Note: Descriptions are shown in the official language in which they were submitted.
CA 02301633 2000-02-24
DESCRTPTION
?iLTROD ARD ICIT F08 IDENTIFYING 119TERAGTIONS BE1'ilEEl~ PROTEINS OR PEPTIDES
The present invention relates to a method and to a kit for the identification
or determination of interactions between prot~ias or peptides by means of
fluorescence-resonance energy transfer.
The identificatiott_and analysis of interactions bets~reen different proteins
or
peptides or fragments thereof represents as important problem of biopiedical
research and biotechnology. At the end of the 1980's a system W88 coo- .
sequently developed, which acquired great research signif~.cance under the
name "Yeast Two-Nyhrid-System" (Fields et ai.,.Natnre, vol.. 340, pp z45~247,
1489). Ti~.ts system is based on the discovery that cellular transcription
activators,. such as e.g. GAL4 or lexA from yeast could be deCOmposed in Lw0 f
independent fvmtctloa domains. Both domains are normally a component .af a
protein in the nucleus of the yeast cell, which binds to.specific actipatiag
sequences~~of diffazeat target genes and regulates the transcription thereof.
Ons domain,~the~.DNA~~binding domain (8D), specifically binds to.a particular
DHA target seqaeace (upstream activating seqaeace) in the vicinity of the
target gene promoter. The oilier domain, the activatins domaf.u (AD),
increases the transcription rate of the target gene by interaction pith the
transcription initiation complex, which is bound to the protater of the
target gene. In the "Yeast Two-Hpbrid System" this structure is utilized by
transcription factors in modified form. The DNA binding domain (BD) of GAL4
or lesA is there expressed as a fuaioa protein with a "baiting protein. or
peptide" in yeast cells. This fusion protein also has a nucleus location
shgnai through which it is transported into the yeast nucleus. The.baiting-
fusi.on protein is bound there to a target sequence (UAS), which in the yeast
strain used is located in the viciaf.ty of prvmaters of two reporter genes
(e. g. anxotrophic i~aarker (HIS3) and enzy~tatic mar$er (lacZ)). This
leads to s'conStellatioa in which the baiting protein or peptide is exposed
~in direct. spatial proximity to the reporter gene promoter. Additionally a
second fusion protein is expressed in the same yeast.cell. Zt Comprises the
activating domain (AD? of GAL4 or LexA sad a catchingJpreying protein or
peptide. It also has a nucleus ~.ocation signal. The catching-fnsiot~ protein
is cvnseqneatly also transported into the yeast nucleus. If the catching
protein and the baiting protein exposed at the BAS physically interact xith
CA 02301633 2000-02-24
- 2 -
One another, then there is an increase in the statistical probability that
the activating domain is located in the vicinity of the reporter gene
promoter. This leads to an increase ~.n the transcription of the reporter
genes, Whose extent is proportional to the strength of the interaction
between the baiting and catching protein. The "Yeast Two-Hybrid
System" can be used both for the quantitative analysis of knoaa bait/catch
pairs and for the identification of unknown catching proteins or peptides_
Possible catthing proteins are e.g. a cDNA library or a combinatory peptide
librazy.
Despite the above-described multiplicity of uses the "Yeast Two-Hybrid System"
suffers from limitations due to the transcription-dependent detection system.
They occur e.g. if the catching aad/nr baiting protein contains localization
signals. These interfering localization signals are e.g. hydrophobic trans-
membrane domains, such as occur in many membrane proteins. They lead to a
transportation of the fusion protein into the cellular membrane, whilst the
nucleus localization signal of the fusion protein is ignored. Therefore
interactions with proteins having such transmembrane domains cannot be detec-
ted. A further problem arises on screening for the analysis of cDNA libra-
ries yr combinatory peptide libraries, which can only be carried out to a
limited Qxtent. Due to their complexity (e. g. more than IO bilZ~,vn pos5xble
variants far a decapQptide) the latter make efficient screening ittpvssible.
The analysis of possiblQ interaction partners has therefore hitherto tdken
place in that a maximum of 50,000 transformed yeast cells per agar plate are
plated out and incubated initially for 3 to 7 days in the incubator. This
weans that a typical test of approximately 5 ~llion possible fusion protein
pairs requires the use of 100 agar plates, which must be thoroughly tested.
