Quinazolines as adenosine receptor antagonists: SAR and selectivity for A2B receptors

Article abstract of DOI:10.1016/S0968-0896(02)00323-1

We have recently reported the discovery of numerous new compounds that are selective inhibitors of all of the subtypes of the adenosine receptor family via a pharmacophore database searching and screening strategy. During the course of this work we made the unexpected discovery of a potent A_2B receptor antagonist, 4-methyl-7-methoxyquinazolyl-2-(2′-amino-4′-imidazolinone) (38, CMB 6446), which showed selectivity for this receptor and functioned as an antagonist, with a binding K_i value of 112 nM. We explored the effects of both substituent- and ring-structural variations on the receptor affinity in this series of derivatives, which were found to be mostly non-selective adenosine receptor ligands with K_i values in the micromolar range. Since no enhancement of A_2B receptor affinity of 38 was achieved, the previously reported pharmacophore-based searching strategy yielded the most potent and selective structurally-related hit in the database originally searched.

Full text of DOI:10.1016/S0968-0896(02)00323-1

Bioorganic & MedicinalChemistry 11 (2003) 77–85
Quinazolines as Adenosine Receptor Antagonists:
SAR and Selectivity for A2B Receptors
Thomas R. Webb,a,* Dmitriy Lvovskiy,a Soon-Ai Kim,b Xiao-duo Ji,b
c
Neli Melman,b JoelLinden and Kenneth A. Jacobsona,b,
*
aChemBridge Corporation, 16981 Via Tazon, Suite G, San Diego, CA 92127, USA
bMolecular Recognition Section, Laboratory of Bioorganic Chemistry, NIDDK, NIH, Bldg. 8A, Rm. B1A 17,
Bethesda, MD 20892-0810, USA
cUniversity of Virginia, Charlottesville, VA, USA
Received 14 February 2002; accepted 11 July 2002
Abstract—We have recently reported the discovery of numerous new compounds that are selective inhibitors of all of the subtypes
of the adenosine receptor family via a pharmacophore database searching and screening strategy. During the course of this work we
made the unexpected discovery of a potent A2B receptor antagonist, 4-methyl-7-methoxyquinazolyl-2-(20-amino-40-imidazolinone)
(38, CMB 6446), which showed selectivity for this receptor and functioned as an antagonist, with a binding Ki value of 112 nM. We
explored the effects of both substituent- and ring-structural variations on the receptor affinity in this series of derivatives, which
were found to be mostly non-selective adenosine receptor ligands with Ki values in the micromolar range. Since no enhancement of
A2B receptor affinity of 38 was achieved, the previously reported pharmacophore-based searching strategy yielded the most potent
and selective structurally-related hit in the database originally searched.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
originally searched, we conducted substructure-based
searching and AR assays on other substructure ‘hits’.
These other ‘hits’ provided a wide range of substituents
on the quinazoline ring. These experiments would give
us information regarding the overall quality of the
pharmacophore-based queries and structure–activity
relationship data. We synthesized five additional com-
pounds for the purpose of testing the importance on one
of the ring nitrogens for adenosine receptor recognition.
The adenosine receptors (ARs) are G protein-coupled
receptors consisting of four subtypes, A1, A2A, A2B, and
A3.1 Untilrecentyl , no selective antagonists of the A2B
subtype were known, among either xanthines or non-
xanthines. It has been difficult to characterize the phar-
macological actions of this AR subtype using only non-
selective agonists and antagonists.
2
Our approach to the identification of novelAR
antagonists has been unified pharmacophore-based
screening.3 During this study we discovered one com-
pound that showed significant binding to all AR sub-
types at 10 micromolar concentration (38, CMB 6446),
but also showed some selectivity for the A2BAR. In order
to determine whether this compound that bound to the
A2BAR was already optimized for potency among the
members of this chemicalseries present in the database
Few A2BAR subtype-selective antagonists are known,
but the biological interest in this subtype is growing.
The alkylxanthine theophylline is a weak, non-selective
AR antagonist, used therapeutically for the treatment of
asthma. The use of theophylline has been associated
with unpleasant side effects, such as insomnia and diur-
esis. The development of a theophylline-like drug with
reduced side effects is desirable. The mechanism of
action of theophylline in asthma has been controversial
for decades,4 and only recently did antagonism of the
A2BAR subtype become a likely mechanism.5,6 The
A2BAR is expressed in some mast cells, such as canine
mastocytoma cells and human HMC-1 cells, in which it
is responsible for triggering acute Ca2+ mobilization.
*Corresponding authors. Tel.:+1-301-496-9024; fax:+1-301-480-
8422; e-mail: twebb@chembridge.com (T. R. Webb); tel.:+1-858-451-
7400; fax:+1-858-451-4701; e-mail: kajacobs@helix.nih.gov (K. A.
Jacobson).
0968-0896/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00323-1

78
T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
Thus, a novelnonpurine A 2BAR antagonist, to com-
plement the recently reported A2BAR selective xan-
thines,2 would be useful as a probe of the relationship of
this receptor subtype to disease states.
In addition to 38, the additionalquinazoline derivatives,
3–53, were tested for affinity in radioligand binding
assays at ARs as shown in Tables 1–5. Compounds 3–
48 were obtained from the database, and the synthetic
procedures for compounds not previously reported are
provided. Scheme 1 shows an overview of the chemistry
that was used in the synthesis of these compounds.
Substituted anilines of general structure I were subjected
to acetone in the presence of iodine to give substituted
1,2-dihydro-2,2,4-trimethylquinoline derivatives of gen-
Results
To further understand the structure–activity relation-
ships of the most potent A2B antagonist discovered (38,
CMB 6446) in our previously reported pharmacophore
searching efforts,3 and to explore structural modifi-
cations that might lead to enhanced selectivity we set
about the iterative collection, synthesis and assay of
structurally related compounds. In order to facilitate
this process we used 2-guanidinylquinazoline sub-
structure searches in ISISBase to search a large avail-
able database of compounds (ꢀ100,000 from Express-
PickTM, available from ChemBridge corporation) the
same database that was originally used for the pharma-
cophore based queries (CNS-SetTM). Compounds that
9
eralstructure II. Derivatives of this type were treated
with cyanoguanidine to give substituted 4-methylquina-
zolyl-2-guanidines (III),10 and then further derivatized
11
to give compounds of the generalstructure IV. Alter-
nately compounds of general structure III could be
selectively mono- or di-alkylated or acylated to give
12
derivatives of the generalstructure V.
