Skip to main content

Full text of "USPTO Patents Application 10555896"

See other formats


wo 2004/099774 PCT/EP2004/004883 

1 

PROCESS FOR THE IDENTIFICATION OF NEW MEDICAL TARGETS 

The present Invention relates to a process for the isolation and identification of 
pharmaceutlcally relevant target compounds (TO) from a sample, wherein said 
target compound(s) bind(s) to a compound of Interest (COI) under physiological 
conditions, said compound of Interest (COI) being associated with a given im-: 
paired condition or disease. Furthennore, the present Invention relates to a proc- 
ess for the identification of a pharmaceutlcally effective compound useful for pre- 
venting and/or treating a given impaired condition or disease, wherein said com- 
pound is identified by its capacity to bind to a relevant target compound (TO) that 
has been identified and isolated according to the invention; 

Most cellular processes are carried out by multiprotein complexes. The 
identification and analysis of their components provides insight into how the 
ensemble of expressed proteins (proteome) is organized into functional units. It 
is the challenge of postgenomic biology to understand how genetic information 
results in the concerted action of gene products In time and space to generate 
function. Dissecting the genetic and biochemical circuitry of a cell is a 
fundamental problem in biology. At the biochemical level, proteins rarely act 
alone; rather they Interact with other proteins to perfoma particular cellular tasks. 
Present knowledge regarding the identity of the building elements of specific 
complexes is limited and Is based on selected biochemical approaches and 
genetic analysis. Previous comprehensive protein-interaction studies are based 
on ex vivo and in vitro systems, such as, for example, the 2-hybrld systems (see 
e.g. Uetz et al.. Nature 10: 623-7 (2000), reviewed In Uetz et al., Cun-. Opin. 
Chem. Biol. 6: 57-62 (2002), patent application WO 00/60066) and need to be 
integrated with more physiological approaches. 

Seraphin, B. and RIgaut, G. (BP 1 105 508 B1) provide a new approach for de- 
tecting and/or purifying biomolecules and/or protein complexes. Their method for 
purifying biomolecules and/or protein complexes comprises three steps: 

(a) providing an expression environment containing one or more heterologous 
nucleic acids encoding one or more sub-units of a blomolecule complex, 



wo 2004/099774 PCT/EP2004/004883 

2 

the sub-units being fused to at least two different affinity tags, one of wliicli 
consists of one or more IgG binding domains of Stapliyiococcus protein A, 

(b) maintaining fine expression environment under conditions tliat facilitate 
expression of tlie one or more subunits In a native fonnn as fusion proteins 
with the affinity tags, and under conditions that allow the formation of a 
complex between the one or more subunits and possibly other compo- 
nents capable of complexing with the one or more subunits, 

(c) detecting and/or purifying the one or more subunits by a combination of at 
least two different affinity purification steps each comprising binding the 
one or more subunits via one affinity tag to a support material capable of 
selectively binding one of the affinity tags and separating the one or more 
subunits from the support material after substances not bound to the sup- 
port material have been removed. 

This method is called TAP purification (Tandem Affinity Purification). 

Gavin et al. (Functional organization of the yeast proteome by systematic analy- 
sis of protein complexes. Nature, vol. 415. January 10, 2002. p.141-147) suc- 
cessfully employed this TAP technology for purifying multiprotein complexes on a 
large scale to systematically analyze protein complexes in Saccharomyces cere- 
visiae. Specifically, they Inserted gene-specific cassettes containing a TAP tag. 
generated by polymerase chain reaction (PGR), which were inserted by homolo- 
gous recombination at the 3'end of the genes. Altogether, they processed 1 ,739 
genes. After growing the yeast cells to mid-log phase, assemblies were purified 
from total cellular lysate by TAP technology. They combined a first high-affinity 
purification, mild elutlon using a site specific protease, and a second affinity puri- 
fication to obtain protein complexes with high efficiency and specificity. The puri- 
fied protein assemblies were separated by denaturing gel electrophoresis, indi- 
vidual bands being digested by trypsin, analyzed by matrix-assisted laser desorp- 
tion/ionizatlon-tlme-of-flight mass spectrometry (MALDI-TOF MS) or Elel<tro- 
spray-mass spectrometry and identified by database search algorithms. 



wo 2004/099774 PCT/EP2004/004883 

3 

Starting with several distinct tagged proteins as entry points to purify a given 
complex, core components can be identified and validated, wtiereas more dy- 
namic, perhaps regulatory components, may be present differentially. 

Thus, the TAP technology allowed to assign cellular functions to new. non- 
annotated gene products, and to understand the context In which proteins oper- 
ate In yeast. TAP technology allows purification of very large complexes. The 
success of the TAP/MS approach for the characterization of protein complexes 
lies in the conditions used for the assembly and retrieval of the complexes. They 
include maintaining protein concentration, localization and post-translatlonal 
modifications in a manner that closely approximates nomnal physi9logy. 

Generally, the eariy phase of screening methods for identifying medically useful 
compounds involves some method wherein a screened compound is character- 
ized with respect to its direct binding interaction with a target compound such as 
a protein, said protein being associated with a given impaired condition or dis- 
ease. While pharmaceutical companies often have large compound pools in the 
range of several million individual compounds, there is a growing need for rele- 
vant targets for testing these libraries. Today, most often, one specific target 
compound is known to be associated with one or more specific diseases. How- 
ever, in most cases it remains unknown if such a given target directly influences a 
disease or whether it acts Indirectly as part of a much larger protein biocomplex, 
wherein multiple components act together through complex interactions such as 
cooperative binding, bridging factors, post translational modifications, allosteric 
structural changes, binding of ions and metabolites to influence the disease proc- 
ess. Because of the complicated and complex interaction of the components in- 
volved, it is highly probable that some or even most of the components of a multi- 
compound biocomplex that is involved in a disease process are potential targets 
for medical drug screening. The isolation and identification of medically important 
protein complexes will provide new insight into the molecular basis of many dis- 
eases and Identify new targets for the therapy and prophylaxis of diseases. 



The identification of further binding partners of a given compound of interest tha 
is associated with a given impaired condition or disease (Indication) will either 
point to (i) a further medical use of the compound of interest when the identified 



wo 2004/099774 PCT/EP2004/004883 

4 

binding partner (by direct and/or indirect binding) tias a l<nown medical use, (ii) 
the identification of new target compounds for drug screening, (iii) potentiai side 
effects of the compound of interest when the identified binding partner is l^nown 
to elicit side effects, or even (iv) the identification of diagnostic agents, when the 
binding partner is found to be suitable for specific and stable binding of the com- 
pound of interest or when the binding partner itself is found indicative of a specific 
disease or condition. 

It is the objective of the present invention to provide a process for isolating and 
Identifying compounds that bind to a compound of interest for the purpose of (i) 
identifying a further medical use of the compound of interest when the Identified 
binding partner (by direct and/or indirect binding) has a l^novwi medical use, (ii) 
for the identification of new target compounds for drug screening, (iii) for the Iden- 
tification of potential side effects of the compound of interest when the identified 
binding partner is known to elicit side effects, or even (iv) for the identification of 
diagnostic agents, when the binding partner is found to be suitable for specific 
and stable binding of the compound of interest or when the binding partner Itself 
Is found Indicative of a specific disease or condition. 

This problem is solved by providing a process for the isolation and identification 
of one or more pharmaceutically relevant target compounds (TC) from a sample 
that directly and/or indirectly bind(s) to a compound of Interest (COI), said com- , 
pound of interest (COI) being associated with a given impaired condition or dis- 
ease, comprising the following steps: 

a) providing said compound of interest (COI), preferably being bound to a 
suitable solid support material, 

b) adding said sample to the compound of Interest (COI) firom step a), 
preferably under physiological conditions, resulting in the direct or indi- 
rect binding to one or more of the components from said sample (CS) 
to the compound of interest (COI-CS complex formation), 



c) isolating and purifying said complex (COI-CS) from step b) and/or its 
components. 



wo 2004/099774 



5 



PCT/EP2004/004883 



d) identifying the component(s) of said complex (COI-CS), 

e) identifying at least one target compound (TC) of said complex (COI- 
CS) that is hitherto unknown to directly or indirectly bind to the com- 
pound of interest (COI), and optionally 

f) further purifying said target compound. • - - 

The basic concept underlying the above process is that a compound (COI) that is 
associated with a given impaired condition or disease is used as a "bair for its 
physiological counterpart(s), the target compound(s). Said target compound binds 
to the bait, and the target compound, preferably the target compound as well as 
all those compounds with an affinity to said target compound (which are also tar- 
get compounds) are Isolated and purified. It is the affinity of the target compound 
and/or of the COI to physiologically related compounds that allows for purifying 
and Identifying disease related complexes. For example, according to the present 
invention one can provide a receptor protein or a small organic drug molecule 
(COI) that is Involved in some impaired condition or disease in a first step, and 
then add said COI as a bait to a sample, e.g. some mammalian cell lysate, and 
isolate and purify any complexes under physiological conditions that result from 
COI-target compound binding. It is critical that if essentially all physiological direqt . 
or Indirect binding partners of a COI shall be detected that isolation and purifica-\ 
tion conditions, in particular also the incubation step b) are physiological or at 
least in the close proximity to the physiological conditions found in the original 
sample that harbors the target compounds. Under these physiological conditions, 
not only the COI-target complex is isolated and purified but also all other compo- 
nents that demonstrate affinity to the components of said complex. These co- 
Isolated and purified compounds are potential new compounds (targets) for medi- 
cal screening assays. 