Thus, the "Yeast Two-Hybrid System" involves high material, time, laboux and
culture space costs for this screening process type. Iu addition, there x,s
no known automated analysis process, which could reduce and standardize the
high laboratory staff costs.
Thus, the problew of the present invention is to pxovide a method and a
corresponding kit for the idQntification of interactions between proteins or
peptides, which at least partly avoid the aforQmentfoned disadvantages and
which can 3.n particular be carried out in an automated manner.
CA 02301633 2000-02-24
- 3 -
They lead to a transportation of the fusion pxotein into the cellular mem-
brane, whilst the nucleus localization signal of the fusion proteia is
ignored. Therefore interactions with proteins havi~tg such transmembrane
domains cannot be detected. A further problem arises on screening far the
analysis of eDPiA libraries or combinatory peptide libraries, which can only
be carried out to a Limited extent. Due to their complexity (e. g. more than
billioa possible variants for a decapeptide) the lattex make efficient
screening iuipvssible. The analysis of possible interaction partners has
therefore hitherto taken place in that a maximum of 50,000 transformed yeast
cells per agar plate are plated out and incubated initially for 3 to 7 days
in the incubator. This means that a typical test in the order of 5 million
possible fusion protein pairs requires the use of 100 agar plates, which roust
be thorpughly tested. Thus, the "Yeast Two-gyb~d System" involves high
material, time, labour and culture space costs for the screening processes.
In addition, there is no known automated analysis process, which would
reduce and standardise the high laboratory staff costs.
International patent application W0 97/27212 describes a screening method
with which pegtide sequences are identified which measurably modify the
phenotype of a cell. The method is based oa the measuretuent of the phenotype
change after expression of the peptide in the cell. Initially no inter-
actions of peptides or proteins are investigated. Only in a second step is
it possible tv carry out a closer investigation of the identified, active
peptide sequences. For thin purpose various Ialown methods can be alter-
natively used, e_g. the absence of aormal cellular functions can be measured,
a two-hybxid system in mammalian cells can be used or a fluorescence-
resonance energy transfQr system can be e~uployrd.
The literature already describe souse test systems for investigating l~oWn
interaction partners, which utilize fluorescence-resonance eaergy transfer.
For example, in international patent app~Cation WO 9I/12530 a fluorescence
test is described for investigating interactions of specific binding partners,
particularly With respect to immunoassays. European patent application EP
0242527 describes an aaalysis process for interacting analytes using the
energy transfer system.
Amended sheet
CA 02301633 2000-02-24
- 3a -
Thus, the probie~ of the present invention is to provide a method and a
corresponding kit for the identification of interactions between proteins or
peptides, which at Least partly avoid the aforementioned disadvantages and
whf.ch can in particular be carxied out i.n an automated manner_
This problem is solved by the subject matter of the claims and their wording
is hereby made by reference into part of the content of the present descrip-
tion_
Amended sheet
CA 02301633 2000-02-24
- 4 -
This problem is in particular soloed by a method in which at least two
proteins or peptides are coupled to different fluorescent components, the
absorption and emission spectra of the fluorescent components overlapping in
such a way that fluorescence-resonance energy tra~zsfer (FRET) is possible and
the components are so brought together by an interaction between the proteins
or peptides that FRET occurs and is measured. The genetic information far at
least two fusion peptides or proteins having a peptide or pxotein fraction
attd in each case one different fluorescent protein yr peptide is present in
an expression systeta introduced into a host ceh, the emis8ion arid absorption
Spectra of~the fluorescent preteins or peptides overlapping i~1 such a way
that fluoresenCe-resonance energy transfer (FRET) occurs and is measured in
the host cell.
Within the scope of the present invention the terms "protein" and "peptide"
are mutually interchangeable and contain an amino acid line-up of any length
and complexity, i.e. dipeptides, vligopeptides, polypegtides, Complete pro-
teins, ~ragments thereof, antibodies, domains, epitopes, etc. They can in
particular be combinatory peptide libraries or the expression products of
cDNA libraries. The protcins/peptides to be investigated for interactions
are subject to no restrictions and can be freely chosen.