We also prepared a series of new substituted quinoline
derivatives (see Scheme 2). Compounds 49–53 were
prepared in order to test the requirement for the N3
nitrogen in AR binding. These substituted 4-methylqui-
noline derivatives were prepared from 2-amino-4-
matched
the
query from either database were selected for assay.
2-guanidinylquinazoline-substructure
Table 1. Affinities of quinazoline derivatives in radioligand binding assays at A1, A2A, and A2B adenosine receptorsaÀc
R0
R00
rA1
rA2A
hA2B
rA1/hA2B
a
b
c
Compd
R
3
4
5
6
7
8
9
—
—
—
—
CH3
CH3
CH3
CH3
OH
H, H
CH3CH2CO, H
C6H5CO, H
CH3CO,CH3CO
H, H
H, H
CH3CO,CH3CO
H, H
CH3CH2CO, H
H, H
CH3,H
CH3CO, H
2:CH3CH2CO
C6H5CO, H
CH3CO,CH3CO
H, H
3.77Æ0.84
2.13Æ0.31
0.988Æ0.171
57% (10À4
36Æ4% (10À4
5.16Æ1.32
23.2Æ8.7
1.22Æ0.14
7.33Æ2.20
43
5.05Æ0.62
4.28Æ1.15
36Æ7% (10À5
39Æ0% (10À5
35Æ2% (10À5
)
)
)
<1
<1
<1
0.296Æ0.063
)
<10% (10À5
<10% (10À5
)
)
—
)
46.0Æ13.0
8.41Æ2.66
<10%(10À4
2.94Æ1.22
9.22Æ1.59
8.41Æ2.72
1.58Æ0.30
5-OCH3
5-OCH3
6-CH3
6-CH3
6-CH3
6-OCH3
6-OCH3
6-OCH3
6-OCH3
6-OCH2CH3
6-CH3,7-CH3
6-CH3,7-CH3
6-CH3,7-CH3
6-CH3,8-CH3
7-CH3
CH3
CH3
CH3
CH3
OCH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH2CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
10Æ4% (10À5
)
<1
<1
—
)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
39Æ5% (10À5
)
)
27Æ9% (10À5
50Æ1% (10À5)
12.2Æ2.9
53Æ2% (10À5
50Æ1% (10À5
1.97Æ0.09
)
)
<10% (10À5
11.7Æ3.1
)
)
>1
5.9
>1
34Æ2% (10À5
)
<10% (10À5
9.54Æ3.41
2.30Æ0.37
2.49Æ0.43
34Æ4% (10À5
)
17.2Æ8.5
6.8Æ1.8
<10% (10À5
3.44Æ0.18
)
0.67
<1
CH3CO, H
2: CH3CH2CO,
H, H
CH3CH2CO, H
CH3CO,CH3CO
H, H
3.90Æ1.76
41Æ2% (10À5
)
38Æ3% (10À4
2.01Æ0.03
10.1Æ5.2
)
)
<10% (10À5
)
2.2Æ0.5
22.5Æ8.1
45Æ6% (10À5)
<1
32Æ7% (10À5
)
7-CH3
7-CH3
9.22Æ2.77
0.359Æ0.095
1.62Æ0.13
14.2Æ3.3
21.4Æ5.7
<10% (10À5
1.36Æ0.31
)
0.551Æ0.102
2.14Æ0.56
12.4Æ7.0
0.26
7.5
7-OCH3
7-OCH3
7-OCH3
7-OCH2CH3
8-CH3
H, H
0.215Æ1.040
<10% (10À5
<10% (10À5
<10% (10À5
C6H5CO, H
CH3CO,CH3CO
CH3CO, H
H, H
)
)
)
47Æ5% (10À4
21.5Æ10.2
3.96Æ1.22
0.413Æ0.087
1.33Æ0.61
32
1.87Æ0.96
47Æ1% (10À5
0.592Æ0.120
1.78Æ0.03
)
<1
0.70
0.75
8-OCH3
8-OCH2CH3
H, H
H, H
0.258Æ0.069
0.624Æ0.250
aDisplacement of specific [3H]R-PIA binding in rat brain membranes, expressed as KiÆSEM in mM (n=3–5), or as a percentage of specific binding
displaced at the indicated concentration (M).
bDisplacement of specific [3H]CGS 21680 binding in rat striatalmembranes. Expressed a KiÆSEM in mM (n=3–6), or as a percentage of specific
binding displaced at the indicated concentration (M).
cDisplacement of specific [3H]ZM241,385 binding at human A2B receptors expressed in HEK cells, in membranes, expressed as KiÆSEM in mM
(n=3–4), or as a percentage of specific binding displaced at the indicated concentration (M).

T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
79
Table 2. Affinities of 4-methylquinazoline imidazolinone derivatives in radioligand binding assays A1, A2A, and A2B receptorsaÀc
R0
rA1
rA2A
hA2B
rA1/hA2B
a
b
c
Compd
R
32
33
34
35
36
37
38
—
6-CH3
6-CH3
H, H
H, H
10.2Æ2.3
1.44Æ0.31
2.91Æ0.38
12.7Æ1.7
5.58Æ1.95
1.35Æ0.26
1.29Æ0.43
23.8Æ4.0
45% (10À5
)
—
3.2
<1
1.3
<1
4.6
12
5.74Æ1.05
1.49Æ0.40
17.7Æ11.9
24Æ2% (10À4
12.9Æ4.1
0.446Æ0.096
¼CHCOOCH3
38Æ8% (10À5
9.44Æ0.11
)
)
6-OCH3
6-OCH2CH3
7-CH3
H, H
H, H
)
14Æ3% (10À5
0.292Æ0.070
0.112Æ0.015
H, H
H, H
7-OCH3
2.4Æ0.3
aDisplacement of specific [3H]R-PIA binding in rat brain membranes, expressed as KiÆSEM in mM (n=3–5).
bDisplacement of specific [3H]CGS 21680 binding in rat striatalmembranes. Expressed a KiÆSEM in mM (n=3–6), or as a percentage of specific
binding displaced at the indicated concentration (M).
cDisplacement of specific [3H]ZM241,385 binding at human A2B receptors expressed in HEK cells, in membranes, expressed as KiÆSEM in mM
(n=3–4), or as a percentage of specific binding displaced at the indicated concentration (M).