The term "isolation and/or identification" as used herein refers to the isolation 
and/or identification of a complex comprising at least one compound of interest 
(COI) being bound to one or more target compounds (TC) and optionally those 
compounds that demonstrate affinity for said complex. Isolation In ttiis context 



wo 2004/099774 PCT/EP2004/004883 

6 

does not necessarily mean connplete purification but merely to a degree of isola- 
tion that allows for the identification of at least one of the target connpounds asso- 
ciated \with said complex. 

The term "target compound" (TC) refers to a compound that demonstrates binding 
affinity to a given compound of interest (COI) directly or indirectly by binding to 
another target compound that binds directly to a given compound of interest. Tar- 
get compounds can be any biomolecules such as a protein, peptide, nucleic acid, 
lipid, small biomolecule or any other molecule present in a living organism. The 
identified "target compound" may also be used as a "compound of interest" for 

i i 

further studies. ! ( 

The "target compound" preferentially binds to an active agent of a pharmaceutical 
composition in vivo. 



The term "complex" as used herein refers to a complex of at least one bio- 
molecule (target compound) (TC) with a compound of interest (COI). ^ 



■J 



The term "compound of interest (COI)" as used herein refers to any type of com- 
pound that can be linked to a given impaired condition or disease. Such a COI 
may be any type of biomolecule. preferably a protein, a peptide, a lipid, a carbo- 
hydrate, or a nucleic acid; or any type of a synthetic compound, such as the active 
agent of a pharmaceutical preparation, preferably a protein, a peptide, a lipid, a • 
nucleic acid, or a synthetic organic drug, more preferably a small molecule or- 
ganic drug. 

Preferably, compounds of interest (Cpl) are selected from the ROTE LISTE 2003. 
Arzneimittelverzeichnis fQr Deutschland. Rote Liste Service GmbH. Frank- 
furt/Main. 

More preferably, said compounds of interest (COI) are associated with diseases 
selected from cancer; neurodegenerative diseases, preferably Alzheimer's dis- 
ease or Parkinson's disease; inflammatory diseases, preferably allergies or 
rheumatoid arthritis; AIDS; metabolic diseases, preferably diabetes mellitus; 
«.H,rr,o. arthPrinsrJRmsis: coronarv and heart diseases; and infectious diseases. 



wo 2004/099774 



7 



PCT/EP2004/004883 



Most preferred COI's for practicing the present invention are selected from ttie 
group consisting of benserazide, sulindac, parthenolide, TNFalplia. TInese are 
presently associated with Parl<inson's disease (benserazide in combination with 
Levodopa) and inflammation (the latter three compounds). 

Through the identification of new physiological complexes that interact with a 
given COI, it is now possible to identify components of said complex which can be 
useful as drug targets and to provide new insight in a disease-related mechanism. 

.; / 

I 

The process of the invention allows to relate known COI to new target com- 
pounds, thereby relating these target compounds to the disease or impaired con- 
dition that is Icnown to be associated with the COI. These new target compounds 
are new screening tools for identifying active agents for treating said disease or 
impaired condition 

Moreover, the process of the present invention allows for identifying new uses of 
given COIs. For example, when a COI fomds a complex with one or more target 
compounds and at least one of these compounds is already known to be associ- 
ated with a disease or impaired condition that has not yet been associated with 
the COI; then this is a dear indication that the COI might have potential new dmg 
use. 

In a first step, the present invention provides a compound of interest. The com- 
pounds of interest may e.g. be selected from any known drug or from any known 
drug targets or known biologically active product, such as a protein. Preferably, 
said compound of interest is bound to a solid support material that is suitable to 
assist the isolation and purification later on during the process of the present in- 
vention. In a further prefen-ed embodiment the COI has a reactive moiety that may 
later be used to attach the COI-bound complex to a solid support or to a further 
reactive component that assists purification and isolation, e.g. an immunoreactive 
moiety and an antibody as reactive component leading to Immunoprecipitation. 



wo 2004/099774 PCT/EP2004/004883 

8 

In a further step of the present invention, a sample, preferably being derived from 
a living organism, Is added to the compound of interest under physiological condi- 
tions. 

"Physiological conditions" for COI-CS (Compound of interest - compounds from 
said sample / target compounds) formation are essential for practicing the present 
invention. "Physiological conditions" are inter alia those conditions which were 
present in the original, unprocessed sample material. Physiological conditions 
include the physiological protein concentration, pH, salt concentration, buffer ca- 
pacity and posttranslational modifications of the proteins mvolved. The term 
"physiological conditions" does not require conditions identical ,to those in the 
original living organism wherefrom the sample is derived but essentially cell-like 
conditions or conditions close to cellular conditions. A person skilled in the art will 
of course realize that certain constraints may arise due to the experimental set up 
which will eventually lead to less cell-like conditions. For example, it will be nec- 
essary to destroy cell walls when taking and processing a sample from a living 
organism to make Its components available for COI-binding and complexing. Suit- 
able variations of physiological conditions for practicing the processes of the in- 
vention will be apparent to those skilled in the art and are encompassed by the 
term "physiological conditions' as used herein. Preferably, the sample is proc- 
essed as a cell iysate or homogenized under mild conditions. In summary. It is to 
be understood that the term "physiological conditions" relates to conditions close 
to cell conditions but does not necessarily require that tinese conditions are Identi- 
cal. (Please, see also Rigaut et al., Nat. Blotechnol. 17, 1030-3 (1999); Puig et 
al., Metinods 24, 218-229 (2001) ; EP-A-1 105508.) 

When said sample is added to the COI under physiological conditions, the COI 
will bind to target compounds in said sample, thus resulting in the direct or indirect 
binding of one or more of the components from the sample (CS) to the compound 
of interest (COI-CS complex formation). This complex comprises at least the 
complex of Interest and one potential target compound but may also comprise a 
multitude of other compounds ttiat demonstrate affinity to the directiy bound target 
compound or the COI. Upon complex formation said complex and/or Its associ- 
ated components are isolated and purified from components which are not asso- 
ciated v^nth the complex. This is done under mild and physiological conditions to 



wo 2004/099774 PCT/EP2004/004883 

9 

leave the complex intact. In a further step, the isolated and purified components 
of the complex are identified. In a last step, components of said complex that 
were hitherto unknown to directly or Indirectly Interact with a compound of interest 
are Identified, and optionally further purified. 

Those newly identified target compounds are highly valuable for dmg screening 
for at least two reasons. First, a target compound identified by the process of the 
invention has strong potential for being used in screening assays for identifying 
active agents useful for treating the impaired conditions or diseases that are as- 
sociated with the original compound of interest used for identifying said target 
compounds. Secondly, if said target compound is identified as a compound that 
is associated with a disease or an impaired condition that has not hitherto been 
associated with the compound of interest, then the compound of interest may be 
of potential dmg use for treating the newly associated medical indication. Thus, 
new medical applications of known active agents can be identified. 

In a further aspect the present invention relates to a process wherein said target 
compounds that are identified according to the process of the present invention 
are further employed for screening assays for identifying new drugs for treating or 
preventing impaired conditions or diseases associated with the COI that was used 
for identifying said target compound. 

In a prefen-ed embodiment the present invention relates to a process for the iden- 
tification of a pharmaceutically effective compound useful for preventing and/or 
treating a given impaired condition or disease comprising the steps of 

(i) selecting one or more phamnaceutically relevant target compounds (TC) 
from a sample that directly or indirectiy bind(s) to a compound of interest 
(COI). said compound of interest (COI) being associated with a given im- 
paired condition or disease, comprising the following steps: 

a) providing said compound of interest (COI). preferably being bound to 
a suitable solid support material, 

b) adding said sample to the compound of interest (COI) from step a), 
preferably under physiological conditions, resulting in the direct or indi- 



wo 2004/099774 PCT/EP2004/004883 

10 

rect binding of one or more of tlie components from said sample (CS) 
to the compound of interest (COI-CS complex fonmation), 

c) isolating and purifying said complex (COI-CS) from step b) and/or its 
components, 

d) identifying tlie component(s) of said complex (COI-CS), 

e) identifying at least one target compound (TC) of said complex (COI- 
CS) that is hitherto unknown to directly or Indirectly bind with the 
compound of interest (COi); 

(ii) employing one or more of the target component(s) (TC) identified in 
step (i) e) in a screening assay for the Identification of pharmaceutical ly 
effective compounds. 

