Applications of parallel synthesis to lead optimization

Article abstract of DOI:10.2533/000942903777679271

Parallel synthesis of focused compound libraries for hit confirmation and lead optimization are certainly important drivers for shortening the lead discovery phase in the pharmaceutical and crop protection industries. In this article we show with permission of Roche and Syngenta three real case studies where Polyphor synthesized focused libraries for lead validation and optimization using high-throughput parallel synthesis and purification techniques. The three examples differ significantly in the synthetic strategies which were employed as well as in the chemical complexity of the final products. A multigeneration approach towards insecticidal triazines, the application of a sequential three-component reaction towards insecticidal and fungicidal thiazoles and finally a multistep synthesis approach of advanced building blocks followed by a two-step final derivatization towards novel antiviral N-hydroxy-indolin-2-ones are presented. In all cases 100-200 analogues were synthesized using parallel synthesis in solution followed by purification of the final products by parallel flash or high-throughput (unattended) HPLC (coupled to MS) within four months. Promising biological results were obtained in all three cases.

Full text of DOI:10.2533/000942903777679271

COMBINATORIAL CHEMISTRY
262
CHIMIA 2003, 57, No. 5
Chimia 57 (2003) 262–269
© Schweizerische Chemische Gesellschaft
ISSN 0009–4293
Applications of Parallel Synthesis to
Lead Optimization
Michael Altorfera, Philipp Ermerta, Jürg Fässlera, Saleem Farooqb, Elke Hillesheimb,
André Jeanguenatb, Klaus Klumppc, Peter Maienfischb, Joseph A. Martinc, John H. Merrettc,
Kevin E.B. Parkesc, Jean-Pierre Obrechta, Thomas Pitternab, and Daniel Obrechta*
Dedicated to Prof. Dr. A. Vasella on the occasion of his 60th birthday
Abstract: Parallel synthesis of focused compound libraries for hit confirmation and lead optimization are
certainly important drivers for shortening the lead discovery phase in the pharmaceutical and crop protection
industries. In this article we show with permission of Roche and Syngenta three real case studies where
Polyphor synthesized focused libraries for lead validation and optimization using high-throughput parallel
synthesis and purification techniques. The three examples differ significantly in the synthetic strategies which
were employed as well as in the chemical complexity of the final products. A multigeneration approach
towards insecticidal triazines, the application of a sequential three-component reaction towards insecticidal
and fungicidal thiazoles and finally a multistep synthesis approach of advanced building blocks followed by
a two-step final derivatization towards novel antiviral N-hydroxy-indolin-2-ones are presented. In all cases
100–200 analogues were synthesized using parallel synthesis in solution followed by purification of the final
products by parallel flash or high-throughput (unattended) HPLC (coupled to MS) within four months. Promising
biological results were obtained in all three cases.
Keywords: High-throughput purification · N-hydroxy-indolin-2-ones as influenza endonuclease inhibitors ·
Insecticidal and fungicidal thiazoles · Insecticidal triazines · Lead validation and optimization ·
Parallel synthesis
1. Introduction
substance library consisting of 15’000
small molecules for general and random different types of parallel synthesis ap-
In addition these examples show three
Since 1996 Polyphor has supported the screening. 15’000 is a comparably small proaches [2] that Polyphor has successfully
pharmaceutical and chemical industry in number, however, this library is derived applied.
its efforts to identify and optimize lead from 63 drug-like scaffolds and is thus very a) multigeneration approaches;
compounds [1].
The process usually starts with high- pharma industry’s internal depositories. As
throughput screening (HTS) of a substance a consequence of HTS the pharmaceutical c) multistep synthesis of advanced build-
diverse and a good complement to the b) application of sequential multi-compo-
nent reactions and
library of small molecules (MW < 600) on and crop protection companies have now
a new target as schematically shown in identified many hits. Hit-to-lead and lead
Fig. 1. In a second step, focused libraries optimization has now become the bottle
are used to optimize the lead structures neck and thus a growing area for out-
ing blocks followed by one or two step
parallel conversion into final products.
identified in the HTS. Polyphor offers a sourcing and an opportunity for start-up 2. Multigeneration Approach
companies.