Table 3. Affinities of 4-methylquinazoline dihydropyrimidine derivatives in radioligand binding assays A1,A2A, and A2B receptorsaÀc
a
b
c
Compd
R
rA1
rA2A
hA2B
rA1/hA2B
39
40
41
—
6-CH3,7-CH3
7-OCH3
1.89Æ0.12
3.89Æ1.15
5.17Æ1.89
8.4Æ0.4
38Æ3% (10À5
31Æ1% (10À5
4.04Æ0.05
)
)
<1
<1
0.16
0.648Æ0.050
19.3Æ8.3
aDisplacement of specific [3H]R-PIA binding in rat brain membranes, expressed as KiÆSEM in mM (n=3–5).
bDisplacement of specific [3H]CGS 21680 binding in rat striatalmembranes. Expressed a KiÆSEM in mM (n=3–6).
cDisplacement of specific [3H]ZM241,385 binding at human A2B receptors expressed in HEK cells, in membranes, expressed as KiÆSEM in mM
(n=3–4), or as a percentage of specific binding displaced at the indicated concentration (M).
Table 4. Affinities of 4-methylquinazoline benzoheptanone derivatives (R0=methyl) in radioligand binding assays A1, A2A, and A2B receptorsa—c
a
b
c
Compd
R
rA1
rA2A
hA2B
rA1/hA2B
42
43
44
45
46
47
48
—
6-CH3
6-OCH3
6-OCH2CH3
6-CH3,7-CH3
7-CH3
2.68Æ0.13
1.65Æ0.24
3.41Æ0.82
2.88Æ0.06
1.95Æ0.19
0.334Æ0.038
1.04Æ0.04
3.1Æ1.0
1.50Æ0.79
8.4Æ2.9
44Æ2% (10À5
46Æ4% (10À5
25Æ4% (10À5
)
)
)
<1
<1
<1
<1
<1
0.23
0.63
5.71Æ2.35
2.08Æ0.27
1.44Æ0.07
1.57Æ0.35
<10% (10À5
)
42Æ8% (10À5
)
4.3Æ2.4
7-OCH3
1.66Æ0.18
aDisplacement of specific [3H]R-PIA binding in rat brain membranes, expressed as KiÆSEM in mM (n=3–5).
bDisplacement of specific [3H]CGS 21680 binding in rat striatalmembranes. Expressed a KiÆSEM in mM (n=3–6).
cDisplacement of specific [3H]ZM241,385 binding at human A2B receptors expressed in HEK cells, in membranes, expressed as KiÆSEM in mM
(n=3–4), or as a percentage of specific binding displaced at the indicated concentration (M).

80
T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
Table 5. Affinities of 4-methylquinoline derivatives in radioligand binding assays A1, A2A, and A2B receptorsa,b,c
a
b
c
Compd
R
rA1
rA2A
hA2B
49
50
52
53
51
COCH3
COOCH3
C(NH)NH2
C(NH)NH-CO2C(CH3)3
C(NH)N (CO2C(CH3)3)2
6.30Æ1.13
21.1Æ7.3
45.7Æ4.6
<10% (10À4
17.2Æ1.9
<10% (10À5
<10% (10À5
<10% (10À5
<10% (10À5
<10% (10À5
)
)
)
)
)
)
19.8Æ2.6
8.75Æ2.77
<10% (10À5
5.06Æ1.69
)
40.7Æ10.1
aDisplacement of specific [3H]R-PIA binding in rat brain membranes, expressed as KiÆSEM in mM (n=3–5), or as a percentage of specific binding
displaced at the indicated concentration (M).
bDisplacement of specific [3H]CGS 21680 binding in rat striatalmembranes. Expressed a KiÆSEM in mM (n=3–6), or as a percentage of specific
binding displaced at the indicated concentration (M).
cDisplacement of specific [3H]ZM241,385 binding at human A2B receptors expressed in HEK cells, in membranes, expressed as KiÆSEM in mM
(n=3–4), or as a percentage of specific binding displaced at the indicated concentration (M).
Scheme 1. General synthetic method for analogues in the quinazoline series. R=alkyl, alkoxy. Reagents: (i) iodine/acetone, reflux; (ii) HCl gas; (iii)
cyanoguanidine; (iv) ethylbromoacetate/DMF or gyl cine/water reflux; (v) acylanhydride/DMSO.
methyl-7-methoxyquinoline (Scheme 2). Compounds 49
and 50 were synthesized with typicalacyal tion methods.
Compound 51 was prepared using di-Boc-protected tri-
flylguanidine,20 and the deprotection of one or both of
the Boc groups gave a mixture of compounds 52 (30%)
and 53 (57%).
In addition to binding effects, the functionaleffect of 38
in inhibiting the A2BAR-mediated rise in intracellular
calcium in HEK-A2B cells was examined. At 100 nM, 38
inhibited calcium mobilization stimulated by NECA
(100 nM) by roughly 70%.
The potency of test compounds at the human A2BAR
was evaluated in binding assays and in a functional
assay. As an initialindication of seelctivity, the binding
was compared at rat A1 and A2AARs. There appeared
to be a great freedom of substitution of both the qui-
nazoline ring system and the pendant guanidino group,
as judged from retention of binding at these three ARs.
Discussion
There is evidence that both the A2BAR and A3AR may
play a role in asthma.5,13À15 The A3AR mediates the
degranulation of rat RBL mast-like cells and is present
in high density in human blood eosinophils. The avail-
ability of antagonists selective for the A2BAR shall pro-
vide an opportunity to explore the importance of these
two AR subtypes in asthma and other inflammatory
diseases.
Compound 38, the previously-identified lead, was 10.6-
and 21.4-fold selective for human A2B versus rat A1 and
A2AARs, with a Ki value of 112 nM. At the human A3AR
a Ki value of 1.85Æ0.62 mM was determined,8 thus selec-
tivity with respect to this subtype was 16.5-fold. A related
derivative, 37, was slightly less selective for the A2AAR.