According to the invention, the term fluorescent component is in particular
understood to mean a material or particles having a fluorescent marking at
the surface. They can he latex particles or matrices, as acre of a conven-
tional nature for automatic protein synthesis.
According to the invention coupling means a more or Less firm binding of the
Amended sheet
CA 02301633 2000-02-24
- 5 -
fluorescent component to the proteins) or peptide(s). This bond tends to
be of a more adsorptive nature or can also be a covalent bond. in a pre-
ferred embodiment the protein or peptide and the fluorescent component can be
present in the form of a fusion protein or peptide. Coupli~xg can also take
place by means of a linker. The latter covers connections of ah types
suitable for linking two mplecules_
The method according to the invention is inter alxa based on the known
phenomenon of fluorescence-resonance energy transfer (FRET), diagrammatically
represented i.n fig. 1. A fluorescent molecule absorbs photons with a charxsc-
teristic wavelength and liberates again the thus absorbed energy within a
very short time by the emission of photons, which gives rise to a measurable
fluorescence. Since during this process there is a certain energy loss by
heat evolution, the e~.tted photon has a characteristic, reduced energy
content and therefore a modified wavelength cotapared with the previously
absorbed photon (Stokes shl.ft). Both parameters, namely the absorption
wavelength and the Stokes shift (and the emission wavelength dependent there-
on) are charatteri-stic parameters of any fluorescent molecule and are a func-
tion of the characteristics thereof. A fluorescence-resonance energy trans-
fer tan be measured if two fluorescent molecules interact in a mixture with
one another and the absorption wave~,ength of one overlaps with the emission
wavelength of the other. In this connection
Amended sheet
CA 02301633 2000-02-24
- 6 -
enzymatic ligation and cloned in an expression vector using standard methods.
It ix also possible to use here a randoptly complex population o~ baiting and
catching proteins in order to obtain a library of possible interaction
partners. According to the invention genetic information and coding DNA
sequence means cDNA or genomic DNA, preferably cDNA.
Known, conventional vectors can be used as expression vectors. However, they
should permit a separate selection of "baiting protein vectors" and "catching
protein vectors" in the ~vst cells, e.g. by different antibiotic resistance
genes (e. g. ampicillin, kanamycin, chloroamphenicol, streptomycin, tetra-
cyclines, sulphonamides) oz different auxotrophic markers (e.g. LEU2, 8IS3 or
TRPl)_ However, it is also possible to combine both fusion proteins together
in a bicistronic expression vector. 5nitable expression vectors for the
expression in E.coli are e.g. pGEMEIC, pUC derivatives, pGEg-2T, pET3b, pQE8
or
pQE42, for the expression in yeast pY100, Ycpadl, pGBT9 or pGAD424, for the
expression in animal cells pKCR, pEFBOS, cDMB and pCEV4. In parti.CUlar, the
Baculovirus expression vector pAcSGNt-A is suitablQ for expression in insect
ce118.
The expert is also aware of methods and host cells for expressing the
expression vector and the fusion protein coded by it. Examples of such cells
are the E.cali strains HB10I, DH1, x1776, JM1O1, Jm109, B121 and SG13009, the
yeast strain Saccharou~Ces cerevisiae, Schi2osaccharo~ces pombe, Y 190,
CG1945, EGY48 or HF7, the animal cells L, 3T3, FM3A, CHO, COs, Vero and Hela
and the insect cells Sf9. Suitable cells, preferably yeast cells, are
(sequentially or synchronously) transformed with the expression vector yr
vectors and preferably propagated under double selection.