The term "employing a target compound" for the identification of phannaceutically 
effective compounds relates to the use of said target compound in a pharmaceuti- 
cal screening assay. The skilled person is well aware of how to set up a pharma- 
ceutical screening assay based on the molecular and physiological characteristics 
of a target compound. Pharmacological validation of potential drug" compounds is 
typically perfonmed by an in vitro binding assay. For example, target compound 
binding to potential drugs can be measured by competition assays, wherein 
known binding agents of a given protein compete for protein binding with a poten- 
tial drug (Competitive binding assay). Suri'ace plasmon resonance can be meas- 
ured to validate TC-drug binding. Drugs or target compounds can be labeled to 
identify drug target complexes. In v/tro and in vivo activity assays are also useful 
to validate drugs. For example, if a protein has an enzymatic activity, then the 
reduction of starting material or the increase of products can be measured or the 
reduction or increase of cofactors such as NADH/NAD, ATP/ADP, etc. if tine po- 
tential drug activates or deactivates tine target compound's biological function 
ttien a cellular functional assay will provide for establishing target compound-dmg 
binding. 

In the art, a wide range of techniques are known for establishing assays and 
screening compound libraries in order to identify potential drugs (Seethala and 
Femandes (eds). Handbook of Drug Screening, Marcel Dekker, New Yori<, 2001). 
Such assays can generally be adapted for rapid screening of large libraries of 



wo 2004/099774 PCT/EP2004/004883 

11 

compounds that were generated by combinatorial chemistry or focused libraries 
of synthetic or natural compounds. 

The most widely used assay techniques are based on radioactivity detection (e.g. 
scintillation proximity assay) or fluorescent detection technologies (e.g. fluores- 
cence intensity, fluorescence polarization, fluorescence resonance energy trans- 
fer or fluorescence correlation spectroscopy). 

in the following, preferred pharmaceutical screening assays will be described that 
can employ one or more of the target component(s) (TC) identified in step 0) e) 
for the identification of pharmaceutically effective compounds. Each of the pre- 
fen-ed assays described below is amendable to high-throughput analysis which 
facilitates the screening of large numbers of conipounds (potential drugs). 

In wHrn binding assav ^protein-p rntein interaction) 

A competitive in vitro binding assay can be used to identify modulators of protein- 
protein or protein-peptide interactions. These modulators can dismpt the interac- 
tion (inhibitors) or stabilize the interaction. 

Briefly, a binding assay Is perfomned in which a purified protein (e.g. cyclin- 
dependent kinase 2/cyclin E complex) is used to bind a fluorescently labeled pep- 
tide. This labeled peptide is contacted with the purified protein in a suitable buffer 
solution that pennits specific binding of the two components to fomi a protein- 
peptide complex in the absence of an added chemical compound. Particular 
buffer conditions can be selected depending of the target protein of Interest as 
long as specific protein-peptide binding occurs in the control reaction. The pro- 
tein-peptide complex has slow rotational mobility compared to the free peptide 
which results in a high fluorescence polarization signal. A parallel binding assay 
is perfomied in which a chemical compound (test agent) is added to the reaction 
mixture. If the chemical compound displaces the labeled peptide, the non-bound 
labeled peptide has a higher rotational mobility than the protein-peptide complex 
resulting in a lower fluorescence polarization signal. On the other hand, if the 
chemical compound stabilizes the protein -peptide interaction, an Increase in the 
polarization signal is observed. The amount of labeled peptide bound to the tar- 



wo 2004/099774 PCT/EP2004/004883 

12 

get protein is determined for the reactions in the absence and presence of the 
chemical compound (test agent). If the amount of bound labeled peptide in the 
presence of the chemical compound is different than the amount of bound, la- 
beled peptide in the absence of the chemical compound, the compound is a 
modulator of the interaction between the protein target and the peptide (Pin et al., 
1999, Analytical Biochemistry 275. 156-161). 



In vitro binding assay (nuclear reo entor - hormnne interaction) 

A competitive in vitro binding assay can be used to Identify modulators of nuclear 
receptors (e.g.. the steroid homnone receptor superfamily). These modulators can 
stimulate receptor functions (agonists) or blocl< receptor functions (antagonists). 
A competitive ligand binding assay does not allow to differentiate between the 
two modes of action. 

Briefly, a binding reaction is performed in which a purified human nuclear recep- 
tor (e.g., glucocorticoid receptor) is used to bind to a fluorescentiy labeled hor- 
mone ligand (e.g.. fluoresceine-dexamethasone). Alternatively, a crude cellular 
extract containing the receptor can be used in the assay. This labeled ligand is 
contacted with the purified protein in a suitable buffer solution that pemnits spe- 
cific binding of the two components to fonn a receptor - ligand complex in the 
absence of an added chemical compound. Particular buffer conditions can be ( 
selected depending of the target protein of interest as long as specific receptor - 
ligand binding occurs in the control reaction. 

The protein - ligand complex has slow rotational mobility compared to the free 
ligand which results In a high fluorescence polarization signal. A parallel binding 
assay is performed In which a chemical compound (test agent) is added to the 
reaction mixture. If the chemical compound displaces the labeled ligand, the non- 
bound labeled ligand has a higher rotational mobility than the receptor - ligand 
complex resulting in a lower fluorescence polarization signal. 



The amount of labeled ligand bound to the target protein is detennined for the 
reactions in the absence and presence of the chemical compound (test agent). 



wo 2004/099774 PCT/EP2004/004883 

13 

This assay can be used to identify molecules that bind to the receptor but does 
not allow to distinguish between agonists and antagonists. For further characteri- 
zation of identified binders a coactivator recruitment assay or a functional cellular 
assay can be used (see below). (Lin et al., 2002, Anal. Biochem. 300. 15-21; 
Parker et al.. 2000, J. Biomol. Screen. 5, 77-88) 

In vitro binding assay (nuclear receptor - coactivator recruitment) 

Ligand-dependent protein - protein interactions between nuclear receptors and 
nuclear receptor cocativators are important for the biological function of nuclear 
receptors. An in vitro binding assay based on fluorescence resonance energy 
transfer can be used to detect and quantify such interactions. This assay format 
can be used to identify agonists, partial agonist and antagonists. 

Briefly, a binding reaction is performed in which a fiuorescently labeled nuclear 
receptor (e.g. estrogen receptor alpha) is used to bind to a fiuorescently labeled 
coactivator or coactivator fragment (e.g. steroid receptor coactivator 1 , SRC-1 ) in 
the presence of a hormone agonist. Close proximity of the nuclear receptor and 
coactivator allows transmission of a FRET signal. Compounds disrupting the re- 
ceptor coactivator complex result in a lower FRET signal (antagonists). 

Alternatively, a labeled nuclear receptor and labeled coactivator can be incubated 
in the absence of a hormone agonist resulting in a low FRET signal. Compounds 
stimulating the association of receptpr and coactivator yield an increased FRET 
signal (agonists). (Zhou et al.. Methods 25, 54-61; Zhou et al.. 1998, Mol. Endo- 
crinol. 12, 1594-1604) 

i ' / 

'. I 

In vitro enzvme activity assav foroteln kinase) 

An in vitm protein tyrosine kinase immunoassay can be used to identify inhibitors 
of kinase activity. 

Briefly, a fluorescein-labeled peptide substrate is incubated with the kinase (e.g. 
Lck), ATP and an antlphosphotyrosine antibody. As the reaction proceeds, the 
phosphorylated peptide binds to the antlphosphotyrosine antibody, resulting in an 



wo 2004/099774 PCT/EP2004/004883 

14 

increase in the polarization signal. Compounds that inhibit the kinase result in a 
low polarization signal. 