However, only a small number of com-
panies has overcome the entry barrier to
Towards Insecticidal Triazines [3]
In collaboration with André Jeanguenat
this market, since it needs time and persist- and Saleem Farooq; Syngenta Crop Pro-
ence to gain the confidence and trust of life tection AG, Research and Technology,
science companies, to prove the capability CH–4002 Basel
to solve challenging chemical problems
efficiently and to keep to the time lines. Be- 2.1. Introduction
*Correspondence: Dr. D. Obrechta
Tel.: +41 61 486 98 98
Fax: +41 61 486 98 99
E-Mail: daniel.obrecht@polyphor.com
aPolyphor AG
Gewerbestrasse 14
CH–4123 Allschwil
cause most projects are highly confidential,
In a random screening program, 1,2,4-
bSyngenta Crop Protection AG
Research and Technology,
CH–4002 Basel
it is difficult to demonstrate a track record. triazines such as 1 and 2 were identified as
We would like to take this opportunity to novel fungicidal lead compounds (Fig. 2).
present, with the permission of Syngenta They displayed an interesting activity
and Roche, three real case studies of fo- against species such as Phytophthora in-
cRoche Discovery Welwyn
40 Broadwater Road
Welwyn Garden City
Herts, AL 3AY, UK
cused libraries.
festans (late blight), Pyrenophora teres (net

COMBINATORIAL CHEMISTRY
263
CHIMIA 2003, 57, No. 5
blotch) and Septoria nodum (glume blotch)
with a new and unknown mode of action.
Substituting the triazine position 3 with an
appropriate phenyl group gave a new class
of compounds (e.g. 3) with a very potent
insecticidal activity, mainly against lepi-
doptera, as well as against some sucking
insects.
Libraries of
small molecules for general and
random screening
High throughput
screening
Focused and customized
libraries of small molecules
for hit confirmation
2.2. Target and Design
Libraries of
Protein Epitope Mimetics
(PEM)
To evaluate the potential of these new
classes of compounds, a library of unsym-
metrical triazines of general formulae I and
II was designed as shown in Fig. 3. In order
to determine which substitution pattern
was most important for biological activity,
it was of particular interest to design a se-
lective synthesis for 3,6-disubstituted and
for 3,5-disubstituted triazines.
and lead optimization
yes
no
Hit identification
Fig. 1. Polyphor libraries in the lead finding and optimization process [1]
2.3. Synthesis
CF3
Based on the library design we devised
a multigeneration approach for R1=SMe;
SOMe; SO2Me and OMe as all intermedi-
ate products were part of the general library
as shown in Scheme 1. Condensation of
(S)-methyl-isothiosemicarbazide 4 [4] with
glyoxals 6 gave separable mixtures of the
two triazines 7 and 8 (7 being generally the
major product) in 60–80% yield. Glyoxals
6 were obtained from the corresponding
methyl ketones 5 by oxidation with SeO2
in a mixture of dioxane and water at 80 °C.
Conversion of 7 and 8 into sulfoxides 9 and
11 and sulfones 10 and 12, and subsequent
transformation into products 13 and 14
was performed using standard procedures
starting from roughly 300 mg of 7 and 8.
Scheme 2 shows the synthesis of the
parent triazines of type I and II with R1:
2,4-(Cl)2-C6H3 and 2,6-(F)2-C6H3. Reac-
tion of amidrazones 15 with glyoxals 6
gave again mixtures of triazines I and II
(~1:9) which were separated by flash chro-
matography.
F
N
F
N
N
N
N
N
N
MeO2S
N
MeO2S
N
1
2
3
Fig. 2. Initial hits based on the triazine scaffold
R2
R2
N
N
N
N
R1
N
R1
N
I
II
In order to circumvent problems of
separation and low yield of isomer I, we
investigated new regioselective approaches
to triazines I. Literature reports [5] sug-
gested that oximes 16 might be suitable pre-
cursors for the regioselective synthesis of I.
Oximes 16 could be efficiently synthesized
by using a trans-oximation reaction with
acetone oxime in dioxane and an aqueous
NaH2PO4 solution as a mild acid catalyst.
After screening numerous reaction con-
ditions in parallel format, we found that
mixing 15 and 16 in 4N HCl/dioxane
(4–5 equiv.) in n-butanol at 80 °C gave
75–85% of triazines I. Reaction of 17 and
18 gave 19 (82%) as shown in Scheme 3.
In summary, a 180-member library of
trazines of type I and II was synthesized
within four months using the synthetic
F
Cl
R1: SMe; SOMe; SO2Me; OMe;
;
Cl
F
Cl
S
R2:
;
tBu ;
;
CF3
Cl
N
N
;
;
R (Me, Cl)
S
CF3
N
S
;
Cl
;
;
N
O
Cl
H
Fig. 3. Design of triazine libraries I and II

COMBINATORIAL CHEMISTRY
264
CHIMIA 2003, 57, No. 5
approaches described in this chapter. In ad-
dition the stage was set for more focused
triazine libraries of type I which were bio-
logically more interesting than the parent
triazine II.
