H), 3.34 (s, 3 H), 3.72-3.78 (m, 2 H), 3.82-3.93 (m, 1 H), 4.02

(s, 3 H), 4.30-4.34 (m, 2 H), 4.97 (d, J ) 5.04 Hz, 1 H), 5.60

(s, 1 H), 7.35-7.41 (m, 2 H), 7.62-7.68 (m, 2 H), 8.81 (s, 1 H),

9.51 (s, 1 H). Anal. (C₂₁H₂₄N₄O₆·0.25H₂O) C, H, N.

2-{[6-Methoxy-7-(2-methoxyethoxy)quinazolin-4-yl]ami-

no}-5-(5-methyl-2-oxo-1,3-oxazolidin-3-yl)benzo-1,4-quin-

one (42). A mixture of 41 (130 mg, 0.30 mmol) and carbon-

yldiimidazole (488 mg, 3 mmol) in 10 mL of dry 1-methyl-2-

pyrrolidinone was stirred under N₂for 27 h. The mixture was

poured into H₂O and extracted with EtOAc and dried. The

solvent was removed and the residue was chromatographed

on a silica gel column, eluting with 4% 2-propanol-CH₂Cl₂.

The solvent from the product fractions was evaporated to give

55 mg (42%) of 42: MS (ESI) m/z 455.1 (M + H)⁺; HRMS

(ESI-FTMS (M + H)⁺) calcd for C₂₂H₂₂N₄O₇455.15613, found

455.15478; ¹H NMR (400 MHz, CDCl₃) δ 1.52 (d, J ) 6.04 Hz,

4 H), 3.49 (s, 3 H), 3.87-3.91 (m, 2 H), 4.08 (s, 3 H), 4.32-

4.35 (m, 2 H), 4.54 (dd, J ) 10.95, 7.93 Hz, 1 H), 4.69-4.79

(m, 1 H), 7.07 (s, 1 H), 7.31-7.33 (s, 1 H), 7.60 (s, 1 H), 8.00

(s, 1 H), 8.68 (s, 1 H), 8.83 (s, 1 H). Anal. (C₂₂H₂₂N₄O₇₎C, H,

N.

2-Chloro-4-hydroxy-3-methoxy-5-nitrobenzaldehyde

(45). To a stirred solution of 43³¹(21.33 g, 77.95 mmol) and

dimethyl sulfate (90 mL, 0.952 mol) in EtOH (192 mL) at 40

°C was added a 40% solution of KOH (140 mL, 98.2 mol)

dropwise over 45 min. The reaction was then stirred at 55 °C

for 1 h. The solvent was evaporated and the residue was

extracted with ether (2×). The ether solution was dried

(MgSO₄) and was passed through a column of Magnesol. The

solvent was removed to give 22.3 g (90.91 mmol)) of 44 as a

colorless oil. This was mixed with H₂O (1.12 mL) and LiCl

(23.12 g, 0.545 mol) in 123 mL of DMF and heated at 110 °C

for 3 h. The dark red mixture was cooled and treated with a

solution of saturated NaHCO₃(59 mL) and H₂O (800 mL). The

aqueous solution was washed with ether (2×), made acidic

with H₂SO₄, and cooled to 4 °C. The resulting solid was

collected by vacuum filtration, washed with H₂O, and dried

in air to give 18.3 g (87%) of 45 as an off white solid: MS (ESI)

m/z 230 (M + H)⁺; ¹H NMR (400 MHz, CDCl₃) δ 4.01 (s, 3 H),

8.54 (s, 1 H), 10.35 (s, 1 H), 11.14 (s, 1 H).

2-Chloro-3-methoxy-5-nitrobenzene-1,4-diol (46). To

compound 45 (17.8 g, 72.47 mmol) were added 1 N NaOH (72.5

mL, 72.5 mmol), H₂O (158 mL), 30% H₂O₂(45 mL), and CH₃-

OH (158 mL). The mixture was stirred at 50 °C for 3.5 h. The

CH₃OH was evaporated and the solution was cooled in an ice

bath. The resulting solid was collected by vacuum filtration,

washed with H₂O, and air-dried to give 7.9 g (50%) of 46 as

an orange solid: MS (ESI) m/z 218 (M + H)⁺;¹H NMR (400

MHz, CDCl₃) δ 4.01 (s, 3 H), 5.43 (s, 1 H), 7.56 (s, 1 H), 10.37

(d, J ) 11.33 Hz, 1 H).

2-{[6-Methoxy-7-(2-methoxyethoxy)quinazolin-4-yl]ami-

no}-5-phenoxybenzo-1,4-quinone (13) (Method D). To a

solution of 1.13 g (2.5 mmol) of 9a in 25 mL of acetone were

added 1.0 g (10.51 mmol) of phenol and 1 g (7.17 mmol) of

potassium carbonate. The reaction was stirred at room tem-

perature overnight. The resulting solid was filtered, washed

with H₂O, and dried to give 1.058 g (91.7%) of 13 as a dark

pink solid: mp 177-180 °C; MS (ESI) m/z 448.2 (M + H)⁺; ¹H

NMR (400 MHz, DMSO-d₆) δ 3.34 (s, 3 H), 3.73-3.77 (m, 2

H), 3.99 (s, 3 H), 4.28-4.35 (m, 2 H), 5.60 (s, 1 H), 7.29-7.34

(m, 3 H), 7.39 (t, J ) 7.43 Hz, 1 H), 7.48 (s, 1 H), 7.56 (t, J )

7.93 Hz, 2 H), 7.82 (s, 1 H), 8.78-8.80 (s, 1 H), 9.16 (s, 1 H).

Anal. (C₂₄H₂₁N₃O₆‚0.50H₂O) C, H, N.