Alternatively, the assay can be configured in a modified indirect fomnat. A fluores- 
cent phosphopeptide is used as a tracer for complex fomnatlon with the antlphos- 
pho-tyrosine antibody yielding a high polarization signal. When unlabeled sub- 
strate is phosphorylated by the kinase, the product competes with the fluorescent 
phosphorylated peptide for the antibody. The fluorescent peptide is then released 
from the antibody into the solution resulting in a loss of polarization signal. 
Both the direct and indirect assays can be used to identify inh|bitors of protein 
tyrosine kinase activity. (Seethala, 2000, Methods 22, 61-70; Seethala and Men- 
zel, 1997, Anal. Biochem. 253, 210-218; Seethala and Menzel, 1998, Anal. Bio- 
chem. 255. 257-262) 

This fluorescence polarization assay can be adapted for the use with protein ser- 
ine/threonine kinases by replacing the antiphophotyrosine antibody vwth an an- 
tiphosphoserine or antiphosphothreonine antibody. (Turek et al.. 2001, Anal. Bio- 
chem. 299. 45-53. PMID 11726183; Wu et al., 2000. J. Bibmol. Screen. 5. 23-30, 
PMID 10841597). 

In vitro activity assay forotei n phosphatase) 

An in vitro protein tyrosine phosphatase immunoassay can be used to identify I 
Inhibitors of phosphatase activity. 

Briefly, a fluorescein-labeled phosphopeptide substrate Is incubated with the 
phosphatase (e.g., T cell PTP) and an antiphosphotyrosine antibody. As the reac- 
tion proceeds, more dephosphorylated peptide is produced which can not bind to 
the antiphosphotyrosine antibody any more, resulting in a decrease in the polari- 
zation signal. For compounds that inhibit the phosphatase ttie polarization signal 
remains high. 

This fluorescence polarization assay can be adapted for the use with protein ser- 
ine/threonine phosphatases by replacing the antiphophotyrosine antibody vwtii an 



wo 2004/099774 PCT/EP2004/004883 

15 

antiphosphoserine or antiphosphothreonine antibody. (Parker et al., 2000, J. 
Biomol. Screen. 5. 77-88) 



In vitro receptor binding assay (GPCR - liaand interaction) 

An in vitro competitive binding assay can be used to Identify modulators (agonists 
or antagonists) of G protein coupled receptors (GPCRs). Briefly, either intact cells 
or receptor-containing membrane fragments (e.g., vesicles bearing the CXCR1 
receptor) and a fluorescently labeled ligand (e.g., interieukin-B) are incubated 
such that specific binding occurs. Addition of test compounds can lead to dis- 
placement of the labeled ligand resulting in a change of the fluorescence signal 
as measured by fluorescence polarization or fluorescence con-elation spectros- 
copy). (Klumpp et al.. 2001. J. Biomol. Screen. 6. 159-170; Banks et al., 2000. J. 
Biomol. Screen. 5, 159-168) 

Such a binding assay can not differentiate between agonists and antagonists, but 
identified binders can be further characterized by functional assays that measure 
production of a second messenger (e.g. cAMP). (Kariv et al.. 1999, J. Biomol. 
Screen. 4. 27-32) 

nelluiar functional assay Huciferase reporter aene system) 

_— — - J 

I 

A cellular assay can be established to identify modulators of signal transduction. 
Briefly, a luciferase reporter construct driven by a suitable promoter element 
(e.g.. NFkB reporter) is transfected into a cell line and the luminescence signal is 
measured In the presence or absence of a cytokines (e.g. interieukin-1 beta). Af- 
ter addition of test agents (chemical compounds) a change of the luminescence 
signal can be recorded Indicating stimulation or Inhibition of reporter gene ex- 
pression. (Davis et al.. J. Biomol. Screen. 7. 67-77; Maffia et al., 1999, J. Biomol. 
Screen. 4, 137-142) 



DMA binding assay 



wo 2004/099774 PCT/EP2004/004883 

16 

An exemplary DNA binding assay can be carried out by contacting a complex 
having DNA binding activity with a radioactive [^^P] end-labeled DNA substrate 
under appropriate conditions and detecting bound protein. The detection of DNA 
bound protein can be carried out, e.g., by filtrating the solution through a nitrocel- 
lulose filter and determining the radioactivity bound to the filter. This assay Is 
based on the retention of nucleic add-protein complexes on nitrocellulose 
vk^ereas free nucleic acid can pass through the filter, (see e.g. Nowock, J. et al.. 
1982, Methods 30: 607-15) 

GTPase assay 

; \ 

An exemplary GTPase assay can be carried out by loading a complex having 
GTPase activity with a radioactivity [gamma^^PJ-labeled GTP substrate under ap- 
propriate conditions and detecting the amount of radioactivity bound to the 
GTPase protein and the release of free radioactive phosphate. The detection of 
the remaining GTP substrate bound to the GTPase protein can be carried out, 
e.g.. by filtrating the solution through a nitrocellulose filter and determining the 
radioactivity bound to the G-protein. (see e.g. Ridley, A. J. et al., 1993, Methods 
12: 5151-60) 

Protease assay 

An exemplary protease assay can be carried out by contacting a complex having 
protease activity with a double labeled peptide substrate with fluorine (e.g. 
EDANS) and quencher chromophores (e.g. DABCYL) under appropriate condi- 
tions and detecting the increase of the fluorescence after cleavage. 

The substrate contains a fluorescent donor near one end of the peptide and an 
acceptor group near the other end. The fluorescence of this type of substrate is 
Initially quenched through intramolecular fluorescence resonance energy transfer 
(FRET) between the donor and acceptor. When frie protease cleaves the sub- 
strates the products are released from quenching and the fluorescence of the do- 
nor becomes apparent. The increase of the fluorescence signal is directly propor- 
tional to the amount of substrate hydrolyzed. (see e.g. Taliani, M. et al, 1996, 
Methods 240: 60-7) 



wo 2004/099774 



17 



PCT/EP2004/004883 



ApoDtosis assay 

An exemplary apoptosis assay can be carried out by contacting a complex having 
apoptosis activity using fluorescent DNA-staining dyes. e.g. propidlum iodide, to 
reveal nuclear morphology substrates under appropriate conditions and detecting 
the amount of apoptotic cells by confocal or transmission electron microscopy. 
The detection of apoptotic cells can be earned out by distinguishing viable from 
apoptotic cells based on morphological alterations typical of adherent cells under- 
going apoptosis becoming rounded, condensed, and detached from the dish, (see 
e.g. Tewari, M. and Dixit, V.IVl.. 1995, J. Biol. Chem., 17 3255-60X 

The samples used in the process of the present invention that comprise the 
potential target compounds are preferably derived from a mammal, preferably 
• from a human, more preferably from a human suffering from said impaired 
condition or disease. 

The term "derive" indicates that the sample is isolated from a mammal and further 
processed to accommodate the technical constraints of the process of the inven- 
tion. Samples from healthy mammalian individuals will provide for target com- 
pounds at regular expression levels and form complexes with further components 
of the sample under regular conditions. If samples are taken from humans with 
an impaired condition or disease, then the compound of interest may be associ-- 
ated with different components and target compounds may be present in different 
concentrations reflecting the cellular conditions of said mammal. 

In a preferred embodiment, the present invention relates to a process, wherein 
said impaired condition or disease and/or said sample is associated with an im- 
paired condition or disease which is selected from cancer; neurodegenerative 
diseases, preferably Alzheimer's disease or Parkinson's disease; inflammatory 
diseases, preferably allergies or rheumatoid arthritis; AIDS; metabolic diseases, 
preferably diabetes mellitus; asthma; artheriosclerosis; coronary and heart dis- 
eases; and infectious diseases. 



wo 2004/099774 PCT/EP2004/004883 

18 

For the formation of the COI-CS complex, It is essential that the complex be 
formed under physiological conditions. Said physiological conditions include the 
cellular content of the sample, the protein concentration, the pH, the buffer capac- 
ity, osmolarity, temperature of the cells from which the sample is derived. As 
mentioned before, physiological conditions according to the present invention do 
not need to be identical to the conditions in complete cells in their natural envi- 
ronment but are merely required to resemble those conditions to an extent that 
; allows for complex fomnation. Preferably, said physiological conditions for forming 
a complex of the present invention consider a physiological pH, buffer, and protein 
content. . _ „ 

: i 

Once the complex is formed under physiological conditions care must be exer- 
cised not to disrupt said complex when isolating and purifying the complex or its 
components. 

One preferred method of practicing the invention involves affinity labeling of the 
target compound or the COI prior to step (b) of the process of the present inven- 
tion. For example, for labeling the TO, cells of the sample, being present, e.g. as 
whole cells, lysates or extracts, are labeled, e.g. by Incubation, with an affinity 
marker, e.g. a cell permeable affinity marker, e.g. biotinylated parthenolide or 
biotinylated cell-permeable caspase inhibitor), and then in step (b) said labeled 
sample is added to the COI for complex formation under physiological conditions. 
Also, the COI can be labeled by conventional techniques. After optional disruption 
of the cell (when whole cells were used) the complexes are isolated and purified 
using solid support material to which the affinity marker has an affinity. 