The N3 was found to be necessary for recognition by the
A2BAR, since 3-deaza analogues did not bind at that sub-
type (cf., 25 and it deaza analogue 51). Nevertheless, ana-
logues lacking the N3 retained affinity for A1 and A2AARs.
Xanthine derivatives displaying selectivity for the A2B
subtype in binding assays were recently reported.2,6
Severalsuch xanthines were found to have antagonistic
2,16
effects in functionalassays.
In addition to xanthine
antagonists of ARs, there is need to identify non-xan-
thine structures. We initially set out to use information
derived from known A3AR antagonists both to identify

T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
81
Scheme 2. The synthesis of the quinoline derivatives. Reagents: (i) acetic anhydride, TEA, cat. DMAP in CH2Cl2; (ii) methylchol roformate, TEA,
CH2Cl2; (iii) di-Boc-N-triflylguanidine, TEA, CH2Cl2; (iv) 5% TFA/CH2Cl2.
new leads for this receptor family and to explore the
pharmacophore relationships within the receptor
family. Further, we also wished to explore the utility
of pharmacophore database queries for the discovery
of new leads.3 An unexpected added benefit of our
approach was the discovery of new selective leads for
related AR subtypes, A1, A2A, and A2B. This ability
to find new structuralseries from pharmacophore
information for an existing bioactive structuralseries
is advantageous in the development of new drug
leads.
One imidazolinone derivative, 38, was roughly an order
of magnitude selective for human A2B versus rat A1 and
A2AARs. Further binding and functionalstudies within
the same species will be required to rigorously establish
the true selectivity of this compound. Interestingly this
compound was selected in the initial pharmacophore-
based searching.3 The full characterization of this com-
pound will require comparative binding studies within
the same species. This derivative functionally antag-
onized the Ca2+ response mediated by the recombinant
human A2BAR expressed in HEK293 cells.
Pharmacophore-based screening of a diverse chemical
library has identified 4-methylquinazoline derivatives as
AR antagonists. Analogues of 4-substituted quinazo-
lines have been selected via 2D substructure queries or
synthesized, then tested in receptor binding, and found
to be generally non-selective. These derivatives distantly
resemble isoquinolines reported as A3AR antagonists.19
Most of the structuralvariation occurred at the 2-posi-
tion, with guanidino and substituted guanidino groups.
Simple functional group substitution of the phenyl ring
was included. Nearly all of the compounds bound to A1,
A2A, and A2BARs in the micromolar range. Dramatic
flexibility of substitution of the 2-guanidino group was
possible with retention of affinity. Bulky substituents
(such as dihydropyrimidines 39–41 and benzocyclo-
heptanones 42–48) did not prevent the binding to ARs.
However the di-Boc derivative 51 did not bind appreci-
ably at any of the ARs examined.
In addition to 38, potent binding to the human A2BAR
(Ki<1 mM) was noted for the unsubstituted guanidine
derivatives 25 and 30 and for the other unsubstituted
imidazolinone derivatives 33 and 37. Compound 25,
which displayed a low degree of selectivity for the
A2BAR, was similar to 38 in the position of substitution
of the quinazoline, that is, a 7-methyl group replaced
the 7-methoxy group. Other analogues substituted at
the 7-position were not A2B-selective. The N-benzoyl
derivative 5 and the unsubstituted guanidine 24 were
potent at A1 and A2AARs and less potent at the A2BAR.
The N-propionylderivative 15 was more potent at the
human A2BAR than at rat A1 and A2AARs. The quina-
zoline N3 was found to be necessary for recognition by
the A2BAR but not by A1 and A2AARs.
The lack of high affinity binding to ARs or distinct SAR
patterns within the series suggests that there may be

82
T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
novelmode of binding. Further experiments wlibl e
needed to define the competitive nature of the antagon-
ism by these quinazoline derivatives.
300 ꢁC, Anal. (calcd/found): 57.14/57.12, H: 57.14/
57.12, and N: 30.30/30.32.
General procedure for the synthesis of substituted
4-methylquinazolyl-2-guanidines
Thus, at least in this case, the pharmacophore-based
query selected the most potent and selective compound
in this structuralseries present in the entire database (of
9600 compounds) that was originally searched. The
search of a larger database (ꢀ100,000 compounds)
using 2D sub-structure queries based on 38 did not
identify more potent or selective compounds within this
particular series. Further experiments will need to be
performed to determine if the results obtained from
these types of queries against other targets may also
yield analogous results. Refinement of searching meth-
ods will be of increasing value to researchers in medicinal
chemistry with the recent advent of substructure- and
similarity-based compound acquisition via a web browser
interface and the availability of larger numbers of diverse
compounds from combinatorialand paraellsynthesis.
Further experiments of this type with purine receptors, as
well as application of the search strategies that we have
been developing, will be the topic of future publications.
Following the procedure described above for 4-methyl-
6-methoxy quinazolyl-2-guanidine and substituting the
following substituted quinoline hydrochlorides (For-
mula II, Scheme 1) for 1,2-dihydro-2,2,4-trimethyl-6-
methoxyquinoline hydrochloride, the corresponding
substituted quinazolyl-2-guanidine compounds (For-
mula III) were obtained.
1
4-Hydroxyquinazolyl-2-guanidine (7). H NMR: d 7.15
(dd, 1H, J=8.0, 8.0 Hz), 7.35 (d, 1H, J=8.0 Hz), 7.5
(brs, 4H), 7.55 (dd, 1H, J=8.0, 8.0 Hz), 7.9 (d, 1H,
J=8.0 Hz), 10.95 (brs, 1H). Mp=286–287 ꢁC. Anal.
.
calcd for C9H9N5O H2O: C, 53.20; H, 4.46; N, 34.46.
Found: C, 52.76; H, 4.42; N, 33.40.
1
4,6,8-Trimethylquinazolyl-2-guanidine (21). H NMR: d
1.08 (t, 6H, J=7.2 Hz), 2.43 (s, 3H), 2.45 (s, 3H), 2.59
(br, 4H), 2.82 (s, 3H), 7.57 (s, 1H), 7.95 (s, 1H), 11.6–
12.0 (br, 2H). Mp=228–230 ꢁC.