NH2 x HI
NH2
O
O
ii
S
i
N
+
R2
R2
CHO
H2N
NHNH2
MeS
6
5
4
thiosemicarbazide
iii
2.4. Biological Results
and Conclusion
R2
In the fungicidal screening, compounds
with a substituent in position 3, such as
S(O)nMe (n = 0, 1, 2) and OMe, displayed
some interesting activities. Main targets
were oomycetes such as Phytophthora in-
festans (late blight), Plasmopora viticola
(grape downy mildew), and Pythium ultimum
(ashy stem blight) as well as Pyrenophora
teres (net blotch) and Septoria (glume
blotch). Among the compounds with insec-
ticidal activity, 3,6-disubstituted triazines
of type I displayed a very potent activity
against key chewing insects such as
Spodoptera littoralis (beet armyworm),
Heliothis virescens (tobacco budworm), and
Plutella xylostella (diamondback moth)
and also against some sucking pests, at a
level comparable to the best commercial
standards.
R2
N
N
N
N
MeS
iv
N
MeS
iv
N
7
8
v
v
R2
R2
R2
R2
N
N
N
N
Me
+
N
Me
N
Me
+
N
vi
S
N
Me
N
S
N
S
N
S
N
O
O
O
O
O
-
11
12
vi
O
9
10
-
R2
R2
N
N
N
N
MeO
N
MeO
N
14
13
Furthermore, this example shows that
multigeneration approaches can be per-
formed in parallel format if optimized reac-
tion conditions can be applied. Thus the
overall number of synthetic steps per library
compound can be significantly reduced as
compared to a linear approach.
i: MeI, dioxane; ii: SeO2, dioxane/H2O, 80 °C; iii: DIEA, DMF or CH2Cl2; then separation;
iv: mCPBA (1.1 equiv.), NaOAc, CH2Cl2; v: mCPBA (2.5 equiv.), NaOAc, CH2Cl2; vi: NaOMe, THF
Scheme 1.
R2
N
NH2 x HX
NH2
N
O
3. Insecticidal and Fungicidal
Thiazoles Using a Novel Sequential
Multicomponent Reaction [6]
+
+
i
N
N
R1
R2
I
+
(~9:1)
(~5:1)
CHO
N
R1
6
15 (X: Cl, Br)
II
ii
In collaboration with Elke Hillesheim,
Thomas Pitterna, and Peter Maienfisch,
Syngenta Crop Protection AG, Research
and Technology, CH-4002 Basel
NH2 x HX
R2
N
O
iii
R1
R2
II
NH2
15 (X: Cl, Br)
CH=NOH
+
N
R1
N
16
I
3.1. Introduction
In a random screening program, the
N-thiazol-2-yl-benzamide 20 was identi-
fied as a novel insecticidal lead compound.
Compound 20 as well as its chloro analogue
21 (Fig. 4) possess good activity against
many chewing and sucking insects and are
strong uncouplers of oxidative phosphory-
lation with IC50 values in the low nM range.
i: DIEA, DMF; ii: acetone oxime, dioxane, 1M NaH2PO4, 60 °C; iii: 4N HCl/dioxane, n-butanol, 80 °C
Scheme 2.
MeO
Br
NH2
NH2
O
F
N
F
N
F
x HCl
i
CH=NOH
OMe
N
N
NC
O
+
S
20: R = Br
21: R = Cl
Br
R
F
N
H
Cl
N
17
18
19
i: 4N HCl/dioxane, n-butanol, 80 °C
Fig. 4. Initial hits based on the aminothiazole
scaffold
Scheme 3.

COMBINATORIAL CHEMISTRY
265
CHIMIA 2003, 57, No. 5
To further exploit the potential of this
chemical class, libraries of compounds of
types III and IV (Fig. 5) were designed
based on a novel sequential three compo-
nent reaction [7] developed by Polyphor.
In library III, R1 comprised aromatic
and heteroaromatic substituents as well
as CF3, whereas in library IV, R1 was
chosen as NH2 (electron-donating group,
H-bond donor). In both libraries E (elec-
tron-withdrawing group) was selected from
CN, CH3CO, 2,4-(Cl)2C6H3CO, 4-NO2-
C6H4CO, CF3CO and 4-CF3-C6H4CO. For
groups Z we chose CO, SO2 and direct
connection of the nitrogen to the 3-chloro-
phenyl group, and R2 was chosen from Me
and EtOCH2 groups.