In a prefen-ed embodiment of the present invention, the COI or target compound 
is bound to a suitable solid support material. This support binding will assist isola- 
tion and purification after complex formation. 

Preferred solid support materials are Sepharose, such as Sepharose 4B, or aga- 
rose or Latex or Cellulose. The matrixes can be coupled by active groups such as 
NHS, Carbodiimide etc. 



wo 2004/099774 PCT/EP2004/004883 

19 

In another prefen-ed alternative the processed sample, e.g, lysate, extract, is 
added to COi's that are bound to solid support. COI's can be coupled to solid 
support by direct coupling, e.g. amino-, sulfhydryl-, carboxyN, hydroxyl-, 
aldehyde-, and ketone groups and by Indirect coupling, e.g. via blotin, blotin being 
covalently attached to COI's and non-covalent binding of blotin to streptavldin 
which is bound to solid support directly. Linkage to solid support material may 
involve cleavable and non-cleavable linkers. Isolation and purification of com- 
plexes does not necessarily involve the removal of the COI from solid support ma- 
terial. Preferably, the COI-solid support linker is cleavable. More preferably, the 
linker comprises an enzyme cleavage site. Also preferred is that the linker com- 
prises a site for indirect coupling, more preferably via a hapten or fluorescent dye 
(e.g. fluorescein covalently bound to drugs such as fluorescein-Taxol, or an anti- 
fluoescein antibody bound to protein A beads.) Once the COI-TC complex Is 
formed under physiological conditions while the COI is bound to solid support, the 
isolation and purification of said complex and its components may proceed. 

Preferred binding interfaces for binding the compound of interest to solid support 
material are linkers with a C-atom backbone. Typically linkers have a backbone of / 
8, 9 or 10 C-atoms. The linkers contain either, depending on the compound to be 
coupled, a carboxy- or amino-acive group. 

Most preferably, the complexes obtained by a process according to the invention 
are isolated and purified at least in part by the TAP technology. 

A preferred process according to the invention that involves isolation and purifica- 
tion of the COI-CS complex and/or its components in step (c) at least in part by 
the TAP technology is a process wherein the compound of interest (COI) provided 
In step (a) is linked to a tandem affinity tag or one or more target compounds (TC) 
in the sample used in step (b) is linked to a tandem affinity tag. 

The term TAP-technology refers to a tandem affinity purification wherein one 
component of a complex comprising at least two components is provided with two 
affinity tags. The TAP technology is e.g. disclosed in EP 1 105508 B1 and is ex- 
emplified by Rlgaut et al., Nat. Blotechnol. 17, 1030-3 (1999) and Puig et al. in 
Methods 24, 218-229 (2001). The tandem affinity purification (TAP) method was 



wo 2004/099774 PCT/EP2004/004883 

20 

used e.g. in: A general procedure for protein complex purification methods 24, 
218-229 (2001), Gavin et al. Functional organization of the yeast protein by sys- 
tematic analysis of protein complexes, Nature, vol. 415, January 2002, 141-147, 
and Rigaut et al., A generic protein purification method for protein complex char- 
acterization and protein exploration, Nature, Biotechnology, vol, 17, October 1999 
1030-1032. While the TAP technology has up to now been used mostly for sam- 
ples, wherein the TAP tag has been added to cell proteins by recombinant meth- 
ods, the present Invention contemplates adding a TAP- tag to the compound of 
interest or target compounds by any suitable method, such as e.g, synthetic 
chemical modification or recombinant modification. 

According to the invention the TAP-tag may be linked to the COI or the target 
compounds. For example, TAP-tags may be linked to target compounds in a 
sample by recombinant techniques such and homologous recombination (see e.g. 
Gavin et al.) or be linked to COI's by direct or Indirect binding (synthetic meas- 
ures; or recombinant measures, If the COI is a peptide or nucleotide). When a 
TAP-tag Is Introduced to target compounds In a sample by recombinant meas- 
ures, it Is preferred to maintain expression of the fusion protein at, or close to, its 
natural level. Indeed, over-expression of the protein often induces its association 
with non-natural partners (heat shock proteins, proteazome). 

In a more prefen-ed embodiment, the TAP tag consists of 2 IgG binding domains 
of staphylococcus aureus protein A (Prot A) and a calmodulin binding peptide ' 
(CBP). preferably separated by a TEV protease cleavage site. If the COI or the 
TCs are peptides, such TAP tags can be positioned on the C as well as the N- 
terminal site of the compound of interest. When using the Prot A module, said 
module needs to be at the extreme N- or C-terminus of a fusion protein or other 
compound of interest. Preferably, both affinity tags are selected for highly efficient 
recovery of proteins present at low concentrations. Prot A binds tightly to an IgG 
matrix requiring the use of the TEV protease to allude material under native condi- 
tions. The eluate of this first affinity purification step is then Incubated with 
calmodulin coated beads in the presence of calcium. After washing, which re- 
moves contaminants and the TEV protease remaining after the first affinity selec- 
tion, the bound material is released under mild conditions with EGTA. Optimized 
conditions have been developed for the generic use of the TAP strategy. The 



wo 2004/099774 PCT/EP2004/004883 

21 

TAP- tag, however, is very tolerant to buffer conditions and changes to be innple- 
mented to optimize recovery of specific complexes. 

Once the COI-CS complex according to any process of the invention is formed, 
either In solution or attached to solid support, the complex and/or its components 
are Isolated and purified from sample components that are not associated with the 
complex. Appropriate methods, especially for isolating and purifying complexes 
bound to solid support material are available to the skilled person and comprise 
e.g. washing, centrifugation, specific affinity purification and elution steps, (e.g. 
see EP 1 105 508 B1, Rigaut et al., Puig et al.. Rigaut et al.) 

i 
J 

The Isolated and purified material can be analyzed In a number of ways. For a 
protein complex or component characterization, proteins are preferably concen- 
trated and fractionated, e.g. on a denaturing gel before identification, e.g. mass 
spectroscopy. Alternatively, Edman degradation or Western blotting may be em- 
ployed. Because the various purification steps are performed In a gentle native 
manner, purified complexes or their components may also be tested for their ac- 
tivities or be used in structural analysis. 

Preferred methods for identifying complex components, are specific antibody bind- 
ing, perferabiy immunoprecipitation, Edman degradation or related chemical 
analysis, Western blot, mass spectroscopy, more preferably matrix-assisted laser 
desorption/ionization-time-of-light mass spectrometry (MALDl-TOF MS). 

The results obtained from the identification techniques are then compared to iden- 
tify target compounds that have hitherto been uni^nown to directly or indirectly 
bind to the compound of interest. This identification can preferably be achieved 
by comparing the chemical structure and/or physical properties of said compo- 
nent(s) with the Information available in sequence databases and/or suitable sub- 
stance libraries. The person skilled In the art Is well aware of how to use modem 
bioinformatlcs for identifying known compounds or Identifying new compounds. 

As mentioned before, the target compounds Isolated and Identified according to 
the present invention are useful for screening assays. 



wo 2004/099774 PCT/EP2004/004883 

22 

Preferably, a screening assay according to the invention comprises 

a) contacting one or more target compound(s) (TC) selected in a process 
according to the invention for isolation and identification of target com- 
pounds, with a compound suspected to be pharmaceutlcally effective, 
and 

b) determining the presence of a chemical and/or physical binding among 
the compound(s) (TC) and the compound of step a(A). 

Compounds suspected to be pharmaceutlcally effective can be derived from natu- 
ral sources such as plants, herbs, and animal sources which have been demon- 
strated to influence mammalian physiology. Typically and preferably, said com- 
pounds are selected from a suitable compound library. Such compound libraries 
are commercially available from e.g. Chemical Diversity Inc., Maybrldge. Tripos, 
Evotec OAI. Most pharmaceutical companies involved In active research have 
suitable compound libraries in which millions of compounds are stored. 

As mentioned before, a process of the present invention is capable of isolating 
more than just the target compound that actively binds to the compound of inter- 
est. Moreover, a process according to the present invention is capable of Isolat- 
ing, purifying and identifying all those components having affinity under physio- 
logical conditions with the COI/TC complex. 

Therefore, in a prefen-ed embodiment the present invention is also directed to a 
process for screening medical compounds comprising the contacting of one. 
some, or all of the components of the identified COI-CS complex. 



Example 1: Identification of the protein complex associated with the drug 
benserazlde 



Identification of the interaction between benserazide and carbonyl 
tase. 



wo 2004/099774 PCT/EP2004/004883 

23 

Carbonyl reductase was surprisingly identified as a novel drug target in a drug 
pulldown assay with immobilized benserazide. 