2-(2-Methoxyethoxy)-5-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}benzo-1,4-quinone (14) (Method

C). To the solution of 0.75 g (1.68 mmol) of 13 in 20 mL of

CH₂Cl₂were added 20 mL (253.6 mmol) of 2-methoxyethanol

and 5 drops of Et₃N. The reaction was stirred at room

temperature overnight. The resulting solid was filtered and

washed with ether to give 0.434 g (60.3%) of 14 as a brown

solid: mp 211-212 °C; MS (ESI) m/z 430.3 (M + H)⁺; ¹H NMR

(400 MHz, CDCl₃) δ 3.45 (s, 3 H), 3.49 (s, 3 H), 3.84 (m, 2 H),

3.89 (m, 2 H), 4.07 (s, 3 H), 4.17 (m, 2 H), 4.34 (m, 2 H), 6.02

(s, 1 H), 7.06 (s, 1 H), 7.32 (s, 1 H), 8.06 (s, 1 H), 8.73 (s, 1 H),

8.81 (s, 1 H). Anal. (C₂₁H₂₃N₃O₇·0.5H₂O) C, H, N.

2-Methoxy-5-{[6-methoxy-7-(2-methoxyethoxy)quinazo-

lin-4-yl]amino}benzo-1,4-quinone (16). A mixture of 9a

(4.55 g, 11.7 mmol), CH₃OH (3.74 g, 11.7 mmol), and Et₃N (3.45

g, 35.1 mmol) in 300 mL of CH₂Cl₂was refluxed for 3 h. The

reaction was cooled and diluted with H₂O. The organic layer

was separated and washed with saturated NaHCO₃solution.

The organic solution was passed through a Magnesol plug,

eluting with CH₃OH-CH₂Cl₂mixtures. The solvent was

evaporated to yield 1.75 mg (39%) of 16: mp 245-248 °C; MS

1

(ESI) m/z 386.3 (M + H)⁺; H NMR (400 MHz, CDCl₃) δ 3.49

(s, 3 H), 3.89 (m, 2 H), 3.92 (s, 3 H), 4.08 (s, 3 H), 4.34 (m, 2

H), 5.99 (s, 1 H), 7.07 (s, 1 H), 7.33 (s, 1 H), 8.07 (s, 1 H), 8.76

(s, 1 H), 8.82 (s, 1 H). Anal. (C₁₉H₁₉N₃O₆) C, H, N.

2-[4-(1H-imidazol-1-yl)phenoxy]-5-{[6-methoxy-7-(2-

methoxyethoxy)quinazolin-4-yl]amino}benzo-1,4-quino-

ne (18) (Method F). To a mixture of 4-imidazol-1-ylphenol

(24.32 mg, 0.152 mmol), 1 N NaOH (116 µL, 0.116 mmol), and

a trace of the phase transfer catalyst Aliquat-336 (tricapry-

lylmethylammonium chloride) in 2 mL of CH₂Cl₂and 1 mL of

H₂O was added 9a (39.29 mg, 0.101 mmol). More CH₂Cl₂was

added to dissolve the solids. The reaction was shaken at room

temperature overnight. The phases were separated. The

organic layers were evaporated and purified by ether-CH₃-

CN recrystallization to yield 36 mg (69%) of 18 as a light brown

solid: mp 124-132 °C; MS (ESI) m/z 514.1 (M + H)⁺, 257.7

(M + 2H)²⁺; ¹H NMR (400 MHz, CDCl₃) δ 3.45-3.51 (s, 3 H),

3.85-3.92 (m, 2 H), 4.03-4.10 (s, 3 H), 4.29-4.38 (m, 2 H),

5.81 (s, 1 H), 7.03 (s, 1 H), 7.26 (s, 1 H), 7.29-7.34 (m, 4 H),

7.50-7.55 (m, 2 H), 7.89 (s, 1 H), 8.18 (s, 1 H), 8.66 (s, 1 H),

8.84 (s, 1 H). Anal. (C₂₇H₂₃N₅O₆·1.25 H₂O) C, H, N.