Experimental
Chemistry
1
4-Ethyl-7-methylquinazolyl-2-guanidine (24). H NMR: d
1.35 (t, 3H, J=7.2 Hz), 2.55 (s, 3H), 3,29 (q, 2H, J=7.2 Hz),
7.46 (dd, 1H, J=8.4, 2.0 Hz), 7.74 (brd, 1H, J=2.0 Hz),
8.16 (d, 1H, J=8.4 Hz), 8.5 (br, 4H). Mp=274–276 ꢁC.
All proton NMR spectra were acquired on a Bruker
AC-300 spectrometer using DMSO-d6 as the solvent,
unless otherwise indicated, and reported as d from TMS
as the internalstandard. Other than compounds 49–53
(see detailed experimental and procedures below) the
compounds used in this study were obtained from the
ChemBridge Corporation collection, described at http://
www.hit2lead.com. General procedures and experi-
mental details are provided below. Following the pro-
cedures described and substituting the corresponding
substituted anilines (Formula I, Scheme 1) for p-meth-
oxyaniline, substituted quinolines (Formula II) were
obtained.8 Compounds 3,17 (10, 13, 14, 16, 17),12 (18,
25),18 (32, 33, 35, 36, 37, 38)11 may be prepared accord-
ing to the literature procedures. The combustion analy-
ticaldata were obtained on new compounds from
NuMega Resonance Labs (San Diego, CA, USA).
1
4-Methyl-8-methylquinazolyl-2-guanidine (29). H NMR:
d 2.48 (s, 3H), 2.71 (s, 3H), 7.15 (br, ꢀ4H), 7.17 (dd,
1H, J=8.1, 6.9 Hz), 7.56 (d, 1H, J=6.9 Hz), 7.81 (d,
1H, J=8.1 Hz). Mp=252–253 ꢁC.
4-Methyl-8-methoxyquinazolyl-2-guanidine (30). 1H NMR:
d 2.69 (s, 3H), 3.91 (s, 3H), 7.17 (dd, 1H, J=7.5, 2.5
Hz), 7.19 (dd, 1H, J=7.5, 6.8 Hz), 7.21 (br, 4H), 7.50
(dd, 1H, J=6.8, 2.5 Hz). Mp=179–181 ꢁC.
1
4-Methyl-8-ethoxyquinazolyl-2-guanidine (31). H NMR:
d 1.42 (t, 3H, J=7.0 Hz), 2.71 (s, 3H), 4,16 (q, 2H,
J=7.0 Hz), 7.17 (dd, 1H, J=7.5, 2.7 Hz), 7.20 (dd, 1H,
J=7.5, 6.4 Hz), 7.33 (br, ꢀ4H), 7.52 (dd, 1H, J=6.4,
2.7 Hz). Mp=218–219 ꢁC.
Preparation of substituted quinazolines
General procedure for the synthesis of 4-methyl-6-
methoxyquinazolyl-2-N-acylguanidines
4-Methyl-6-methoxyquinazolyl-2-guanidine. A solution
was prepared of 0.1 mol1,2-dihydro-2,2,4-trimethy-l6-
methoxyquinoline hydrochloride in 100 mL of water,
and was treated with 0.1 molof dicyandiamide. The
resulting mixture was refluxed for approximately four
hours untilthe evolution of gas ceased. The hot reaction
mixture was decanted from oils, cooled and treated with
concentrated KOH with constant stirring untila pH of
between 10 and 11 was obtained. The precipitate was
isolated by filtration, washed several times with iso-
propanol, recrystallized from DMSO, then washed with
isopropanoland dried under vacuum. The desired pro-
duct of 4-methyl-6-methoxy quinazolyl-2-guanidine was
obtained at a yield of 84%, a melting point of 298–
A suspension of 4-methyl-6-methoxyquinazolyl-2-guan-
idine (0.01 mol) was prepared in 30 mL DMSO at
ambient temperature. The suspension was treated with
the corresponding anhydride (0.005 mol) and allowed to
stir for 2 h. The reaction mixture was filtered and the
residue recrystallized from dioxane yielding the desired
4-methyl-6-methoxyquinazolyl-2-N-acylguanidine.
General procedure for the preparation of substituted
4-methylquinazolyl-2-N-acylguanidines
Following the procedures described for 4-methyl-6-meth-
oxyquinazolyl-2-N-acylguanidine above and substituting

T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
83
the following substituted quinazolyl-2-guanidines (For-
mula III, Scheme 1) for 4-methyl-6-methoxyquinazolyl-
2-guanidine, the corresponding substituted 4-methyl-
quinazolyl-2-N-acylguanidines (Formula V) were
obtained with the listed physical and spectral properties.
4-Methylquinazolyl-2-N,N-diacetylguanidine (6). 1H NMR:
d 2.26 (bs, 6H), 2.88 (s, 3H), 7.60 (dd, 1H, J=8.0, 7.5
Hz), 7.81 (d, 1H, J=8.6 Hz), 7.92 (dd, 1H, J=8.6, 7.5
Hz), 8.21 (d, 1H, J=8.0 Hz), 11.4–11.9 (br, ꢀ2H).
Mp=196–197 ꢁC. Anal. calcd for C14H15N5O2: C,
58.94; H, 5.30; N, 24.55. Found: C, 58.59; H, 5.23; N,
24.43.
4-Methylquinazolyl-2-N-propionylguanidine (4). 1H NMR:
two forms: d 1.07 (t, 3H, J=7,5 Hz), 2.56 (br, 2H), 2.87
and 2.92 (s, 3H), 7.58 and 7.64 (dd, 1H, J=8.2, 7.5 Hz),
7.75–8.0 (m, 2H), 8.20 and 8.25 (d, 1H, J=8.2 Hz), 8.5
(br, ꢀ2H), 11.1–11.6 (br, ꢀ1H). Mp=153–155 ꢁC.
6-Methoxy-4-methylquinazolyl-2-N,N-dipropionylguani-
1
dine (15). H NMR: d 1.10 (t, 6H, J=7.2 Hz), 2.62 (bs,
4H), 2.70 (s, 3H), 3.97 (s, 3H), 7.12–8.15 (m, 3H), 11.6–
12.0 (br, 2H). Mp=156–158 ꢁC.