Coupling of the compound of interest fCQI^ and wasliina of coupled beads: 

The compound benserazide was immobilized on NHS-activated beads (Affi Gel 
10, BioRad) via its NHa-group. 300 pi of the beads (both with the Immobilized 
benserazide and control beads) were washed in 10 ml washing buffer A (50 mM 
Tris. pH 7,5: 0,1 M NaCI, 0.15 % Igepal. 1,5 mM MgCb, 0,1 mM DTT) for 5-10 
min at 4 °C and centrifugation for 5 min at 1000 rpm in a Heraeus Varifuge 3 
OR). 

Incubation of beads with Ivsate: 

Mouse liver cell lysate (60-100 mg total protein) and 125 pi 50x protease inhibitor 
tablets (Roche. Complete. EDTA free) \A^ere added to the beads and the sus- 
pension was Incubated for 1 h at 4 "C (while rotating). The suspension was 
washed 1-3 x with 10 ml washing buffer B (50 mM Tris, pH 7,5; 0,1 M NaCI. 0.15 
% Igepal, 1,5 mM MgCb, 0,1 mM DTT, 1 x Protease Inhibitor tablet (Complete, 
EDTA free. Roche)) by rotating the suspension for 5-10 min at 4 "C and centrifu- 
gation at 10.000 rpm at 4 "C. The beads were transferred to a 1 ml MoBiTec col- 
umn and connected to a 10 ml syringe. 10 ml washing buffer B were added. 

Drug elution: 

Elution was peri'ormed by adding a 5-10 fold excess of the drug relative to the 
beads capacity. The drug was dissolved in 500 pi washing buffer B and incubated 
with beads on a rotating platform for 1 h at 4°C and subsequently eluted in an 
eppendorf tube. 300 pi of washing buffer B were added to the beads and eluted 
immediately in the same eppendorf tube. The beads were washed with 5 ml 
washing buffer B using a syringe. 



Acidic elution: 



wo 2004/099774 PCT/EP2004/004883 

24 

500 \il acidic buffer (0,1 M NaOAc, pH 4,0) were added to the beads, the beads 
were rotated at 4^*0 for 10-15 min. After elution in an eppendorf tube the beads 
were washed with 10 ml H2O using a syringe, 

BoilinQ of beads: 

300 pi of 2x sample buffer (NuPage LDS sample buffer) + 100 mM DTT were 
added to the beads. After boiling for 10-15 min at 95 °C the suspension was 
eluted in an eppendorf tube. 

The sample was run on a Coomassie gel and the proteins were identified by 
massspectrometry analysis as described below. 

Carbonyl reductase (CBR1) was identified as a binding partner of benserazide 
Determination of the inhibitory effect of benserazide on carbonyl reductase 

Determination of carbonvl-reductase activity: 

The carbonyl reductase activity was evaluated spectrophotometrically according 
to the methods of Iwata et al. 1990. Eur. J. Biochem 193, p. 75-81. : Inazu N, 
Ruepp B., Wirth H.. Wermuth B. 1992. BBA, 1116, p. 50-56 and Imamura et al. . 
1993. Arch. Biochem. Biophys. 300. p. 570-576. The oxidation rate of NADPH 
was recorded in the presence of the specific substrate menadione at 340 nm at 
room temperature on a Jenway 6505 UVA/IS Spectrophotometer. 

The standard assay mixture consisted of 100 mM sodium phosphate buffer pH 
7.0, 0.12 mM NADPH, 0.25 mM menadione. The reaction was started by adding 
5-20 pg of E. CO// expressed His-tagged human carbonyl reductase (CBR1) or 
altematively by adding 1 0 pi of mouse live lysate extract (total protein concentra- 
tion of 15 mg/ml). The total volume of the reacHon mixture was 1 ml. The change 
of the absorbance was monitored at 340 nm. 



Inhibition of carbonvl-reductase activitv with benserazide: 



wo 2004/099774 PCT/EP2004/004883 

25 

The Inhibition of carbonyl reductase was determined using the assay described in 
example 1. In addition to menadione as a substrate, the assay mixture was sup- 
plemented with 0, 0.5, 1, 2, 3, 6, or 7.5 mM benserazide. The inhibition experi- 
ment was performed with mouse liver lysate as well as with recombinant CBR1 in 
protein supplemented probes. The results for mouse liver lysate and recombinant 
CBR1. are presented In table 1 . These results demonstrate that benserazide has 
a profound inhibitory Impact on carbonyl reductase activity. 



Table 1 



benserazide (mM) 


initial rate-C^'/mm) 




carbonyl reductase 


mouse liver lysate 


0 


24.0 


16.3 


0.5 


n.d. 


12.1 


1 


n.d. 


8.6 


2 


13.7 


1.2 


3 


n.d. 


1.3 


6 


3.5 


n.d. 


7.5 


0.3 


n.d. 



Identification of proteins binding to carbonyl reductase 

Subsequently, human carbonyl reductase was TAP-tagged at the amino-temninus 
and expressed In a human neuronal cell line (SK-N-BE2 cells). The protein com- 
plex was purified according to TAP-technology procedures (see also O/0009716 / 
EP 1 105 508 81 and Rigaut, G et.al. (1999), Nature Biotechnology, vol. 17 (10): 
1030-1032). 

For expression of the TAP-tagged cattonyl reductase, the cell line was infected 
with a MoMLV-based recombinant virus construct. 

For the preparation of the vector, 293 gp cells were grown to 100% confluency. 
They were split 1:5 on poly-L-lysine plates (1:5 diluted Poly-L-Lysine [0.01% 
stock solution, Sigma P-4832] in PBS. left on plates for at least 10 min.). 



wo 2004/099774 PCT/EP2004/004883 

26 

On Day 2 63 pg retroviral vector DNA together v«th 13 pg of DNA of plasmid en- 
coding an appropriate envelope protein were transfected into 293 gp cells (So- 
mia. NV et al (1999) Proc. Natl. Acad. Sci. USA 96: 12667-12672; Somia, NV et 
al., (2000) J. Virol. 74: 4420-4424). 

On Day 3, the medium was replaced with 15 ml DMEM + 10% FBS per 15-cm 
dish. 

On Day 4, the medium containing the viruses (supernatant) was harvested (at 24 
h following medium change after transfection). When a second collection was 
performed, DMEM 10 % FBS was added to the plates and the plates were incu- 
bated for another 24 h. 

For collecting the supernatant was filtered through a 0.45 micrometer filter (Com- 
ing GmbH, cellulose acetate, 431155). 

The filter was placed into konical polyallomer centrifuge tubes (Beckman, 
358126) that were placed in buckets of a SW 28 rotor (Beckman). 

The filtered supernatant was ultracentrifuged at 19400 rpm in the SW 28 rotor for 
2 hours at 21 "C. The supernatant was discarded. The pellet containing the vi- 
ruses was resuspended in a small volume (for example 300 pi) of Hank's Bal- 
anced Salt Solution [Gibco BRL, 14025-092] by pipetting up and down 100-times 
using an aerosol-safe tip. These viruses were used for transfection as described 
below. 

Cells that were infected were plated one day before infection into one well of a 6- 
well plate. 4 hours before infection the old medium on the cells was replaced with 
fresh medium. Only a minimal volume was added, so that the cells were com- 
pletely covered (e.g. 700 pi). During infection tiie cells were actively dividing. 

To the concentrated virus, polybrene (hexadimethrine bromide; Sigma, H 9268) 
was added to achieve a final concentration of 8 pg/ml (this is equivalent to 2.4 pi 
of the 1 mg/ml polybrene stock per 300 pi of concentrated retrovirus). The virus 
was incubated in polybrene at room temperature for 1 hour. 



wo 2004/099774 



27 



PCT/EP2004/004883 



For infection, tlie virus/polybrene mixture was added to the cells and incubated at 
37 "^C at the appropriate CO2 concentration for several hours (e.g. over-day or 
over-night). 

Following infection, the medium on the infected cells was replaced with fresh me- 
dium. The cells were passaged as usual after they became confluent. The cells 
contained the retrovirus integrated Into their chromosomes and stably expressed 
the gene of interest. 

Purification or protein complexes: 

For purifying the protein complex associated with carbonyl reductase the follow- 
ing protocols were used. 

For the purification of cytoplasmic TAP-tagged proteins 5x10° adherent cells 
(corresponding to 40 15 cm plates) were used. The cells were harvested and 
washed 3 times in cold PBS (3 min, 1300 rpm, Heraeus centrifuge). -The cells 
v^^ere frozen in liquid nitrogen and stored at -80*'C, or the TAP purification was 
directly continued. 

