WO1997045745A1

WO1997045745A1 - Ex vivo/in vitro assay systems

Info

Publication number: WO1997045745A1
Application number: PCT/GB1997/001423
Authority: WO
Inventors: Michael White; Douglas Hurd
Original assignee: Amersham Pharmacia Biotech Uk Limited
Priority date: 1996-05-24
Filing date: 1997-05-23
Publication date: 1997-12-04

Abstract

A method of identifying interacting pairs of proteins comprises testing for possible interactions between the proteins in an ex vivo system, selecting pairs of proteins which apparently interact and performing a further test in an in vitro system in which the proteins are labelled such that a detectable signal is produced as a result of the proximity of the labels when the proteins interact.

Description

EX VIVO/IN VITRO ASSAY SYSTEMS

This invention is concerned with protein-protein interactions More specifically, the invention relates to improved methods for determining whether pairs of proteins are capable of interacting

The identification of a pair of interacting proteins can give valuable information for use, for example, in the elucidation of gene function and can also provide a new target for a high throughput drug screen Protein-protein interactions can be analysed either by ex vivo or by in vitro methods (4) In vitro methods are useful in some circumstances, for instance they allow a potential drug to be tested without the added complication of assessing whether the drug can cross a cell membrane They can also confirm that a suspected protein-protein interaction does take place, and may suggest the role of post translational modifications in regulating a protein-protein interaction (3) Ex vivo methods include the two-hybrid system (2), and the methods of Gramatikoff et al (9) or Bunter and Kingston (6) and have the advantage over in vitro methods that the conditions are closer to normal physiological conditions There are also several powerful ex vivo techniques which allow the researcher to clone from a cDNA library, a protein which interacts with a known protein In the two-hybrid system of Fields and Song (2), also described in US 5,283,173, two chimeric genes which encode fusion proteins are used to test for an interaction between a known protein and a protein of interest The first chimeric gene codes for a known protein, often called the bait protein, fused to the DNA-binding domain of a transcription factor The second chimeric gene codes for a protein of interest fused to the transcriptional activation domain Additionally, the protein of interest may not be known and could be derived from a cDNA library In a suitable host cell, if the protein of interest and the bait protein do interact they bring into proximity the DNA-binding and transcriptional activation domains This proximity is sufficient to cause transcription of a reporter gene placed under the control of a promoter containing a binding site for the DNA-binding domain

A modified version of the two-hybrid system is described in WO 96/30507 and by Yavuzer & Godmg (8) In this system a promoter capable of initiating transcription in vitro is present in the chimeric gene which codes for the protein of interest and such a promoter may also be present in the chimeric gene encoding the known protein The promoter enables the rapid transcription and translation of the proteins in vitro Thus, when a pair of interacting proteins is identified in a host cell by a two- hybrid system, by removing the chimeric genes (which may be in the form of plasmids) from the host cell and expressing the interacting proteins in vitro, the proteins may be further investigated Preferably, the in vitro promoter is upstream of the gene encoding the known protein or the protein of interest but downstream of the sequence encoding the DNA- binding domain or the transcriptional activation domain This means that the interacting proteins can be expressed in vitro independently rather than as the fusion proteins of the two-hybrid system

This invention aims to provide an assay system which uses the benefits of ex vivo and in vitro investigations to identify pairs of interacting proteins Such an assay system may then provide the basis for high throughput drug assays using pairs of interacting proteins

The present invention provides a method of investigating whether two proteins are capable of interacting with one another, which method comprises a) providing nucleic acids encoding a first protein and a second protein, b) performing an ex vivo test involving producing the first and second proteins using the nucleic acids, to indicate whether the first and second proteins may be capable of interacting with one another, c) performing a further test to indicate whether the first and second proteins are capable of interacting with one another in vitro, comprising the steps of i) producing the first and second proteins using the nucleic acids, n) labelling the first protein from i) with a first label capable of participating in a proximity detection assay, in) labelling the second protein from i) with a second label capable of participating in a proximity assay with the first label, iv) contacting the labelled first protein and the labelled second protein in vitro under conditions to allow protein-protein interactions, v) detecting an interaction between the labelled first protein and the labelled second protein by means of the proximity of the first label to the second label when the first and second proteins interact The invention therefore allows an apparent interaction between two proteins ex vivo to be confirmed in vitro Pairs of proteins showing an interaction in the ex vivo test are selected for further testing in vitro The method enables the identification of pairs of interacting proteins, with greater certainty and speed compared with other methods

The proximity of the first protein label to the second protein label brought about by an interaction between the two proteins results in the production of a detectable signal This may be achieved by eg a scintillation proximity assay (SPA) system, in which the first protein label is a radiolabel suitable for use in SPA and the second protein label is a fluorescer comprised in a solid phase The detectable signal is light energy emitted when the labelled first protein interacts with the second protein which is immobilised on the solid phase, and the radiolabel is thus brought sufficiently close to the fluorescer. Scintillation proximity assay technology is described in US 4,568,649.