E
E
S
S
Z
Cl
Z
Cl
N
R1
N
H2N
N
N
R2
R2
Library IV
Library III
R1: 4-Cl-C6H4; 2-thienyl; CF3
E: CN; CH3CO; 2,4-(Cl)2-C6H3CO; CF3CO; 4-NO2-C6H4CO; 4-CF3-C6H4CO
R2: H, Me, CH2OEt
Z: CO, SO2, direct bond between NR2 and 3-chlorophenyl group
Fig. 5. Design of aminothiazole libraries III and IV
3.2. Synthesis
For the synthesis of the libraries III and
IV we took advantage of a novel sequential
three-component reaction starting from
amidines 22 or isothiourea 25, isothio-
cyanates 23 and ECH2X (24, E: electron-
withdrawing group) as shown in Scheme 4.
Yields ranged from 40–80%.
R3N=C=S
E
NH x HCl
i
23
E-CH2-X
24
S
+
R1
NHR3
R1
NH2
Scheme 5 shows a possible mechanism
for the formation of thiazoles of type 27 via
intermediates V, VI and VII. Thiouronium
salt 25 is first reacted with isothiocyanates
23 in the presence of DIEAor DBU in DMF
to form the reactive intermediates of type
V, which can be isolated if required or
reacted in situ with 24 in the presence of
a second equivalent of base. Alkylation
proceeds most probably at the sulfur atom
to form intermediates VI, which react via
an intramolecular Knoevenagel-type con-
densation via intermediates VII to form
after elimination of 1-mercapto-hexane
products of type 27. Scheme 6 describes
the three-component reaction of 22 or 25
with 2,4-dimethoxy-benzyl-isothiocyanate
28 and 24 to yield intermediate thiazoles
29. Thiazoles 29 were either treated with
3-chlorobenzene sulfonyl chloride followed
by removal of the 2,4-dimethoxy-benzyl
group with trifluoroacetic acid in CH2Cl2
to yield sulfonamides 30 or treated with
3-chlorobenzoyl chloride followed by de-
protection of the intermediates with trifluo-
roacetic acid in CH2Cl2 to yield amides 31.
These transformations were performed
in parallel format with purification of 30
and 31. Overall yields were in the range of
40–70%. Subsequent alkylation of 30 and
31 with either MeI or ClCH2OEt in DMF
and K2CO3 gave products 32 and 33, which
were again purified by parallel flash chro-
matography. A library of roughly 150 com-
pounds was prepared by Polyphor and
screened by Syngenta.
N
22
26 (Library III)
R3N=C=S
NH x HBr
E
S
i
23
+
n-hexyl-S
NH2
NHR3
E-CH2-X
H2N
N
25
24
27 (Library IV)
R1: 4-Cl-C6H4, thiophen-2-yl, CF3 ; R3: 3-Cl-C6H4; 2,4-dimethoxy-benzyl;
X: Cl, Br, I
i: 22 or 25 and 23, DMF, DBU or DIEA; then 24, DBU or DIEA
Scheme 4.
E
NH
S
24
NH
S
25 +
23
NHR3
n-hexyl-S
N
NHR3
n-hexyl-S
N
H
VI
V
H
E
B:
NHR3
S
27
H2N
n-hexyl-S
N
H+
B: DIEA or DBU
VII
Scheme 5.

COMBINATORIAL CHEMISTRY
266
CHIMIA 2003, 57, No. 5
3.3. Biological Results and
Conclusions
NH x HCl
NH2
NH x HBr
The library was evaluated in a standard
HTS screening for insecticidal, herbicidal
and fungicidal activity. None of these com-
pounds reached the insecticidal activity of
the lead structures 20 and 21. However,
some thiazoles of type III displayed in-
teresting fungicidal (nine compounds) and
herbicidal (three compounds) activity. The
detected fungicidal activity was mainly on
one pathogen, Pyricularia oryzae (rice
blast), whereas herbicidal activity was ob-
served on Stellaria media (common chick-
weed). The most potent compounds were
those with R1 = CF3 and E = CN, 4-NO2-
C6H4-CO or 2,4-(Cl)2C6H3CO. Among all
compounds prepared, best activity was
observed for thiazole 34 (Fig. 6).
or
E
R1
OMe
n-hexyl-S
NH2
i, ii
S
22
25
H2N/R1
+
N
N
OMe
H
OMe
29
N=C=S
MeO
v, iv
iii, iv
28
E
E
O
S
O
S
O
Cl
S
Cl
H2N/R1
N
H
H2N/R1
N
H
N
N
vi
vi
E
E
O
O
S
O
S
S
Cl
Cl
H2N/R1
N
R2
H2N/R1
N
R2
N
NC
O
N
S
Cl
33
32
N
H
F3C
N
i: DIEA or DBU, DMF; ii: DIEA or DBU, DMF, ECH 2X (24); iii: 3-Cl-C6H4COCl, CH2Cl2,
pyridine, DMAP; iv: TFA, CH2Cl2, H2O; v: 3-Cl-C6H4SO2Cl, CH2Cl2, pyridine, DMAP;
vi: K2CO3, DMF, MeI or ClCH2OEt
34
Fig. 6. Most active analogue 34
Scheme 6.