2-[2-Fluoro-1-(fluoromethyl)ethoxy]-5-{[6-methoxy-7-

(2-methoxyethoxy)quinazolin-4-yl]amino}benzo-1,4-

quinone (26) (Method E). A mixture of 9a (60 mg, 0.15

mmol), 1,3-difluoro-2-propanol (147.84 mg, 1.54 mmol), and

NaOPh(3H₂O) (51 mg, 0.3 mmol) in 2 mL of CH₂Cl₂was

shaken at room temperature overnight. The reaction was

diluted with H₂O, and the layers were separated. The organic

layer was washed three times with saturated K₂CO₃and was

passed through a plug of Magnesol, eluting with CH₂Cl₂

followed by 1% CH₃OH-CH₂Cl₂. The solvent was removed in

vacuo to yield 18 mg (26.7%) of 26: MS (ESI) m/z 450.2 (M +

H)⁺; HRMS (ESI-FTMS (M + H)⁺) calcd for C₂₁H₂₁F₂N₃O₆

450.14712, found 450.14647; ¹H NMR (400 MHz, CDCl₃) δ 3.49

(s, 3 H), 3.89 (m, 2 H), 4.07 (s, 3 H), 4.34 (m, 2 H), 4.61-4.73

(m, 2 H), 4.75-4.91 (m, 3 H), 6.13 (s, 1 H), 7.04 (s, 1 H), 7.32

2-Chloro-1,3,4-trimethoxy-5-nitrobenzene (47). Com-

pound 46 (7.8 g, 35.53 mmol) in 77 mL of DMF was stirred

with dimethyl sulfate (11.2 g, 88.81 mmol) and K₂CO₃(14.73

g, 106.57 mmol) at 80 °C for 1 h. The mixture was poured into

H₂O and the resulting solid was collected by vacuum filtration,

washed with H₂O, and air-dried to give 8.0 g (91%) of 47 as a

gray solid: MS (APCI) m/z 247.1; MS (APCI) m/z 247.1; ¹H

NMR (400 MHz, CDCl₃) δ 3.96 (t, J ) 8.0 Hz, 9 H), 7.17 (s, 1

H).

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7579

4-Chloro-2,3,5-trimethoxy-phenylamine (7e). A solution

of 47 (8.0 g, 32.31 mmol), Fe (10.83 g, 193.33 mmol), and AcOH

(11.1 mL, 193.83 mmol) in CH₃OH (429 mL) was refluxed with

mechanical stirring for 2 h. The reaction was then treated with

NaOH (10 M, 19.38 mL, 193.83 mmol) and filtered. The solid

was washed with EtOAc. The filtrate was concentrated,

redissolved in EtOAc, washed with saturated NaHCO₃, dried

(MgSO₄), and concentrated to give 6.17 g of 7e as a light tan

oil. This material was used without additional purification.

7.21 (s, 1 H), 7.26-7.36 (m, 2 H), 7.56 (d, J ) 12.34 Hz, 1 H),

8.65 (s, 1 H).

N-(4-Chloro-2,5-dimethoxyphenyl)-6-methoxy-7-[(1-

methylpiperidin-4-yl)methoxy]quinazolin-4-amine (58a).

To a stirred mixture of 56 (2.07 g, 4.27 mmol) and 57a (1.09

g, 8.6 mmol) in THF (4.0 mL) under nitrogen at 25 °C was

added sodium bis(trimethylsilyl)amide (1.0 M in THF, 10.6

mL, 10.6 mmol) over 30 s. The reaction mixture was refluxed

for 30 min, cooled, and partitioned between CH₂Cl₂and H₂O.

The CH₂Cl₂layer was washed with brine, dried over MgSO₄,

and evaporated. A solution of the resulting gum in TFA (25

mL) was stirred at 55-60 °C for 45 min and concentrated to

dryness. The residue was partitioned between CH₂Cl₂and

aqueous NaHCO₃. The CH₂Cl₂layer was washed with brine,

dried over MgSO₄, and evaporated. The residue was purified

on a flash column of silica gel (2 × 20 cm), eluting with 3:1

CH₂Cl₂-EtOAc and 25:25:1 CH₂Cl₂-EtOAc-CH₃OH to yield

1.75 (87%) of 58a as a solid: MS (ESI) m/z 473.2 (M + H)⁺,

237.1 (M + 2H)²⁺; ¹H NMR (400 MHz, DMSO-d₆) δ 1.28-1.45

(m, 2 H), 1.76 (d, J ) 10.07 Hz, 3 H), 1.87 (t, J ) 10.83 Hz, 2

H), 2.16 (s, 3 H), 2.79 (d, J ) 11.08 Hz, 2 H), 3.75 (s, 3 H),

3.80 (s, 3 H), 3.93 (s, 3 H), 3.99 (d, J ) 6.04 Hz, 2 H), 7.15 (s,

1 H), 7.22 (s, 1 H), 7.38 (s, 1 H), 7.79 (s, 1 H), 8.32 (s, 1 H),

9.18 (s, 1 H). Anal. (C₂₄H₂₉ClN₄O₄·0.2CH₄O) C, H, N.

2,3-Dimethoxy-5-{[6-methoxy-7-(2-methoxyethoxy)-

quinazolin-4-yl]amino}benzo-1,4-quinone (49). To a solu-

tion of 9e (800 mg, 1.91 mmol) in CH₂Cl₂(115 mL) were added

CsCO₃(825 mg, 2.53 mmol) and CH₃OH (57 mL, 1.45 mol).

The reaction mixture was stirred at room temperature for 2.5

h and filtered through a short column of silica gel. The solvent

evaporated and the residue was chromatographed on silica gel,

eluting with CHCl₃-EtOAc from 7:3 to 5:5. Product fractions

were collected and concentrated. The residue was stirred in

small amount of CH₃CN. The resulting solid was filtered to

yield 0.2 g (25%) of 49 as a red crystalline solid: MS (ESI)

1

m/z 416.1 (M + H)⁺; H NMR (400 MHz, CDCl₃) δ 3.45-3.50

(s, 3 H), 3.85-3.90 (m, 2 H), 3.98 (s, 3 H), 4.07 (s, 3 H), 4.20

(s, 3 H), 4.30-4.38 (m, 2 H), 7.06 (s, 1 H), 7.32 (s, 1 H), 7.93-

7.93 (s, 1 H), 8.66 (s, 1 H), 8.80 (s, 1 H). Anal. (C₂₀H₂₁N₃O₇.

0.5H₂O) C, H, N.