The fluorescence of the cells is then excited at the absorption maximum of
one of the two fluorescent proteins (preferably at the maximum of FP-A),
measured at the emission maximum of the other fluorescent protein (preferably
of FP-B) and the ~ueasnred result is used as a selectf.on criterion in the
ipso,
lation and cloning of cells contafning a potential interaction pair. This
measurement i.s calibrated bQforehand by cells in which PP-A and FP-B are
expressed cloned in different vectors, but witheut fusing with a further
protein domain (negative control). The positive control can be constituted
CA 02301633 2000-02-24
by a fusion protein in which FP-A and FP-B (~rithont baiting or catching pro-
tein) are coupled to one another in the above-described manner and con-
sequently there is a maxiirium spatial proximity between FP-A and FP-B,
Fluorescence is preferably exca.ted by laser. FRET measurement takes place
using suitable filter combinations, which ran be chosen by the expert as a
function of the absorption and emission maxima of the fluorescent proteins
used_ In accordance with the bases of FRET each interaction between the
fusion proteins mast lead to a FRET rise, which can be detected after
appropriate calibration of the system. For the investigation of known bait/
catch pairs this detection preferably takes place by fluorescence microscopy.
For the screening of cDNA libraries ox combinatory peptide libraries measure-
ment preferably takes place by fluorescence-controlled cell sorting (FACS).
For further identification of the interaction partners determined by FRET
they are plated out on agar plates and plasmide-DtiA is iso).ated from the
individual colonies and can then be sequenced using standard methods.
The method according to the invention has e.g. the following preferred
applications:
- A known pair of possible interaction partners is jointly cxgressed in a
host cell and the resulting fluorescence spectrum is compared with the nega-
tive and positive controls or other interaction partners.
- A baiting-fusion protein is transformed or tx~ansfected with a randomly
complex mixture of catching-fusion proteins (jointly or sequentially) in
suitable host cells and preferably in such a way that each cell only carries
one defined pair of possible interaction partners or a small group of such
pairs. From the resulting host cell populations are then isolated the Cells
which have a significant FRET and consequently e~cpress a potential inter-
action pair_ By isolating the eacprassion vectors Contained xn these cells it
is easily possible to determine the sequence of the participating interacta.on
partners. The mixture of catching proteins could e_g. be made o~ cpNA
libraries or combinatory peptide libraries.
CA 02301633 2000-02-24
g _
The baiting protein could also be represented by a population of different
proteins_ It is conseQuently e.g. possible to determine an interaction
matrf.x, e.g. in that a cbNA library is searched for possible interaction
pairs between unknown proteins.
- A known pair of interacting fusion proteins is jointly expressed in host
cells and the fluorescence transfer between the two fusion proteins is deter-
mined_ Then a third protein or peptide is expressed in the same host cells
and its influence on FRET is trade an analysis criterion. This makes it pos3-
ible to datvrmine proteins or peptides which disturb or intensify an existing
interaction between known interaction partners.
The advantages of the method according to the invQntion axe that unlike is
the "Yeast Two-Hybrid System" it is independent of transcr~,ption-regulating
mechanisms and the location of fusion proteins in the nucleus is nn~.mportant.
Thus, e.g. also intact membrane proteins can be used as baiting proteins.
The possibility of an automated cell, sorting additionally exists, which
permits the obtaW ing of a high screening complexity. Thxough the combin-
ation of a pharmacologically interesting target sequence and a combinatory
peptide library it is possible to screen several million peptide ligands per
time unit. Compared with known methods with which thin task might not be
performable, the method according to the invention is characterized by an
enormous time and labon.r saving.
The kit according to the invention (test IQt) is described in claim 8.
Preferred embodiments of this kit can be gathered from the dependent clairus
9 to 16.. The wording of claims 8 to 16 xs hereby made by reference into part
of the content of this description. Express reference is made to previous
parts of the description directly or indirectly linked with the features of
claims 8 to 16_
The claimed kit is used fOx performing the claimed method in a particularly
simple manner. the kit preferably contains in a suitable container a first
vector, which carries the host cell-optimized coding sequence of a FRET donor
(under the control of a suitable promoter) and in which there is a cloning
CA 02301633 2000-02-24
_ c~
possibility fer the introduction of a coding sequence, which codes for a
first test protein/test peptide (i.e. a baiting or catching protein or
peptide).
There is also an associated second vector with host cell-optimized coding
sequence for the associated FRET acceptor and cloning possibility for the
coding sequence of a second test protein/test peptide (corresponding inter-
action partner for the bait-catch pair zo be formed).