4. Parallel Synthesis of a New
Class of Influenza Endonuclease
Inhibitors [8][9]
processing enzymes containing two diva- a parallel synthesis approach as indicated
lent metal ions, particularly Mn2+ and Mg2+ in Fig. 8:
ions.
HTS for endonuclease inhibitors gave
–
The β-ketoamide and hydroxamic acid
groups are both acidic and potentially
troublesome functional groups, for which
suitable protection would be required.
The final molecules would be highly
polar and good isolation and purifica-
tion methods would be required.
In collaboration with K. Klumpp, J.A. rise to two different lead series belonging
Martin, J.H. Merrett and K.E.B. Parkes, to the ketobutanoates of type 35 and the
Roche Discovery Welwyn, 40 Broadwater N-hydroxy-imides of type 36 as shown in
Road, Welwyn Garden City, Herts, AL7 Fig. 7. Both lead series are potential chela-
–
3AY, UK
tors of divalent cations. Based on these
and further molecular modeling studies the
Fig. 9 shows that the three sub-libraries
4.1. Introduction
Roche group proposed tetramic acid de- of type 37, 38 and 39 could be synthesized
Influenza virus infects an estimated rivatives derived from N-hydroxy-indolin- from a common precursor of type IX.
120 million people in US, Europe and Japan 2-ones of type VIII as a novel pharma- The synthesis of the key building blocks
in a typical year. The influenza endonucle- cophore combining structural features of of type IX is shown in Schemes 7–10.
ase is an attractive target for several reasons: both the ketobutanoates 35 and the
a) It is a key component of the viral tran- N-hydroxy-imides 36.
The synthesis started from commercial
aniline derivative 40, which was Boc-pro-
scription mechanism, which has no cel-
Based on the general formula VIII, tected to give 41 as shown in Scheme 7.
lular counterparts and should therefore Polyphor proposed to synthesize three sub- Treatment of 41 with 2.2 equiv. of sodium
provide a good potential for discovering libraries 37, 38 and 39 as shown in Fig. 7. dimethylmalonate in DMSO gave a high
selective, non-toxic drugs.
A literature review indicated that com- yield of the nucleophilic substitution
b) In contrast the neuraminidase inhibitors pounds of type 37–39 had not been synthe- product 42, which was saponified using aq.
that do not prevent the formation of new sized before. As discussed with Roche we 2N NaOH/MeOH at 80 °C to give after
virus particles, but interfere with virus planned to have R1 as the low variation sub- acidification the phenylacetic acid 43 in
release from host cells and are therefore stituent (R1: Me, Et and Ph) and R3, R4 and 60–70% overall yield from 40. In order
virustatic, endonuclease inhibitors, due R5 as high variation substituents (typically to synthesize the whole library 50g of acid
to the block of viral transcription, are 25 or 50 variations) as shown in Fig. 8.
expected to have a virucidal effect.
43 was required.
The conversion of 43 to 44 offered two
major challenges:
4.2. Target Mechanism and
Inhibitor Design
Endonuclease of influenza RNA poly- 2-one derivatives VIII would offer several
merase belongs to the group of phosphate- synthetic challenges especially in view of
4.3. Synthesis
a) Partial reduction of the nitro group into
the corresponding hydroxylamine fol-
lowed by cyclization to the N-hydroxy-
indolin-2-one.
We anticipated that N-hydroxy-indolin-

COMBINATORIAL CHEMISTRY
267
CHIMIA 2003, 57, No. 5
b) Identification of a suitable protective
group for the hydroxamic acid group
which would be stable enough for the
subsequent steps and could be removed
quantitatively in the last step.