2-({6-Methoxy-7-[(1-methylpiperidin-4-yl)methoxy]-

quinazolin-4-yl}amino)-5-{2-[methyl(phenyl)amino]-

ethoxy}benzo-1,4-quinone (62) (Method G). To a solution

of 107 mg (0.243 mmol) of 59a in 3 mL of CH₂Cl₂was added

158 mg (0.485 mmol) of cesium carbonate and 367 mg (2.43

mmol) of N-methyl-2-(phenylamino)ethanol. The reaction was

stirred at room temperature for 2 h and 20 min. It was then

washed with brine, dried over Na₂SO₄, and concentrated. The

product was purified using a Gilson HPLC. The product

fractions were neutralized with NaHCO₃and extracted with

CH₂Cl₂. The resulting material was filtered through a plug of

Magnesol eluting first with EtOAc (discard) and then with

CH₃OH-EtOAc-Et₃N mixtures. The filtrate was concentrated

in vacuo to yield 25 mg (18%) of 62: MS (ESI) m/z 558.2 (M +

H)⁺, 279.6 (M + 2H)²⁺; HRMS (ESI-FTMS (M + H)⁺) calcd

for C₃₁H₃₅N₅O₅558.27110, found 558.272; ¹H NMR (400 MHz,

CDCl₃) δ 1.88-2.10 (m, 7 H), 2.33 (s, 3 H), 2.90-2.95 (m, 2

H), 3.06-3.11 (s, 3 H), 3.89 (t, J ) 5.79 Hz, 2 H), 4.02-4.08

(m, 5 H), 4.16 (t, J ) 5.92 Hz, 2 H), 5.93 (s, 1 H), 6.71-6.81

(m, 3 H), 7.05 (s, 1 H), 7.24-7.30 (m, 3 H), 8.04 (s, 1 H), 8.73

(s, 1 H), 8.80 (s, 1 H). The purity of 62 was evaluated by two

2-[2-Fluoro-1-(fluoromethyl)ethoxy]-3-methoxy-5-{[6-

methoxy-7-(2-methoxyethoxy)quinazolin-4-yl]amino}-

benzo-1,4-quinone (50). To a solution of 9e (650 mg, 1.55

mmol) in CH₂Cl₂(93 mL) was added CsCO₃(671 mg, 2.06

mmol) and 1,3-difluoro-2-propanol (4.46 g, 46.45 mmol). The

reaction mixture was stirred at room temperature overnight.

The mixture was filtered through a short column of silica gel,

eluting with CHCl₃/EtOAc ) 1:1. The solvent was removed

and the residue was stirred with ether. The resulting solid

was filtered to yield 0.25 g (33.6%) of 50 as a red solid: MS

(ESI) m/z 480.1 (M + H)⁺; HRMS (ESI-FTMS (M + H)⁺) calcd

1

for C₂₂H₂₃F₂N₃O₇480.15768, found 480.15833; H NMR (400

MHz, CDCl₃) δ 3.45-3.50 (s, 3 H), 3.87-3.91 (m, 2 H), 4.06

(s, 6 H), 4.28-4.39 (m, 2 H), 4.65-4.73 (m, 2 H), 4.78-4.87

(m, 2 H), 5.07-5.22 (m, 1 H), 7.04 (s, 1 H), 7.33 (s, 1 H), 7.91-

7.94 (s, 1 H), 8.59 (s, 1 H), 8.82 (s, 1 H). The purity of 50 was

evaluated by two HPLC systems and found to be 100% (system

C, t_R) 3.89 min) and 89% (system D, t_R) 12.2 min).

(4-Chloro-2, 5-dimethoxyphenyl)(4-methoxybenzyl)-

amine (54). Under N₂, to a stirred solution of the p-anisal-

dehyde 53 (35.4 g, 260 mmol) in CH₂Cl₂(750 mL) were added

7a (46.9 g, 250 mmol), sodium triacetoxyborohydride (79.5 g,

375 mmol), and acetic acid (21.5 mL, 375 mmol). The reaction

mixture was stirred at room temperature for 2.5 h. H₂O was

added and the mixture was adjusted to pH 9-10 with K₂CO₃.

The CH₂Cl₂layer was washed with H₂O, dried, and concen-

trated. The residue was dissolved in 3:1 hexanes-EtOAc (500

mL) and passed through a 8.0 × 4.0 cm pad of silica gel. The

solvent was evaporated to yield 77.1 g (96%) of 54 as a white

solid: mp 53-63 °C; MS (ESI) m/z 308 (M + H)⁺; ¹H NMR

(400 MHz, CDCl₃) δ 3.77 (d, J ) 5.54 Hz, 6 H), 3.80 (s, 3 H),

4.26 (s, 2 H), 4.57 (bs, 1 H), 6.25 (s, 1 H), 6.75 (s, 1 H), 6.84-

6.94 (m, 2 H), 7.24-7.34 (m, 2 H).

HPLC solvent systems and found to be 98% (system G, t_R

2.03 min) and 98% (system B, t_R) 2.97 min).

)

2-[2-Fluoro-1-(fluoromethyl)ethoxy]-5-{[6-methoxy-7-

(3-pyrrolidin-1-ylpropoxy)quinazolin-4-yl]amino}benzo-

1,4-quinone (65). A solution of 59b was prepared by CAN

oxidation of 58b (850 mg, 1.8 mmol) and extracting in 500 mL

of CHCl_3.To the solution were added 1,3-difluoro-2-propanol

(1.39 mL, 18.0 mmol) and NaOPh(3H₂O) (306.2 mg, 1.8 mmol).