The coding sequences for the test proteins/test peptides in the first and/or
second vQCtor may already be present. Preferably also Integrated CDNA
libraries or combinatory oligonncleotide Libraries are contained. In such
cases e.g. a cbftA library (e.g. from the human brain) is present.with one of
the FRET partners fused in a first vector and a second vector in which the
user can clone an owa protei.n/peptide, e.g. as bait, with the other FRET
partner,
The invention is described hereinafter relat~.ve to the drawings, wherein
shoo:
Fig. 1 A diagrammatic representation of fluorescence~resonance energy
transfer (FILET) .
Fig. 2 A diagrammatic representation of the energy shift.
Fig. 3 A diagran~uatic representation of FRET by mesas of the interact~.ou
of baiting protein with catching protein.
Fig, 4. A gene map Of the expression vector pGBT9.
Fig. 5 The gene map of the e~cpression vector pGAD42~.
The inventf.on is described by means of the followiag examples.
E~CAMpLE I
Initially a suitable vector system is formed. For this purpose and for
CA 02301633 2000-02-24
- 10 -
reasons of siuiplicity use is made of the plasmides pG8T9 (~K I605-A),
pGAD424 (#'K1605-B) of Clontech (Palo Alto, US) and the plasmides pRSET
$-P4-3 and pRSET $-S65T (Helm and Tsien, Curz-ent Biology 1996, 6: 178-182)_
pGBT9 is a possible "baiting protein vector" of the Yeast Two-Hybrid System
marketed by Clontech. The vector contains an ampicilLin resistance gene for
selection in bacteria, a TRP1 gene for anxotrophic selection in yeast and an
expression cassette for the DNA-binding domain of the GAL4 transcription
factor (fig. 4, Clontech Hatchu~aker "GAIk Two-gybrid Vectors $andbook,
~pT3062-1) wader the control of an alcohol dehyd=ogenase promoter (ADH1)
for expression in yeast. The GAL4-DNA-binding domain is an essential, com-
ponent of the Yeast Two-Hybrid System.
For the purpose of the invention the GAL4 DNA-binding domain together with
its nucleus location signal is removed, is that it is cut out by moans of a
HinDIII/EcoRI restriction digestion.
pGAD4~24 is a possible "catching protein vector" of the Yeast Two-Hybrid
System (fig. 5). This vector has a similar structure to pGBT9, except that
instead of the TBPI ~uarker it contains a L$U2 gene for selection in yeast and
in place of the GAL4 DNA-binding domain ~t contains the GAL4 activating
domain.
For the purpose of the invention also from this vector is removed by a
HiaDIYI/EcoRI digestion the GAL4 activating doma~.n. For this purpose
initially further HinDIII restriction sites present in the vector are removed
by partial digestion, filling the restricting site with Pfu polymerase and
subsequent "blunt end Iigation".
Suitable coding sequences of GFP variants S65T and P4-3 (Heim and Tsien,
Current Biology 6: 178-182, 1996) are now amplified by PCR and provided at
their S' and 3' ends With AinDIxI or EcoRI restriction sites. For this
purpose use is made of the oli.gonncleotide primers GFP-Eco RI-3'(3'-
CGGGAATTCTTTGTATAGTTCATCCAT-3'), and GFP-HinDIII-S'-
TCCAAGCTTATGAGTAAAGGAGAAGAACTT-3') (Carl Both GmbR, Karlsrnhe). As the
coding sequences of P4-3 and S65T are identical in their 5' and 3'-terminal
CA 02301633 2000-02-24
- 11 -
e.g, in that a cDNA library is searched for possible interaction pairs
between unknown proteins.
- A known pair of interacting fusion proteins is jointly expressed in host
cells and the fluorescence transfer between the two fusion proteins is detex.-
wined. Then a third protein or peptide is expressed in the same host cells
and its influence on FRET is made an analysis criterion. This makes it poss-
ible to determine proteins or peptides which disturb or intensify an existing
interaction between known interaction partners.
The advantages of the method according to the invention are that unlike in
the "Yeast Two-hybrid System" it is independent of transcription-regulating
mechanisms and the location of fusion proteins in the nucleus i.s unimportant.
Thus, e.g. also intact membrane proteins can be used as baiting proteins.