N-Hydroxy-imides
N
Ketobutanoates
O
OH
OH
Among the methods known for the
reduction of aromatic nitro groups into a
hydroxylamine group we focused on the
hydrogenation procedure described by
Kende [10] using Pt/C and DMSO. After
screening several reaction conditions we
found that using a solvent mixture of
EtOH/THF 2:1 and DMSO in the pres-
ence of Pt/C and H2, followed by cycliza-
tion using acetic acid gave the desired
N-hydroxy-indolin-2-one derivative in
roughly 65% yield along with 5–10% of the
corresponding lactam, obtained from over-
reduction of the hydroxyl amine to the ani-
line, as a mixture of solids of low solubility
which was difficult to further purify. This
mixture was treated directly with pivaloyl
chloride in CH2Cl2 in the presence of di-
isopropylethylamine (DIEA) to give 44 in
high yield after chromatography on SiO2
along with the lactam. The pivaloyl group
not only proved to be an excellent protec-
tive group for the hydroxamic acid moiety
but also made all derivatives highly soluble
and amenable to chromatographic purifi-
cation. With the key building block 44 in
hand we tried to introduce the β-ketoamide
group as shown in Scheme 8.
O
O
N
O
35
OH
Merck: IC50: 21.3µM
36
Flutimide: fungal methabolite
Merck: IC50: 5.5 µM
tetramic acids derived from N-hydroxy-indolin-2-ones
R1
R1
OH
O
R2
R2
OH
O
N
OH
N
OH
VIII
R1
N
R1
R1
N
OH
O
OH
OH
O
O
O
R5
O
O
O
S
N
H
R3
N
N
R4HN
N
OH
H
H
OH
OH
38
39
37
In a first series of experiments we
attempted the conversion of 44 into 45a–c
applying reaction conditions known for
similar transformations such as 1.1 equiv.
of dicyclohexylcarbodiimide, the corre-
sponding acid R1COOH (R1: Me, Et, Ph)
and N,N-dimethylaminopyridine (DMAP)
in CH2Cl2. Under these reaction conditions
only very low yields (<10%) of compounds
45a–c could be obtained. In addition, it
turned out that the separation of 45a–c
from unreacted 44 was very difficult. After
screening several reaction conditions in
parallel format it was found that reaction
of 44 with 2.5 equiv. of the corresponding
R1COX (X= Cl or R1CO2) in a mixture
of DIEA and DMAP in CH2Cl2 gave good
yields of the corresponding compounds
46a–c, which could be easily purified. We
believe that the impressive improvement
of the yields is based on the fact that the in-
termediate products 45a–c are immediately
acylated by the excess of reagent under the
reaction conditions to form irreversibly
46a–c. These products were obtained as
mixtures of (E)- and (Z)-isomers. It turned
out that compounds 46a (R2 = Me) and 46b
(R2 = Et) were not stable enough for the
subsequent transformations and had to
be transacylated into derivatives 47a and
47b using pivaloyl chloride and NaCN or
Fig. 7. Design of the N-hydroxy-indolin-2-one scaffold
R1
low variation site
β-keto amide moiety
OH
O
N
high variation site
RHN
hydroxamic acid moiety
OH
Fig. 8. Potential sites of variation in the N-hydroxy-indolin-2-one scaffold
R1
OCOR'
37, 38, and 39
O
N
H2N
IX
OPG
key precursors for parallel synthesis
PG: protective group
Fig. 9. Key precursors for parallel synthesis

COMBINATORIAL CHEMISTRY
268
CHIMIA 2003, 57, No. 5
tetrabutylammonium cyanide in a mixture
of pyridine and CH2Cl2. In contrast 46c
(R2 = Ph) was stable enough for the subse-
quent transformations.
CH(COOMe)2
NO2
F
F
NO2
ii
i
BocHN
BocHN
NO2
H2N
Cleavage of the Boc-group of 46c, 47a
and 47b was performed with 4N HCl/diox-
ane or with trifluoroacetic acid in CH2Cl2,
followed by extraction under basic condi-
tions to yield the building blocks 48a–c as
shown in Scheme 9. Building blocks 48a–c
constituted the key precursors for the par-
allel synthesis of the three sub-libraries
49a–c, 50a–c and 51a–c and were synthe-
sized on a 5g scale.
The parallel syntheses of the three sub-
libraries (anilides 49a–c, ureas 50a–c and
sulfonamides 51a–c) were performed under
standard conditions (see Scheme 9) and
the products were purified by parallel flash
chromatography on SiO2. At that stage
we aimed at having 100 mg of each of the
intermediates.