The reaction mixture was stirred at room temperature for 2

h. The solvent was concentrated to a volume of 80 mL at 40

°C. The reaction mixture was filtered through a plug of

Magnesol, eluting with CH₂Cl₂, EtOAc, 25% 2-propanol-

EtOAc, and then CH₂Cl₂-EtOAc-2-propanol-Et₃N (200:200:

100:6) to yield 97.4 mg (10.7%) of 65 as a red solid: MS (ESI)

m/z 503.1 (M + H)⁺, 252 (M + 2H)²⁺; HRMS (ESI-FTMS)

calcd for C₂₅H₂₈F₂N₄O₅503.21005, found 503.21081; ¹H NMR

(400 MHz, CDCl₃) δ 1.81 (s, 4 H), 2.08-2.24 (m, 2 H), 2.56 (s,

4 H), 2.69 (t, J ) 7.18 Hz, 2 H), 4.08 (s, 3 H), 4.28 (t, J ) 6.55

Hz, 2 H), 4.66-4.72 (m, 2 H), 4.77-4.91 (m, 3 H), 6.13 (s, 1

H), 7.06 (s, 1 H), 7.33 (s, 1 H), 8.10 (s, 1 H), 8.69 (s, 1 H), 8.84

(s, 1 H). The purity of 65 was evaluated by two HPLC systems

and found to be 86% (system C, t_R) 2.03 min) and 76%

(system D, t_R) 7.79 min)

N-(4-Chloro-2,5-dimethoxyphenyl)-7-fluoro-6-methoxy-

N-(4-methoxybenzyl)quinazolin-4-amine (56). A mixture

of 54 (30.8 g, 100 mmol), 55,³³and pyridine (0.65 mL) in tert-

butyl alcohol (240 mL) under nitrogen was stirred at reflux

temperature for 24 h. The tert-butyl alcohol was removed at

reduce pressure. The residue was stirred with CH₂Cl₂and

diluted with NH₄OH. The insoluble material was removed by

filtration. The CH₂Cl₂layer was washed with brine, dried over

MgSO₄, and evaporated to give 148.4 g of a dark red gum. The

gum was dissolved into 40:1 CH₂Cl₂-EtOAc (30 mL) and

placed on a silica gel column (3.6 × 4.2 cm), which was eluted

with 40: 1 CH₂Cl₂-EtOAc to yield 31 g (80%) of 56 as a white

Acknowledgment. The authors wish to thank Drs.

Tarek Mansour, Philip Frost, Lee Greenberger, and

Janis Upeslacis for their support and encouragement.

We also would like to thank the members of the Wyeth

Discovery Oncology group for in vivo evaluations, and

1

solid: mp amorphous; H NMR (400 MHz, DMSO-d₆) δ 3.34

(d, J ) 2.01 Hz, 6 H), 3.70 (d, J ) 5.29 Hz, 6 H), 5.27 (s, 2 H),

6.63 (d, J ) 9.57 Hz, 1 H), 6.78-6.86 (m, 2 H), 7.12 (s, 1 H),

7580 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24

Wissner et al.

metastatic colorectal cancer [see comment]. J. Clin. Oncol. 2003,

21, 60-65.

members of the Wyeth Chemical Technologies group for

analytical and spectral determinations.

(20) Johnson, D. H.; Fehrenbacher, L.; Novotny, W. F.; Herbst, R.

S.; Nemunaitis, J. J.; Jablons, D. M.; Langer, C. J.; DeVore, R.

F., III; Gaudreault, J.; Damico, L. A.; Holmgren, E.; Kabbinavar,

F. Randomized phase II trial comparing bevacizumab plus

carboplatin and paclitaxel with carboplatin and paclitaxel alone

in previously untreated locally advanced or metastatic nonsmall-

cell lung cancer. J. Clin. Oncol. 2004, 22, 2184-2191.

(21) Torrance, C. J.; Jackson, P. E.; Montgomery, E.; Wissner, A.;

Kinzler, K. W.; Vogelstein, B.; Frost, P.; Discafani, C. M.

Combinatorial Chemoprevention of Intestinal Neoplasia. Nat.

Med. 2000, 6, 1024-1028.

Supporting Information Available: Additional synthetic

procedures and characterization data. This material is avail-

able free of charge via the Internet at http://pubs.acs.org.

References

(1) Carmeliet, P.; Jain, R. K. Angiogenesis in cancer and other

diseases. Nature 2000, 407, 249-257.

(2) Terman, B. I.; Dougher-Vermazen, M.; Carrion, M. E.; Dimitrov,

D.; Armellino, D. C., Gospodarowicz, D.; Bohlen, P. Identification

of the KDR tyrosine kinase as a receptor for vascular endothelial

cell growth factor. Biochem. Biophys. Res. Commun. 1992, 187,

1579-1586.

(3) Folkman, J. Role of angiogenesis in tumor growth and metasta-

sis. Sem. Oncol. 2002, 29, 15-18.

(4) Ferrara, N. Vascular endothelial growth factor as a target for

anticancer therapy. Oncologist 2004, 9 Suppl 1, 2-10.

(5) Hicklin, D. J.; Ellis, L. M. Role of the vascular endothelial growth

factor pathway in tumor growth and angiogenesis. J. Clin. Oncol.

2005, 23, 1011-1027.

(6) Kerbel, R. S. Inhibition of tumor angiogenesis as a strategy to

circumvent acquired resistance to anti-cancer therapeutic agents.

Bioessays 1991, 13, 31-36.

(7) Browder, T.; Butterfield, C. E.; Kraling, B. M.; Shi, B.; Marshall,

B.; O’Reilly, M. S.; Folkman, J. Antiangiogenic scheduling of

chemotherapy improves efficacy against experimental drug-

resistant cancer. Can. Res. 2000, 60, 1878-1886.