The possibility of an automated cell sorting additionally e~.sts, which
permits the obtaining of a high screening complexity. Through the combin-
ation of a pharmacologically interesting target sequence and a combinatory
peptide library it is possible to screen several million peptide ~.igands per
t3~e LZnit. Compared with known methods with which this task might not be
performable, the method according to the invention is characterized by an
enormous time and labour saving.
The kit according to the itwention (test kit) is described in claim 7.
Preferred embodiments of this kit can be gathered from the dependent claims
8 to I5. The wording of claims 7 to I5 is hereby taade by reference into part
of the content of this description. Express reference is trade to previous
parts of the description directly or indirectly linked with the features of
claims 7 to 15.
The claimed kit is used for performing the claimed method in a particularly
simple manner. The kd.t preferably contains iu a suitable container a first
vector, which carrzes the host cell-optimized coding sequence of a FRET
donor (under the control of a suitable prouaoter) and in which there is a
cloning possibility for the introduction of a coding sequence, which codes
Amended sheet
CA 02301633 2000-02-24
-lz-
for a first test protein/test peptide (i.e_ a baiting or catching protein or
peptide).
There is also an associated second vector with host cell-optimized coding
sequence for the associated FRET acceptor and cloning possibility for the
coding sequence of a second test protein/test peptide (corresponding inter-
action partner for the bait-catch pair to be formed).
The coding sequences for the test proteias/test peptides in the first and/ox
second vector map already be present. Preferably also 3.ntegrated cDIZA
libraries or combinatory oligonucleotide 7.ibraries are contained. In such
cases e.g. a cDNA library (e.g. from the human brain) is present with one of
the FRET partners fused in a first vector and a second vector in which the
user can clone an own protein/peptide, c.g. as bait, with the other FRET
partner.
The invention is described hereinafter relative to the dxawings, wherein show:
Amended sheet
CA 02301633 2000-02-24
- 13 -
cells, once with the baiting plasmide pGBT-BFP-P2g.2 and a further time with
the empty baiting plasmide vector pGBT-BFP. The fluorescence spectra of the
tr~m transformants are then compared is order to exclude those interaction
candidates rahich interact directly with the blue fluorescent protein, but not
with the baiting protein.
Candidates checked in this way are they plated out again on selective minimum
mediuta plates and the sequence of the potentially interacting catching domain
is determined by standard molecular biology methods.
EgAMpLE 2
Use is made of two vectors, the first being selected from the commercially
~evailable expression vectors fvr EBfP or EGFP, e.g. pEBFP-N1, pE$FP-NZ,
pEBFP-N3, pEBFP-CI, pEBFP-CZ, pEBFP-C3, pECFP, pECFP-GI (Clontech) and the
SeCOnd from the commercially available expression vectoxs for EGFP or EYFP,
e.g. pEGFP-NI, pEGFP-N2, pEGFP-N3, pEGFp-C1, pEGFp-C2, pEGFp-C3, pEYFP-N1,
pEYFP N2, pEYFP-N3, pEYFP-C1, pEYFP-CZ, pEYfP-G3.
In one of the tyro vectors a DNA, sequence is cloned in the usual way and this
codes a baiting protein, so that the baiting protein is read off as a fusion
protein wa.th the fluorescent protein. In the other vector a cDNA library ox
a combinatory oligonucleotide library is Cloned, so that the second fluore-
scent protein is read off as a fusion protein with in each case a protein
fragment or combinatory peptide. Both vectors together are then transfected
in known manner iu suitable mammalian cells, e.g. COS-7, NG-108, H1H/3T3, etc.
Following an adequate incubation time the cehs are then analyzed by FREx
microscopy (see Clegg, in "Fluorescence Imaging Spectroscopy and tsicroscopy",
pp I79-236, John Wiley & Sons, 1996) or in the above-described manner are
individualized by flnorescence,activated cell sorting (FRCS). The cells in
which FRET occurs are then used in the conventional manner, e.g. by cytoplast
iso7.ati.on with a miervpipette conventionally used in electrophps5.ology and
subseqaQnt RT-PCR used fvr determining the basic catch sequences.