The cells were lysed in 10 ml CZ lysis buffer (50 mM Tris, pH 7.5, 5 % Glycerol. 
0.2 % IGEPAL, 1.5 mM MgCl2. 100 mM NaCI, 25 mM NaF, 1 mM Na3V04, 1 mM 
DTT, containing 1 tablet of protease inhibitor cocktail (Roche) per 25 ml of buffer) 
by pipetting 2 times up and down, followed by a homogenizing step (10 strokes in 
a dounce homogenizer with tight pestle). The lysate was incubated for 30 min on 
ice. After spinning for 10 min at 20000 g the supematant was subjected to an 
ultracentrifugation step of 1 h at 100 000 g. The supematant was frozen in liquid 
nitrogen and stored at -SO'C, or the TAP purification was directly continued. 

The lysates were thawn quickly in a 37*^0 waterbath. 0.4 ml of unsettled rabbit 
IgG-Agarose beads (Sigma, washed 3 times in CZ lysis buffer) were added, and 
incubated for 2 h while rotating at 4C. Protein complexes bound to the beads 
were obtained by centrifugation (1 min, 1300 rpm, Heraeus centrifuge). The 
beads were transfen-ed into 0.8 ml Mobicol Ml 002 columns (Pierce) and washed 



wo 2004/099774 PCT/EP2004/004883 

28 

With 10 ml CZ lysis buffer (containing 1 tablet of Protease inhibitor cocktail 
(Roche) per 50 ml of buffer). After an additional washing step with 5 ml TEV 
cleavage buffer (10 mM Tris, pH 7.5, 100 mM NaCI, 0.1 % IGEPAL. 0.5 mM 
EDTA. 1 mM DTT) the protein-complexes were eluted from the beads by adding 
150 pi TEV cleavage buffer, containing 5pl of TEV-protease (GibcoBRL, Cat.No. 
10127-017). For better elution the columns were incubated at 16"C for 1 h (shak- 
ing with 850 rpm). The eluate was applied on fresh Mobicol columns containing 
0.2 ml settled calmodulin affinity resin (Stratagene, washed 3 times with CBP 
wash buffer). 0.2 ml 2 times CBP binding buffer (10 mM Tris. pH 7.5, 100 mM 
NaCI, 0.1 % IGEPAL, 2 mM MgAc, 2 mM imidazole, 4 mM CaCia, 1 mM DTT) 
were added followed by an incubation of 1 h at 4°C while rotating. Protein- 
complexes bound to the calmodulin affinity resin were washed w/ith 10 ml CBP 
wash buffer (10 mM Tris, pH 7.5, 100 mM NaCI, 0,1 % IGEPAL, 1 mM MgAc, 1 
mM imidazole. 2 mM CaCIa, 1 mM DTT). They were eluted by the addition of 600 
pi CBP elution buffer (10 mM Tris. pH 8.0, 5 mM EGTA) for 5 min at 37°C (shak- 
ing with 850 rpm). The eluates were transferred into a siliconized tube and lyophl- 
llzed. The calmodulin resin was boiled for 5 min in 50 pi 4x Laemmli sample 
buffer. The fractions were combined and applied on gradient NuPAGE gels (Invi- 
trogen, 4-12%. 1.5 mm, 10 well). 

For the purification of membrane TAP-tagged proteins 5 x 108 adherent cells (cor- 
responding to 40 15 cm plates) were used. The cells were harvested and washed 
3 times in cold PBS (3 min, 1300 rpm, Heraeus centrifuge). The cells were frozen 
in liquid nitrogen and stored at -80°C, or the TAP purification was directly contin- 
ued. 

The cells were lysed in 1 0 ml membrane lysis' buffer (50 mM Tris, pH 7.5, 7.5 % 
glycerol, 1 mM EDTA, 150 mM NaCI, 25 mM NaF, 1 mM UasVO^, 1 mM DTT, 
containing 1 tablet of protease Inhibitor cocktail (Roche) per 25 ml of buffer) by 
pipetting 2 times up and down, followed by a homogeniang step (10 strokes In a 
dounce homogenizer with tight pestle). After spinning for 10 min at 1300 rpm 
(Heraeus centrifuge) the supernatant was subjected to an ultracentrifugatlon step 
of 1 h at 100000 g. The "default" supernatant was frozen In liquid nitrogen and 
stored at -80*C. or the TAP purification was directly continued. TTie "membrane" 
pellet was resuspended in 7.5 ml membrane lysis buffer (+ 0.8% IGEPAL) by 



wo 2004/099774 PCmP2004/004883 

29 

pipetting, followed by resuspension through a gauge needle for 2 times. After 
incubation for 1 h at 4'C (while rotating) the lysate was cleared by a centrlfugatlon 
step of 1 h at 100000 g. The "membrane" supematant was frozen In liquid nitro- 
gen and stored at -80*C, or the TAP purification was directly continued. 

The lysates were thawn quickly In a 37'C waterbath. 0.4 ml of unsettled rabbit 
IgG-Agarose beads (Sigma, washed 3 times in Membrane lysis buffer) were 
added and incubated for 2 h rotating at 4*C. Protein complexes bound to the 
beads were obtained by centrifugation (1 min, 1300 rpm. Heraeus centrifuge). 
The beads were transferred into 0.8 ml Mobicol M1002 columns (Pierce) and the 
membrane fractions were washed with 10 ml membrane lysis buffer (containing 
0.8% IGEPAL and 1 tablet of Protease inhibitor cocktail (Roche) per 50 ml of 
buffer). The default fractions were treated the same way but the buffer contained 
only 0.2% IGEPAL. After an additional washing step with 5 ml TEV cleavage 
buffer (10 mM Trie, pH 7.5, 100 mM NaCI, 0.5 mM EDTA, 1 mlVI DTT, containing 
0.5% IGEPAL for the membrane fraction and 0.1% IGEPAL for the default frac- 
tion), the protein-complexes were eluted from the beads by adding 150 pi TEV 
cleavage buffer, containing 5 pi of TEV-protease (GibcoBRL, Cat.No. 10127- 
017). For better elution the columns were incubated at 16°C for 1 h (shaking with 
850 rpm). For the membrane fraction 3 additional pi of TEV-protease were added 
and incubated for another hour. The eluate was applied on fresh Mobicol col- 
umns containing 0.2 ml settled calmodulin affinity resin (Stratagene, washed 3 
times with CBP wash buffer). 0.2 ml 2 times CBP binding buffer (10 mM Tris, pH 
7.5, 100 mM NaCI. 0,1 % IGEPAL, 2 mM MgAc. 2 mM imidazole, 4 mM CaC!2. 1 
mM DTT) was added followed by an incubation of 1 h at 4°C rotating. Protein- 
complexes bound to the calmodulin affinity resin were washed with 10 ml of CBP 
wash buffer (10 mM Tris, pH 7.5, 100 mM NaCI, 0.1 % IGEPAL. 1mM MgAc, 1 
mM imidazole, 2 mM CaCb. 1mM DTT). They were eluted by addition of 600 pi 
CBP elution buffer (10 mM Tris, pH. 8.0. 5 mM EGTA) for 5 min at 37*C (shaking 
with 850 rpm). The eluates were transfen-ed into a siliconized tube and lyophi- 
lized. The calmodulin resin was boiled for 5 min in 50 pi 4x Laemmll sample 
buffer. The fractions were combined and applied on gradient NuPAGE gels (Invi- 
trogen. 4-12%. 1.5 mm, 10 well). 

The composition of the protein complex was analyzed as described below. 



wo 2004/099774 



30 



PCT/EP2004/004883 



Gel-separated proteins were reduced, alkylated and digested In gel essentially by 
following the procedure desCTibed by Shevchenko et al. (Shevchenko. A.. Wllm. 
M.. Vorm. O., Mann. M. Anal. Chem. 1996. 68, 850-858). Briefly, gel-separated 
proteins were excised from the gel using a clean scalpel, reduced using 10 mM 
DTT (in 5 mM ammonium bicarbonate, 54 'C, 45 min) and subsequently alky- 
lated with 55 mM iodoacetamlde (in 5 mM ammonium bicarbonate) at room tem- 
perature in the dark (30 min). Reduced and alkylated proteins were digested in 
gel with porcine trypsine (Promega) at a protease concentration of 12.5 ng/pl in 5 
mM ammonium bicarbonate. Digestion was allowed to proceed for 4 hours at 37 
"C and the reaction was subsequently stopped using 5 pi 5% formic add. 

Gel plugs were extracted twice with 20 pi 1% TFA and pooled with acidified di- 
gest supematants. Samples were dried in a vaccum centrifuge and resuspended 
in 13 pi 1%TFA. 