(22) Wissner, A.; Overbeek, E.; Reich, M. F.; Floyd, M B.; Johnson,

B. D.; Mamuya, N.; Rosfjord, E. C.; Discafani, C.; Davis, R.; Shi,

X.; Rabindran, S. K.; Gruber, B. C.; Ye, F.; Hallett, W. A.;

Nilakantan, R.; Shen. R., Wang, Y.-F.; Greenberger, L. M.; Tsou,

H.-R. Synthesis and Structure-Activity Relationships of 6,7-

Disubstituted 4-Anilinoquinoline-3-carbonitriles. The Design of

an Orally Active, Irreversible Inhibitor of the Tyrosine Kinase

Activity of the Epidermal Growth Factor Receptor (EGFR) and

the Human Epidermal Growth Factor Receptor-2 (HER-2). J.

Med. Chem. 2003, 46, 49-63.

(23) Rabindran, S. K.; Discafani, C. M.; Rosfjord, E. C.; Baxter, M.;

Floyd, M. B.; Golas, J.; Hallett, W. A.; Johnson, B. D.; Nilakan-

tan, R.; Overbeek, E.; Reich, M. F.; Shen, R.; Shi, X.; Tsou, H.-

R.; Wang, Y.-F.; Wissner, A. Antitumor Activity of HKI-272, an

Orally Active, Irreversible Inhibitor of the HER-2 Tyrosine

Kinase. Cancer Res. 2004, 64, 3958-3965.

(24) Tsou, H.-R.; Overbeek, E.; Hallett, W. A.; Reich, M. F.; Floyd,

M. B.; Johnson, B. D.; Michalak, R. S.; Nilakantan, R.; Discafani,

C.; Golas, J.; Rabindran, S. K.; Shen, R.; Shi, X.; Wang, Y.-F.;

Upeslacis, J.; Wissner, A. Optimization of 6,7-Disubstituted-4-

(arylamino)quinoline-3-carbonitriles as Orally Active, Irrevers-

ible Inhibitors of HER-2 kinase Activity. J. Med. Chem. 2005,

48, 1107-1131.

(25) McTigue, M. A.; Wickersham, J. A.; Pinko, C.; Showalter, R. E.;

Parast, C. V.; Tempczyk-Russell, A.; Gehring, M. R.; Mrocz-

kowski, B.; Kan, C. C.; Villafranca, J. E.; Appelt, K. Crystal

Structure of the Kinase Domain of Human Vascular Endothelial

Growth Factor Receptor 2: A Key Enzyme in Angiogenesis.

Struct. Fold Des. 1999, 7, 319-330.

(26) Hennequin, L. F.; Thomas, A. P.; Johnstone, C.; Stokes, E. S.

E.; Ple´, P. A.; Lohmann, J.-J. M;Ogilvie, D. J.; Dukes, M.; Wedge,

S. R. Curwen,; J. O.; Kendrew; Lambert-van der Brempt, C.

Design and Structure-Activity Relationship of a New Class of

Potent VEGF Receptor Tyrosine Kinase Inhibitors. J. Med.

Chem. 1999, 42, 5369-5389.

(27) Wissner, A.; Berger, D. M.; Boschelli, D. H.; Floyd, M. B. Jr.,

Greenberger, L. M.; Gruber, B. C.; Johnson, B. D.; Mamuya, N.;

Nilakantan, R.; Reich, M. F.; Shen, R.; Tsou, H.-R.; Upeslacis,

E.; Wang, Y.-F.; Wu, B.; Ye, F.; Zhang, N. 4-Anilino-6,7-

dialkoxyquinoline-3-carbonitrile Inhibitors of Epidermal Growth

Factor Receptor (EGF-R) Kinase and Their Bioisosteric Rela-

tionship to the 4-Anilino-6,7-dialkoxyquinazoline Inhibitors. J.

Med. Chem. 2000, 43, 3244-3256.

(28) Shewchuk, L.; Hassell, A.; Wisely, B.; Rocque, W.; Holmes, W.;

Veal, J.; Kuyper, L. F. Binding Mode of the 4-Anilinoquinazoline

Class of Protein Kinase Inhibitor: X-ray Crystallographic Stud-

ies of 4-Anilinoquinazolines Bound to Cyclin-Dependent Kinase

2 and p38 Kinase. J. Med. Chem. 2000, 43, 133-138.

(29) Stamos, J.; Sliwkowski, M. X.; Eigenbrot, C. Structure of the

Epidermal Growth Factor Receptor Kinase Domain Alone and

in Complex with a 4-Anilinoquinazoline Inhibitor. J. Biol. Chem.

2002, 277, 46265-46272.

(30) Wedge, S. R.; Ogilvie, D. J.; Dukes, M.; Kendrew, J.; Curwen,

J. O.; Hennequin, L. F.; Thomas, A. P.; Stokes, E. S. E.; Curry,

B.; Richmond, G. H. P.; Wadsworth, P. F. ZD4190: An Orally

Active Inhibitor of Vascular Endothelial Growth Factor Signal-

ing with Broad-spectrum Antitumor Efficacy. Cancer Res. 2000,

60, 970-975.

(8) Kerbel, R. S.; Kamen, B. A. The anti-angiogenic basis of

metronomic chemotherapy. Nature Rev. Cancer 2004, 4, 423-

436.