1
4-Methylquinazolyl-2-N-benzoylguanidine (5). H NMR:
d 2.92 (s, 3H), 7.43–7.46 (m, 2H), 7.51 (dd, 1H, J=8.5,
7.5 Hz), 7.60 (dd, 1H, J=8.0, 7.5 Hz), 7.91–7.97 (m, 2H),
8.18–8.24 (m, 3H), 9.3 (bs, 1H), 10.0 (bs, 1H), 12.1 (bs, 1H).
Mp=170–172 ꢁC. Anal. calcd for C17H15N5O: C, 66.87;
H, 4.95; N, 22.94. Found: C, 66.37; H, 5.09; N, 22.60.
4,6,7 - Trimethylquinazolyl - 2 - N,N - dipropionylguanidine
1
(20). H NMR: d 1.08 (t, 6H, J=7,2 Hz), 2.43 (s, 3H),
2.45 (s, 3H), 2.59 (br, 4H), 2.82 (s, 3H), 7.57 (s, 1H),
1
7.95 (s, 1H), 11.6–12.0 (br, 2H). H NMR (DMSO-d6):
2.25 (bs, 6H), 2.49 (s, 3H), 2.84 (s, 3H), 7.43 (d, 1H,
J=8.4 Hz), 7.60 (s, 1H), 8.09 (d, 1H, J=8.4 Hz), 11.6
(br, ꢀ2H). MS (EI) m/z 299 [M]+. Mp=169–171 ꢁC.
Anal. calcd for C18H23N5O2: C, 63.32; H, 6.79; N,
20.51. Found C, 63.20; H, 6.59; N, 20.46.
1
4,6-Dimethylquinazolyl-2-N-propionylguanidine (11). H
NMR: d 1.06 (t, 3H, J=7,5 Hz), 2.35 (q, 2H, J=7.5
Hz), 2.49 (s, 3H), 2.81 (s, 3H), 7.64–7.69 (m, 2H), 7.86
(s, 1H), 9.9–11.5 (br, ꢀ3H). Mp=179–180 ꢁC. Anal.
calcd for C14H17N5O: C, 61.98; H, 6.32; N, 25.81.
Found: C, 61.86; H, 6.01; N, 25.56.
4,7-Dimethylquinazolyl-2-N,N-diacetylguanidine (23). 1H
NMR: d 2.25 (bs, 6H), 2.49 (s, 3H), 2.84 (s, 3H), 7.43 (d,
1H, J=8.4 Hz), 7.60 (s, 1H), 8.09 (d, 1H, J=8.4 Hz),
11.6 (br, ꢀ2H). MS (EI) m/z 299 [M]+. Mp=180–
182 ꢁC. Anal. calcd for C15H17N5O2: C, 60.19; H, 5.72;
N, 23.40. Found: C, 59.80; H, 5.61; N, 23.13.
4,6,7-Trimethylquinazolyl-2-N-acetylguanidine (19). 1H
NMR, two forms: d 2.28 (bs, 3H), 2.44 (s, 3H), 2.46 (s,
3H), 2.83 and 2.87 (s, 3H), 7.59 and 7.71 (s, 1H), 7.94
and 7.99 (s, 1H), 8.4 (bs, ꢀ2H). MS (EI) m/z 271 [M]+.
Mp=248–249 ꢁC.
4-Methyl-7-methoxyquinazolyl-2-N,N-diacetylguanidine
1
(27). H NMR: d 2.28 (bs, 6H), 2.92 (s, 3H), 3.96 (s,
3H), 7.13 (s, 1H), 7.17 (d, 1H, J=8.4 Hz), 8.09 (d, 1H,
1
4,7-Dimethylquinazolyl-2-N-propionylguanidine (22). H
NMR: d 1.06 (t, 3H, J=7.2 Hz), 2.35 (q, 2H, J=7.2
Hz), 2.49 (s, 3H), 2.79 (s, 3H), 7.33 (d, 1H, J=8.4 Hz),
7.58 (s, 1H), 8.00 (d, 1H, J=8.4 Hz), 9.9–11.1 (br,
ꢀ3H). MS (EI) m/z 271 [M]+. Mp=159–161 ꢁC.
J=8.4 Hz), 11.5–12.1 (br, ꢀ2H). Mp=176–178 ꢁC.
Preparation of 4-methyl-6-methoxyquinazolyl-2-(20-
amino-40-imidazolinone) (35).
4-Methyl-7-methoxyquinazolyl-2- N-benzoylguanidine
1
A mixture of 4-methyl-6-methoxyquinazolyl-2-guani-
dine (0.01 mole) and ethyl 2-bromoacetate (0.006 mole)
in 50 mL of dimethylformamide was refluxed for 4 to 6 h,
allowed to cool and poured into water. The resulting
precipitate was filtered and the residue recrystallized
from dioxane yielding the desired 4-methyl-6-methoxy-
quinazolyl - 2 - (20 - amino - 40 - imidazolinone) with the
(26). H NMR: d 2.91 (s, 3H), 3.95 (s, 3H), 7.51 (d,
1H, J=2.5 Hz), 7.64 (dd, 1H, J=9.0, 2.5 Hz), 7.7 (bs,
1H), 7.89 (d, 1H, J=9.0 Hz), 8.1–8.35 (brm, 4H), 10.9–
13.4 (br, ꢀ2H). MS (EI) m/z 335 [M]+. Mp=236–237ꢁC.
4-Methyl-7-ethoxyquinazolyl-2-N-acetylguanidine (28).
1H NMR: d 1.40 (t, 3H, J=7.2 Hz), 2.05 (s, 3H), 2.80
(s, 3H), 4.19 (q, 2H, J=7.2 Hz), 7.39 (d, 1H, J=2.5
Hz), 7.50 (dd, 1H, J=9.0, 2.5 Hz), 7.73 (d, 1H, J=9.0
Hz), 9.8 (br, ꢀ3H). Mp=154–155 ꢁC. Anal. calcd for
17
reported physicaland spectroscopic properties.
General procedure for the preparation of substituted 4-
methylquinazolyl-2-(20-amino-40-imidazolinones).
ꢂ
C14H17N5O2 1/3H2O: C, 57.33; H, 6.07; N, 23.88.
Found: C, 57.17; H, 5.60; N, 23.51.