Peptide samples were injected into a nano LC system (CapLC. Waters or Ulti- 
mate, Dionex) which was directly coupled either to a quadnjpole TOP (QTOF2, 
QTOF Ultima. QTOF Micro. Micromass or QSTAR Pulsar. Sclex) or ion trap 
(LCQ Deca XP) mass spectrometer. Peptides were separated on the LC system 
using a gradient of aqueous and organic solvents (see below). Solvent A was 5% 
acetonitrile in 0.5% formic acid and solvent B was 70% acetonitriie in 0.5% fomriic 
acid. 



Table 2 



Time (min) 


% solvent A 


% solvent B 


0 


95 


5 


5.33 


92 


8 


35 


50 


50 


36 


20 


80 


40 


20 


80 


41 


95 


5 


50 


95 


5 



spectrometer. 



wo 2004/099774 



31 



PCT/EP2004/004883 



The peptide mass and fragmentation data generated in tlie LC-MS/I\/IS experi- 
ments were used to query fasta formatted protein and nucleotide sequence data- 
bases maintained and updated regularly at tiie NCBI (for the NCBInr, dbEST and 
the human and mouse genomes) and European Bioinfonnatics Institute (EBI, for 
the human, mouse, Drosophila and C. elegans proteome databases). Proteins 
were identified by correlating the measured peptide mass and fragmentation data 
with the same data computed from the entries in the database using the software 
too! Mascot (Matrix Science, Perl<ins, D. N.. Pappin, D. J., Creasy, D. M., Cottrell, 
J. S., Electrophoresis 1999, 20. 3551-67). Search criteria varied depending on 
which mass spectrometer was used for the analysis. i 

: I 

! .. I 

Proteins identified are: 

- E1 -component of the alpha-l<etoglutarate dehydrogenase complex: alpha- 

ketoglutarate dehydrogenase or oxoglutarate dehydrogenase (OGDH; EC 
1.2.4.2) 

- E2-component of the alpha-ketoglutarate dehydrogenase complex dihydrolipoyi 

sucdnyltransferase (OMIM-No. 126063; OMIM:"Online Mendelian Inheri- 
tance In Man", database available at the National Center for Biotechnology 
Information, www.nd3i.nlm.nih.gov) 

- E3-component of the alpha-ketoglutarate dehydrogenase complex: dihydrolipoyi 

dehydrogenase (OMIM-No. 246900) 

The a-ketoglutarate dehydrogenase complex is a multienzyme complex consist- 
ing of 3 protein subunits, alpha-ketoglutarate dehydrogenase (Elk. or oxogluta- 
rate dehydrogenase; OGDH); dihydrolipo^ sucdnyltransferase (E2k. or DLST); 
and dihydrolipoyi dehydrogenase (E3). The complex catalyzes a key reaction in 
the Krebs tricartjoxylic add cycle. 

Alpha-ketoglutarate dehydrogenase (E1k) catalyzes the conversion of alpha- 
ketogluterate to succinyl coenzyme A, a critical step in the Krebs tricarboxylic 
acid cycle. Deficiencies in the activity of this enzyme complex have been ob- 
served in brain and peripheral cells of patients with Alzheimer's disease. 



wo 2004/099774 PCT/EP2004/004883 

32 

The DLST gene maps to 14q24.3 and the E3 gene maps to chromosome 7. A 
second related sequence, possibly a pseudogene, was identified and mapped to 
chromosome 10, pointed to a possible significance to the finding of a reduction in 
the activity of this complex In Alzheimer disease brain and cultured skin fibro- 
blasts from Alzheimer disease patients. (Reference: OMIM 203740), 

The association between carbonyl reductase and alpha-ketoglutarate dehydro- 
genase points to a role of the carbonyl reductase in protecting alpha-ketoglutarate 
dehydrogenase from inactivation by reactive metabolites. 

i 

For example, 4-hydroxy-2-nonenal (HNE) is a highly toxic product^of lipid peroxi- 
dation. HNE inhibits mitochondrial potent inhibitor of mitochondrial respiration. 
HNE inhibits alpha-KGDH. 

Example 2: identification of the protein complex associated with the drug 
parthenollde 

Identification of the interaction between oarthenolid e and IKKbeta 

Parthenolide Is a natural compound that can be isolated from the medicinal herb 
feverfew (Tanacetum parthenium). It Is known from traditional medicine that 
parthenolide has anti-inflammatory properties. In order to identify the molecular ^ 
(intracellular) target for this comound a parthenolide affinity reagent was synthe- 
sized. 

The experimental procedure was carried out as described in Kwok et al. 2001, 
Chemistry & Biology 8, 759-766) 

Biotinylated parthenolide was synthesized by oxidation v^th selenium dioxide and 
tert-butylhydroperoxide to produce the allylic alcohol. The next steps were esteri- 
fication of the allylic alcohol with 12-(Fmoc amino) dodecanoic acid (Mitsunobu 
conditions), removal of the Fmoc group with tetrabutylammonium fluoride and 
coupling with biotin using N-[dimethylamino)-1H-1.2,3-triazolo-[4.5-blpyridino-1- 
yimethylene]-N-methylmethanaminimum hexa-fluorophosphate N-oxide/di- 



wo 2004/099774 PCT/EP2004/004883 

33 

isopropylethylamine. The biotlnylated parthenolide product was verified by nu- 
clear magnetic resonance (NMR) and electrospray nnass spectroscopy. 

This affinity reagent was used to isolate proteins that bind to parthenolide from 
human cervical carcinoma cells (HeLa). Affinity purification utilized steptavidin 
resin which tightly interacts with biotin (steptavidin-resin pull-down experiment). 
IKKbeta was identified as a parthenolide binding protein. 

Identification of proteins binding to IKKbeta 

The IKKbeta protein was fused to the TAP-affinity tag and expressed in Hel< 293- 
cells. TAP purification followed by mass spectrometry analysis identified a protein 
complex that contained the IKKalpha protein. 

The identification was carried out essentially as described in Example 1 . 
As a cell line, Hek 2g3-cells were used. 

Screen for inhibitors of IKKalpha 

The IKKalpha protein is a kinase. Kinases are considered a target class that is 
pharmaceutically attractive. For kinases that play a role in disease pathways en- 
zymatic assays can be designed that allow for the identification of inhibitors. A , 
number of inhibitors against kinases have been developed that have utility In 
treating diseases (e.g. cancer or inflammation). 

In particular, an enzymatic assay for the IKKalpha kinase was described that al- 
lowed the identification of small molecule inhibitors (Burke et al., 2003, JBC 278, 
1450-1456; PMID: 12403772). In this assay the enzyme-catalysed phosphoryla- 
tion of a GST-lkappaBalpha substrate was performed by adding purified IKKal- 
pha enzyme and radioactively labeled gamma pP]ATP in a suitable buffer. 
Reaction samples were analyzed by SDS-polyacrylamide gel electrophoresis and 
the radioactivity incorporated into the substrate protein was quantified by autora- 
diography. 



wo 2004/099774 PCT/EP2004/004883 

34 

Alternatively, a 17-amino acid peptide corresponding to amino acids 26-42 of 
IkappaBaipha can be used as substrate (PIVIID: 9575145; PIVIID: 10593898). The 
samples are analyzed by HPLC analysis (PMID: 9207191 ) and the amount of 
IKK-catalyzed incorporation of ^^P into the peptide substrate is quantified by liquid 
scintillation counting. 

Alternatively, a non-radioactive kinase assay can be used to identify IKKalpha 
inhibitors. This assay is fluorescence-based and as a readout the change of fluo- 
rescence polarization is measured (PMID: 10803607; PMID: 11020319). This 
assay can be performed in a homogeneous way, a simple mix-and-read fonmat, 
where no separation steps are required and therefore can be used for high 
throughput screening (HTS) of small molecule libraries. The Fluorescence Polari- 
zation (FP) -based protein kinase assay uses fluorescein-labeled phosphopep- 
tides bound to an anti-phosphotyrosine antibody (or antl-serine / anti-threonine 
antibodies for serine/threonine kinases). Phopsphopeptides generated by a ki- 
nase compete for this binding. In kinase reactions, polarization decreases with 
time as reaction products displace the fluorescein-labeled phosphopeptide from 
the anti-phosphotyrosine (or anti-phosphoserine/threonine) antibodies. For IK- 
Kalpha a fluorescein-labeled peptide corresponding to amino acids 26-42 of 
IkappaBaipha containing phosphoserine at position 32 or 36 is used as tracer 
molecule. Non-fluorescent non-phosphorylated peptides of the same sequence 
serve as substrate for the IKKalpha kinase. Once these substrate peptides are . 
phophorylated by the kinase, they displace the fluorescent phosphopeptide tracer 
from the anti-phosphoserine antibody and the polarization signal decreases.