(9) Ferrara, N.; Hillan, K. J.; Gerber, H. P.; Novotny, W. Discovery

and development of bevacizumab, an anti-VEGF antibody for

treating cancer. Nat. Rev. Drug Dis. 2004, 3, 391-400.

(10) Sweeney, P.; Karashima, T.; Kim, S. J.; Kedar, D.; Mian, B.;

Huang, S., Baker, C.; Fan, Z.; Hicklin, D. J.; Pettaway, C. A.;

Dinney, C. P. Anti-vascular endothelial growth factor receptor

2 antibody reduces tumorigenicity and metastasis in orthotopic

prostate cancer xenografts via induction of endothelial cell

apoptosis and reduction of endothelial cell matrix metallopro-

teinase type 9 production. Clin. Cancer Res. 2002, 8, 2714-2724.

(11) Pavco, P. A.; Bouhana, K. S.; Gallegos, A. M.; Agrawal, A.;

Blanchard, K. S.; Grimm, S. L.; Jensen, K. L.; Andrews, L. E.;

Wincott, F. E.; Pitot, P. A.; Tressler, R. J.; Cushman, C.;

Reynolds, M. A.; Parry, T. J. Antitumor and antimetastatic

activity of ribozymes targeting the messenger RNA of vascular

endothelial growth factor receptors. Clin. Cancer Res. 2000, 6,

2094-2103.

(12) Tokunaga, T.; Abe, Y.; Tsuchida, T.; Hatanaka, H.; Oshika, Y.;

Tomisawa, M.; Yoshimura, M.; Ohnishi, Y.; Kijima, H.; Yamaza-

ki, H.; Ueyama, Y.; Nakamura, M. Ribozyme mediated cleavage

of cell-associated isoform of vascular endothelial growth factor

inhibits liver metastasis of a pancreatic cancer cell line. Int. J.

Oncol. 2002, 21, 1027-1032.

(13) Takei, Y.; Kadomatsu, K.; Yuzawa, Y.; Matsuo, S.; Muramatsu,

T. A small interfering RNA targeting vascular endothelial

growth factor as cancer therapeutics. Cancer Res. 2004, 64,

3365-3370.

(14) Underiner, T. L.; Ruggeri, B.; Gingrich, D. E. Development of

vascular endothelial growth factor receptor (VEGFR) kinase

inhibitors as anti-angiogenic agents in cancer therapy. Cur. Med.

Chem. 2004, 11, 731-745.

(15) Boyer, S. J. Small molecule inhibitors of KDR (VEGFR-2)

kinase: An overview of structure activity relationships. Curr.

Top. Med. Chem. 2002, 2, 973-1000.

(16) Laird, A. D.; Cherrington, J. M. Small molecule tyrosine kinase

inhibitors: Clinical development of anticancer agents. Exp. Opin.

Invest. Drugs 2003, 12, 51-64.

(17) Steward, W. P.; Thomas, A.; Morgan, B.; Wiedenmann, B.;

Bartel, C.; Vanhoefer, U.; Trarbach, T.; Junker, U.; Laurent, D.

Lebwohl, D. Expanded phase I/II study of PTK787/ZK 222584

(PTK/ZK), a novel, oral angiogenesis inhibitor, in combination

with FOLFOX-4 as first-line treatment for patients with meta-

static colorectal cancer. J. Clin. Oncol. (2004 ASCO Annual

Meeting Proceedings) 2004, 22, 3556.

(31) Borgulya, J.; Bruderer, H.; Bernauer, K.; Zuercher, G.; Mose,

D.-P. Catechol-O-methyltransferase-inhibiting Pyrocatechol De-

rivatives: Synthesis and Structure-Activity Studies. Helv.

Chim. Acta 1989, 72, 952-968.

(32) Wissner, A.; Floyd, M. B; Rabindran, S. K.; Nilakantan, R.;

Greenberger, L. M.; Shen, R.; Wang, Y.-F.; Tsou, H.-R. Syntheses

and EGFR and HER-2 Kinase Inhibitory Activities of 4-Anili-

noquinoline-3-Carbonitriles. Analogues of Three Important 4-A-

nilinoquinazolines Currently Undergoing Clinical Evaluation as

Therapeutic Antitumor Agents. Bioorg. Med. Chem. Lett. 2002,

12, 2893-2897.

(33) Barker, A. J.; Brown, D. S. Preparation of Aminoquinazolines

Useful in the Treatment of Cancer. Eur. Pat. Appl. EP602851,

1994.

(34) Magee, P. S. Exploring the Chemistry of Quinones by Computa-

tion. Quant. Struct. Act. Relat. 2000, 19, 22-28.

(18) Raymond, E.; Faivre, S.; Vera, K.; Delbaldo, C.; Robert, C.; Spatz,

A.; Bello, C.; Brega, N.; Scigalla, P.; Armand, J. P. Final results

of a phase I and pharmacokinetic study of SU11248, a novel

multi-target tyrosine kinase inhibitor, in patients with advanced

cancers. Proc. Am. Soc. Clin. Oncol. 2003, 22, 192 (abstr 769).

(19) Kabbinavar, F.; Hurwitz, H. I.; Fehrenbacher, L.; Meropol, N.

J.; Novotny, W. F.; Lieberman, G.; Griffing, S.; Bergsland, E.

Phase II, randomized trial comparing bevacizumab plus fluo-

rouracil (FU)/leucovorin (LV) with FU/ LV alone in patients with

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones

Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7581

(35) Rozeboom, M. D.; Tegmo-Larsson, I.-M.; Houk, K. N. Frontier

Molecular Orbital Theory of Substituent Effects on Regioselec-

tivities of Nucleophilic Additions and Cycloadditions to Benzo-

quinones and Naphthoquinones. J. Org. Chem. 1981, 46, 2338-

2345.