The final deprotection of products
49a–c, 50a–c and 51a–c into the corre-
sponding final products 52a–c, 53a–c and
54a–c again proved challenging and the
solution is described in Scheme 10. As
the final products were expected to be high-
ly polar and difficult to purify by chro-
matography we used N,N-dimethylamino-
ethyl amine or N,N-dimethylamino-propyl
amine in CHCl3 or CH2Cl2 as deacylating
agent yielding the corresponding acylated
by-products (Scheme 10) which were sol-
uble in aqueous HCl solution, whereas
the products 52a–c, 53a–c and 54a–c were
soluble in CHCl3 or CH2Cl2. After evapo-
ration of the solvent the products could
be generally obtained in pure form as solids
by precipitation from mixtures of ethyl-
acetate/hexane or by purification on reverse
phase preparative HPLC. Using this proto-
col we were able to deliver 131 compounds
in quantities between 6–71 mg and purities
of 75–99% to Roche for biological testing.
This program was performed within four
months.
41
42
40
iii
COOH
iv, v
O
N
BocHN
BocHN
NO2
OCOtBu
44
43
i: Boc2O, THF, 80 °C; ii: CH2(COOMe)2, NaH, DMSO; iii: aq. NaOH, MeOH, 80 °C;
iv: Pt/C(5%),H2, DMSO, EtOH; then AcOH; v: t-BuCOCl, DIPEA, CH2Cl2
Scheme 7.
O
O
O
O
O
O
R1
N
R1
N
R1
R1COX
(2.5 equiv.)
iii
O
ii
i
N
OPiv
BocHN
BocHN
BocHN
OPiv
OPiv
44
46a-c
47a, 47b
R1
N
OH
a: R1= Me; b: R1= Et; c: R1= Ph
O
BocHN
OPiv
45a-c
i: R1COOH, DCC, DMAP, CH2Cl2; ii: 46a: X = OCOCH3; 46b: X = Cl; 46c: X = Cl, DMAP, DIEA,
CH2Cl2, THF; 0 °C-r.t.; iii: t-BuCOCl, tetrabutylammonium cyanide or NaCN, pyridine, CH2Cl2
Scheme 8.
R1
N
OCOR2
O
O
ii
R3
N
H
OPiv
49a-c
R1
O
O
O
O
O
O
R1
N
R2
OCOR2
O
R1
N
R2
4.4. Biological Results
From 131 compounds tested, 26 com-
pounds inhibited influenza endonuclease
with an IC50 < 50µM. More importantly,
10 compounds (e.g. 55, 56, and 57) showed
a significant antiviral activity in cell cul-
tures as shown in Fig. 10. The most in-
teresting compound of this first series of
compounds was 57 having an EC50 of 6µM.
In summary we have developed a new
approach to N-hydroxy-indolin-2-one de-
rivatives of type VIII amenable to parallel
synthesis and within four months three in-
teresting novel lead compounds (55–57)
were generated from a library of 131 com-
pounds showing significant antiviral activ-
O
i
iii
R4HN
N
H
N
OPiv
H2N
BocHN
50a-c
R1
OPiv
OPiv
47a (R1= Me; R2= tBu)
47b (R1= Et; R2= tBu)
46c (R1= Ph; R2= Ph)
48a-c
OCOR2
O
O
O
iv
S
R5
N
N
OPiv
H
51a-c
i: 4N HCl/dioxane or trifluoroacetic acid CH2Cl2; then extraction with aq. sat. NaHCO solution
3
and CH2Cl2;ii: R3COCl, pyridine (DIEA), DMAP, CH 2Cl2; iii: R4N=C=O, CHCl3 (ethanol-free), 70 °C;
iv: R5SO2Cl, pyridine (DIEA), DMAP, CH 2Cl2
Scheme 9.

COMBINATORIAL CHEMISTRY
269
CHIMIA 2003, 57, No. 5
vantage of common intermediates and
common reactive building blocks and
thus lower the overall synthetic effort
and cost by significantly reducing the
number of steps per final compound.
Parallel approaches can also be success-
fully applied for synthesizing highly
complex products derived from ad-
vanced building blocks obtained by
multi-step synthesis. In these cases the
parallel steps are preferably performed
at the very end of the synthesis as shown
in section 4.