Following the procedures described above for 35 and
substituting the following substituted quinazolyl-2-
guanidines (Formula III, Scheme 1) for 4-methyl-6-
methoxyquinazolyl-2-guanidine, the corresponding sub-
stituted 4-methylquinazolyl-2-(20-amino-40-imidazolin-
one) (32, 33, 36, 37, and 38) were obtained with the
General procedure for the preparation of substituted 4-
methylquinazolyl-2-N,N-diacylguanidines.
Following the procedures described for 4-methylquin-
azolyl-2-N-acylguanidines except 0.002 mol of the
anhydride was used, and substituting the following
substituted quinazolyl-2-guanidines (Formula III,
Scheme 1) for 4-methyl-6-methoxyquinazolyl-2-guani-
dine, the corresponding substituted 4-methylquinazolyl-
2-N,N-diacylguanidines (Formula V) were obtained
with the listed spectral and physical properties.
10
reported spectraland physicalproperties.
4-Methyl-6-methoxyquinazolyl-2-N-methylguanidine (13).
A mixture of 4-methyl-6-methoxyquinazolyl-2-guani-
dine (0.01 mol) and iodomethane (0.015 mol) in 30 mL
of dry dimethylformamide was heated under reflux for
10 h. The mixture was allowed to cool, the precipitate

84
T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
was isolated by filtration and washed with acetone
yielding the desired product.11
J=9.6 Hz), 10.81 (s, 1H), 11.95 (s, 1H). MS (CI/NH3)
m/z 431 (M+H+)+.
6-Methyl-4-methoxyquinazolyl-2-guanidine. A suspen-
sion of 4-hydroxy-6-methylquinazolyl-2-guanidine (5 g)
in 100 mL of methanolwas treated with 100 mg of
p-toluenesulfonic acid and refluxed for 15 h. The pre-
cipitate was isolated by filtration, washed with water,
dried and recrystallized from dimethylformamide yield-
ing 4 g (80% yield) of the desired 6-methyl-4-methoxy-
qunolyl-2-guanidine. Mp 265–266 ꢁC.
N-(7-Methoxy-4-methyl-quinolin-2-yl)-guanidine (52)
and N-tert-butyloxycarbonyl-N’-(7-Methoxy-4-methyl-
quinolin-2-yl)-guanidine (53). A solution of 51 (8 mg,
0.019 mmol) in 5% trifluoroacetic acid in CH2Cl2 (1 mL)
was stirred for 2 h at room temperature. The solvent was
evaporated under reduced pressure and the residue was
purified by preparative thin-layer chromatography (ethyl
acetate/hexanes/methanol=10:10:1) to give 52 (0.8 mg,
30%) and 53 (1.6 mg, 57%) as a white solid.
Preparation of substituted quinolines
52. 1H NMR (CD3OD) d 2.68 (s, 3H), 3.95 (s, 3H), 6.82
(d, 1H, J=0.8 Hz), 7.19 (dd, 1H, J=2.6, 9.2 Hz), 7.38
(d, 1H, J=2.6 Hz), 7.95 (d, 1H, J=9.2 Hz). MS (CI/
NH3) m/z 231 (M+H+)+. FAB+ exact mass: calcd for
C12H15N4O 231.1246. Found: 231.1238.
N-(7-Methoxy-4-methyl-quinolin-2-yl)-acetamide
(49).
To a solution of 2-amino-4-methyl-7-methoxyquinoline
(9.4 mg, 0.04 mmol) in CH2Cl2 (1 mL) was sequentially
added triethylamine (14 mL, 0.10 mmol) and acetic
anhydride (6.6 mL, 0.07 mmol) and DMAP (1 mg, 0.008
mmol). The reaction mixture was stirred at room tem-
perature for 1 h, and the solvent was removed under
a stream of nitrogen. The residue was purified by
preparative thin layer chromatography (ethyl acetate/
hexanes/methanol=10/10/1) to give 49 (9.8 mg, 84%)
53. 1H NMR (CD3OD) d 1.55 (s, 9H), 2.66 (s, 3H), 3.97
(s, 3H), 6.85 (s, 1H), 7.15 (dd, 1H, J=2.6, 9.2 Hz), 7.32
(d, 1H, J=2.6 Hz), 7.91 (d, 1H, J=9.2 Hz). MS (CI/
NH3) m/z 331 (M+H+)+. FAB+ exact mass: calcd for
C17H23N4O3 331.1770. Found: 331.1761.
1
as a white solid. H NMR (CDCl3) d 2.25 (s, 3H), 2.68
(s, 3H), 3.94 (s, 3H), 7.13 (s, 1H), 7.82 (dd, 1H, J=1.1,
8.4 Hz), 8.17 (s, 1H), 9.16 (bs, 1H). MS (CI/NH3) m/z
231 (M+H+)+. FAB+ exact mass: calcd for
C13H15N2O2 231.1134. Found: 231.1138.
Pharmacology
R-PIA, and 2-chloroadenosine were purchased from
Sigma-RBI (St. Louis, MO, USA). Other synthetic reagents
were purchased from Aldrich (Milwaukee, WI, USA).
N-(7-Methoxy-4-methyl-quinolin-2-yl)-carbamic
acid
methyl ester (50). To a solution of 2-amino-4-methyl-7-
methoxyquinoline (10 mg, 0.053 mmol) in CH2Cl2 (1
mL) was added triethylamine (14 mL, 0.1 mmol) and
methylchol roformate (6 mL, 0.08 mmol). The reaction
mixture was stirred at room temperature for 1 h, and the
solvent was removed by a nitrogen stream. The residue
was purified by preparative thin layer chromatography
(ethylacetate:hexanes:methano=l 10:15:1) to give 50
(10.5 mg, 81%) as a white solid. 1H NMR (CDCl3) d 2.67
(d, 3H, J=0.8 Hz), 3.83 (s, 3H), 3.93 (s, 3H), 7.09 (dd,
1H, J=2.6, 9.1 Hz), 7.14 (d, 1H, J=2.5 Hz), 7.59 (bs,
1H), 7.81 (d, 1H, J=9.1 Hz), 7.94 (s, 1H). MS (CI/NH3)
m/z 247 (M+H+)+. FAB+ exact mass: calcd for
C13H15N2O3 247.1083. Found: 247.1072.