(36) Takahashi, N.; Schreiber, J.; Fischer, V.; Mason, R. P. Formation

of Glutathione-Conjugated Semiquinones by the Reaction of

Quinones with Glutathione: An ESR Study. Arch. Biochem.

Biophys. 1987, 252, 41-48.

(37) FRED; Open Eye Scientific Software: Santa Fe, NM (www.eye-

sopen.com)

(38) Rush, T. S.; III.; Manas, E.; Tawa, G.; and Alvarez, J. Solvation-

Based Scoring for High Throughput Docking in Virtual Screen-

ing in Drug Discovery; Alvarez, J. and Shoichet, B., Eds.; CRC

Press: Boca Raton, FL, 2005; pp 241-269.

(39) Loganzo, F. and Hardy, C. A A Sensitive, Time-resolved Fluo-

rometric Assay for Detection of Inhibitors of Phosphotyrosine

Kinases. Am. Biotech. Lab. 1998, 16, 26-28.

(40) Hennequin, L. F.; Stokes, E. S. E.; Thomas, A. P.; Johnstone,

C.; Ple, P. A.; Ogilvie, D. J.; Dukes, M.; Wedge, S. R.; Kendrew,

J.; Curwen, J. O. Novel 4-Anilinoquinazolines with C-7 Basic

Side Chains: Design and Structure Activity Relationship of a

Series of Potent, Orally Active, VEGF Receptor Tyrosine Kinase

Inhibitors. J. Med. Chem. 2002, 45, 1300-1312.

(41) Bold, G.; Altmann, K.-H.; Bruggen, J.; Frei, J.; Lang, M.; Manley,

P. W.; Traxler, P.; Wietfeld, B.; Buchdunger, E.; Cozens, R.;

Ferrari, S.; Furet, P.; Hofmann, F.; Martiny-Baron, G.; Mestan,

J.; Rosel, J.; Sills, M.; Stover, D.; Acemoglu, F.; Boss, E.;

Emmenegger, R.; Lasser, L.; Masso, E.; Roth, R.; Schlachter, C.;

Vetterli, W.; Wyss, D.; Wood, J. M. New Anilinophthalazines as

Potent and Orally Well Absorbed Inhibitors of the VEGF

Receptor Tyrosine Kinases Useful as Antagonists of Tumor-

Driven Angiogenesis. J. Med. Chem. 2000, 43, 2310-2323.

(42) Dougher, M.; I. Terman, B. I. Autophosphorylation of KDR in

the kinase domain is required for maximal VEGF-stimulated

kinase activity and receptor internalization. Oncogene 1999, 18,

1619-1627.

(43) Sybyl 6.9; Tripos, Inc : St. Louis, MO.

(44) OMEGA; Open Eye Scientific Software: Santa Fe, NM.

(45) Davidson, N. Statistical Mechanics; McGraw-Hill: New York,

1962. McQuarrie, D. A. Statistical Mechanics; Harper & Row

Inc., New York, 1976.

(46) Butenandt, A.; Biekert, E.; Haerle, E. Preparation and conver-

sions of 4,5,8-trihydroxy-6-mercaptoquinoline-2-carboxylic acid.

Chem. Ber. 1964, 97, 285-294.

JM050559F

Article Doi

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones as covalent-binding, irreversible inhibitors of the kinase domain of vascular endothelial growth factor receptor-2

DOI: 10.1021/jm050559f

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm050559f

Full text of DOI:10.1021/jm050559f

Products guided by the article

R&D Labs maybe for 870962-50-8

Relevant to this article

LIGATING PROPERTIES OF THIONITROSOAMINES. I. NEUTRAL MONONUCLEAR N-THIONITROSODIMETHYLAMINE-PALLADIUM(II) AND -PLATINUM(II) COMPLEXES

New 8-substituted xanthiene derivatives as potent bronchodilators

Copper-Catalyzed Lactamization of (E)-2-(2-Bromophenyl)-3-arylacrylamides for the Synthesis of (E)-3-Arylideneindolin-2-ones

First examples of intramolecular addition of primary amidyl radicals to olefins

A facile, regioselective and controllable bromination of aromatic amines using a CuBr₂/Oxone system

Article Doi

2-(Quinazolin-4-ylamino)-[1,4]benzoquinones as covalent-binding, irreversible inhibitors of the kinase domain of vascular endothelial growth factor receptor-2

DOI: 10.1021/jm050559f

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm050559f

Full text of DOI:10.1021/jm050559f

Products guided by the article

R&D Labs maybe for 870962-50-8

Relevant to this article

LIGATING PROPERTIES OF THIONITROSOAMINES. I. NEUTRAL MONONUCLEAR N-THIONITROSODIMETHYLAMINE-PALLADIUM(II) AND -PLATINUM(II) COMPLEXES

New 8-substituted xanthiene derivatives as potent bronchodilators

Copper-Catalyzed Lactamization of (E)-2-(2-Bromophenyl)-3-arylacrylamides for the Synthesis of (E)-3-Arylideneindolin-2-ones

First examples of intramolecular addition of primary amidyl radicals to olefins

A facile, regioselective and controllable bromination of aromatic amines using a CuBr2/Oxone system

A facile, regioselective and controllable bromination of aromatic amines using a CuBr₂/Oxone system