R1
N
R1
OCOR2
OH
O
i
O
R6HN
R6HN
N
OPiv
OH
–
52a-c: R6=R3CO
53a-c: R6=R4NHCO
54a-c: R6=R5SO2
49a-c: R6=R3CO
50a-c: R6=R4NHCO
51a-c: R6=R5SO2
R1: a=Me, b=Et, c=Ph
i: Me2N(CH2)2NH2, CHCl3, 4 °C to rt; then extraction with 1N HCl and CH2Cl2
by-products are soluble in acidic aqueous phase:
O
O
N
N
Acknowledgements
N
N
Ph
H
H
The authors would like to thank the follow-
ing collaborators from Polyphor: Drs. M. Billet,
W. Ganci, P. Jeger, F. Lach, C. Ludin, A. Luther,
T. Maetzke, K. Sekanina, and Ms. F. Abdennour,
C. Abrecht, Y. Bänninger, C. Barth, A. Bickel,
D. Dändliker, S. Dillinger, M. Fischesser,
A. Gomez, L. Kiefer, B. Maret, M. Müller, O.
Roger, A. Ruetsch, A. Salomon, M. Schnecken-
burger, K. Steinmetz; and Mrs H. Behra, L. Cor-
nu, N. Desjonqueres, S. Gable, J.-F. Gosalbes,
S. Luisoli, and F. Tritsch. Furthermore we would
like to thank Sir Jack Baldwin, FRS (University
of Oxford, UK), Prof. Dr. J.A. Robinson
(University of Zürich) and Prof. Dr. A. Vasella
(ETH, Zürich) for stimulating discussions and
valuable advice.
Scheme 10.
Me
Ph
N
OH
O
OH
O
O
O
O
N
OH
S
S
N
H
N
H
OH
56
55
IC50 = 48µM
IC50 = 9µM; EC50 = 21µM
Received: February 14, 2003
Ph
N
OH
O
[1] P. Ermert, D. Obrecht, J.-P. Obrecht,
K. Sekanina, Chimia 2000, 54, 198–202.
[2] R.E. Dolle, K.H. Nelson, Jr., J. Comb.
Chem. 2002, 4, 369–418; D. Obrecht,
J.-M. Villalgordo, ‘Solid-Supported Com-
binatorial and Parallel Synthesis of Small-
Molecular-Weight Compound Libraries’,
Tetrahedron Organic Series, Vol 17, Perg-
amon, 1998; K. Frobel, T. Krämer, Chemie
in unserer Zeit 1996, 30, 270–285.
[3] C.Abrecht, P. Ermert, P. Jeger, D. Obrecht,
A. Jeanguenat, S. Farooq, 10th IUPAC
International Congress of the Chemistry
of Crop Protection, Basel, 2002, Poster 6
(abstract 3a.06).
O
HOOC
N
N
H
OH
H
57
IC50 = 3µM; EC50 = 6µM
compounds show antiviral activity in cell cultures
Fig. 10. Most active library members
ity in whole cells. Based on our synthetic final products. Nevertheless some general
approach optimization of this lead series conclusions can be made:
[4] W.W. Paudler, T.–K. Chen, J. Heterocycl.
Chem. 1970, 7, 767.
should prove straightforward.
–
All three projects were completed within
four months. The time invested per
compound was significantly reduced
by switching from parallel flash chro-
matography to the highly efficient (un-
attended) high-throughput preparative
HPLC-system, operating on normal and
reversed phase.
[5] H. Yamanaka, Heterocycles 1989, 29, 2249.
[6] C. Abrecht, P. Ermert, D. Obrecht, E.
Hillesheim, T. Pitterna, P. Maienfisch,
10th IUPAC International Congress of
the Chemistry of Crop Protection, Basel,
2002, Poster 34 (abstract 3a.34).
[7] D. Obrecht, P. Ermert, Fifth International
Electronic Conference on Synthetic
Organic Chemistry (ECSOC-5); www.
mdpi.org/ecsoc-5/, September 1–30, 2001,
[B0005]
[8] P. Ermert, J. Fässler, K. Klumpp, J.A.
Martin, J.H. Merrett, D. Obrecht, K.E.B.
Parkes, J. Med. Chem. accepted for publi-
cation, 2003.
[9] Presented in part at JCCF14 (Japan
Combinatorial Chemistry Focus Group)
on April 22nd, 2002, Osaka, Japan.
[10] A. Kende, Synth. Commun. 1990, 20,
2133–38.
This example demonstrates that parallel
synthesis approaches can also be success-
fully applied to highly complex and synthet-
ically challenging compound libraries.
5. Summary and Outlook
–
–
Positive screening results were obtained
from all three projects
Parallel synthesis of focused compound
libraries for hit confirmation and lead opti-
mization are certainly important drivers
for shortening the lead discovery phase in
the pharmaceutical and crop protection in-
dustries. We have shown three examples in
this area which differ significantly in the
synthetic strategies which were employed
as well as in the chemical complexity of the
In general, 100 to 200 compounds per
library were synthesized and purified
in average amounts of 25 mg per
compound in average purities of 90%
(analysed by HPLC).
Multigeneration and sequential multi-
component reactions are well suited
for parallel synthesis as they take ad-
–