Ki values of quinazoline derivatives were determined in
displacement of binding of the non-selective radioligand
[3H]ZM241385 (4 - (2 - [7 - amino - 2 - {furyl}{1,2,4}-
triazolo{2,3 - a}{1,3,5}triazin - 5 - ylaminoethyl) - phenol
(Tocris, Ballwin, MO, USA) at the human A2BAR stably
expressed in HEK-293 cell membranes.7 In order to deter-
mine selectivity, the compounds were evaluated using
standard binding assays at A1, A2A, and A3ARs. The
screening2 utilized rat brain A1/A2AARs (with radioligands
[3H]R-PIA, Amersham, Arlington Heights, IL, USA and
[3H]CGS21680, Perkin–Elmer, Boston, MA, USA8
The functionaleffects in inhibiting the rise in intrace-l
lular calcium elicited by the non-selective agonist
NECA in HEK cells expressing the human AAR were
examined, by methods described.6 The test compound
was incubated with cells at 37ꢁ for 2 min. Then the cells
(1 million in 2 mL) were transferred to a stirred cuvette
maintained at 37 ꢁC within an Aminco SLM 8000
spectrofluorometer (SML instruments, Urbana, IL,
USA). The ratios of indo-1 fluorescence obtained at
400 and 485 nm (excitation, 332 nm) was recorded using
a slit width of 4 nm. NECA was added after a 100 s
equilibration period.
N,N0-Di-tert-butyloxycarbonyl-N00-(7-Methoxy-4-methyl-
quinolin-2-yl)-guanidine (51). 2-Amino-4-methyl-7-meth-
oxyquinoline (19 mg, 0.1 mmol) was added to a solution
of di-BOC-protected-triflylguanidine (36 mg, 0.1 mmol)
and triethylamine (16 mL, 0.12 mmol) in CH2Cl2 (1 mL),
and the mixture was stirred at refluxing temperature
until all triflylguanidine was consumed (the reaction was
monitored by TLC). After completion of the reaction,
the mixture was diluted with CH2Cl2 (2 mL) and the
organic layer was washed with 2 M sodium bisulfate,
saturated sodium bicarbonate, brine, and dried over
sodium sulfate, filtered, concentrated in vacuo. The
residue was purified by preparative thin layer chroma-
tography (ethylacetate/hexanes/methano=l 10:30:1) to
give 51 (16 mg, 40%) as a white solid. 1H NMR
(DMSO-d6) d 1.52 (s, 9H), 2.63 (s, 3H), 3.89 (s, 3H),
7.15–7.19 (m, 1H), 7.19 (s, 1H), 7.93 (s, 1H), 7.94 (d, 1H,
Acknowledgements
We are gratefulfor the hepl of Sergey Shorshnev of
ChemBridge Moscow for his help with NMR spectral
interpretation.

T. R. Webb et al. / Bioorg. Med. Chem. 11 (2003) 77–85
85
References and Notes
9. Cliffe, W. H. J. Chem. Soc. 1933, 35, 1327.
10. Knovenagel, E. Ber. 1921, 54B, 1722.
¨
1. Jacobson, K. A.; Knutsen, L. J. S. In P1 and P2 Purine and
Pyrimidine Receptors; Abbracchio, M. P., Williams, M., Eds.;
Handbook of Experimental Pharmacology, Volume 151/I: Pur-
inergic and Pyrimidinergic Signaling I; Springer: Berlin, Ger-
many, 2001; p 129.
2. Kim, Y.-C.; Ji, X.-D.; Melman, N.; Linden, J.; Jacobson,
K. A. J. Med. Chem. 2000, 43, 1165.
3. Webb, T. R.; Melman, N.; Lvovskiy, D.; Ji, X.-D.; Jacob-
son, K. A. Bioorg. Med. Chem. Lett. 2000, 10, 31.
4. Feoktistov, I.; Polosa, R.; Holgate, S. T.; Biaggioni, I.
Trends Pharmacol. Sci. 1998, 19, 148.
5. Kirshenbaum, A. S.; Hettinger, B.; Day, Y.-J.; Gilfillan, A.
M.; Metcalfe, D. D.; Kim, Y.-C.; Linden, J.; Jacobson, K. A.
2001, Amer. Acad. Allergy, Asthma and Immunol., 57th Annual
Meeting, 2001 (abstr 2000803).
6. Auchampach, J. A.; Jin, J.; Wan, T. C.; Caughey, G. H.;
Linden, J. Mol. Pharmacol. 1998, 52, 846.
7. Ji, X.-D.; Jacobson, K. A. Drug Des. Discov. 1999, 16, 217.
8. Olah, M. E.; Gallo-Rodriguez, C.; Jacobson, K. A.; Stiles,
G. L. Mol. Pharmacol. 1994, 45, 978.
11. Shikhaliev, Kh. S.; Falaleyev, A. V.; Ermolova, G. I.;
Solov’ev, A. S. Chem. Heterocycl. Comp. 1999, 35, 818.
12. Shikhaliev, Kh. S.; Falaleyev, A. V.; Ermolova, G. I.;
Solov’ev, A. S. Khim. Khim. Tekhnol. 1998, 41, 116.
13. Fozard, J. R.; Hannon, J. P. Pulm. Pharmacol. Ther. 1999,
12, 111.
14. Salvatore, C. A.; Tilley, S. L.; Latour, A. M.; Fletcher, D. S.;
Koller, B. H.; Jacobson, M. A. J. Biol. Chem. 2000, 275, 4429.
15. Feoktistov, I.; Polosa, R.; Holgate, S. T.; Biaggioni, I.
Trends Pharmacol. Sci. 1998, 19, 148.
16. Jacobson, K. A.; IJzerman, A. P.; Linden, J. Drug Devel.
Res. 1999, 47, 45.
17. Arya, V. P.; David, J.; Grewal, R. S.; Marathe, S. B.;
Patil, S. D. Indian J. Chem. 1977, 15B, 1129.
18. Rosowsky, A.; Nadel, M. E.; Modest, E. J. J. Heterocycl.
Chem. 1972, 9, 625.
19. van Muijlwijk-Koezen, J. E.; Timmerman, H.; Link, R.;
van der Goot, H.; Ijerman, A. P. J. Med. Chem. 1998, 41, 3994.
20. Feichtinger, K.; Zapf, C.; Sings, H. L.; Goodman, M. J.
Org. Chem. 1998, 63, 3804.