Alternatively, the detectable signal may be a change in an existing signal output, eg. fluorescence. Fluorescence resonance energy transfer (FRET) is a method which works on this principle and is described by Erickson and Cerione (12). It employs two different fluorescent molecules, a donor and an acceptor, such that when these are in sufficient proximity to one another the fluorescence of the donor molecule is absorbed by the acceptor molecule and light of a longer wavelength is emitted. Thus, when there is an interaction between two proteins each of which is labelled with one of these fluorescent molecules, a detectable signal is produced.

By such methods as are described above, the in vitro test in the method according to the invention may be performed as a homogeneous assay without the need for a separation step after the first and second proteins have been contacted under conditions to allow protein-protein interactions. The preferred format for the in vitro test in the method according to the invention is a homogeneous format.

As noted above, the invention may involve immobilising the second protein on a solid phase. The solid phase may be of any suitable material and it may take a variety of forms, for example it may be a surface or surfaces of a vessel, or particles such as beads. In the embodiment of the invention described above using scintillation proximity assay technology, the solid phase will comprise a suitable material impregnated with a fluorescer and preferably the solid phase will be scintillation proximity assay beads. The use of a solid phase to immobilise one of the proteins in the in vitro test in the method according to the invention will provide an improved method for use in high throughput screening.

In a preferred method according to the invention, step b) is performed using a two-hybrid assay, although other systems known in the field, such as those of Gramatikoff (9) and Bunter & Kingston (6), could be used. A particularly preferred adaptation of the two-hybrid assay for use in the method according to the invention is the system described by Yavuzer et al in WO 96/30507 and (8), as already discussed herein. The Yavuzer et al system employs a promoter which allows in vitro expression of one or both of the chimeric genes which encode the proteins being tested. This makes the transition from the ex vivo test to the in vitro test in the method according to the invention much more straightforward, since it can be done by simply isolating the chimeric genes from the host cell eg. by harvesting the plasmids containing the genes, and transferring them into an in vitro system.

Alternatively, where a suitable in vitro promoter is not already present, the transfer from ex vivo to in vitro testing could be brought about by amplification of the first and second protein coding regions from the two- hybrid system, preferably by the polymerase chain reaction (PCR), using a primer which contains a promoter having activity in vitro. The nucleic acids thus produced are capable of being expressed in vitro.

As a further alternative, the first and/or second proteins may be produced in an ex vivo system eg. bacterial cells such as E.coii, and then isolated for use in the in vitro test according to the invention.

When a solid phase is used in the in vitro part of the assay method according to the invention, it may need to be adapted in some manner to facilitate immobilisation of the second protein. The solid phase may thus have attached to it a ligand capable of specifically binding to the second protein. Additionally, the second protein may be provided with an attachment label which specifically binds to the ligand on the solid phase. Suitable attachment labels for the second protein and ligands for the solid phase, which labels and ligands are members of specific binding pairs, are well known in the art. Examples include biotin and streptavidin; for example the second protein may be labelled with biotin and the solid phase coated with streptavidin.

The second protein can be immobilised on the solid phase by a variety of methods, including any of the following:

1 ) By a biotin-streptavidin linkage as already described and as 5 illustrated in Figure 2. Avidin could also be used in place of streptavidin.

2) By a metal ion-histidine link. If the second protein includes a tag of multiple histidine residues (eg. 6), it can then be immobilised on the solid phase via a metal ion-histidine link.

3) By an antibody. The solid phase can be coated first in i o protein A which will enable an antibody specific for the second protein to bind via its Fc portion. The second protein can then be immobilised on the surface by binding to the antibody. For this purpose, an epitope tag recognised by the antibody may be incorporated into the second protein either by means of gene fusion or by direct attachment to the second

15 protein. Alternatively, the antibody could recognise an epitope within the second protein itself.

Labelling of the proteins may be carried out during expression of the genes in vitro. For example, the first protein may be expressed by in vitro transcription of the gene encoding it, followed by in

20 vitro translation during which ³⁵S-methionine is incorporated into the protein. Other suitable labels can be incorporated into the first protein in this way. ³H or ¹²⁵l as well as ³⁵S are examples of radiolabels suitable for use in scintillation proximity assays. The second label, or an attachment means for the second label, may be incorporated into the second protein

25 during translation in vitro eg. by including tRNA-biotinylated lysine in the in vitro translation reaction (5). Where the first and/or second proteins are produced by eg. E.coli for the in vitro test, labelling could be carried out by including the label in the growth medium, eg. in a form of a radiolabelled amino acid for the first protein label.

30 Alternatively, labelling of the first and/or second protein could be carried out chemically in vitro once the protein has been produced Where FRET is used, this is the preferred labelling method

Finally, the in vitro protein-protein interaction test can be carried out by bringing the appropriately labelled first protein into contact with the labelled second protein, in a suitable aqueous environment, and the reaction monitored for any signal produced

The invention is further illustrated in the attached figures, in which

Figure 1 illustrates vectors for use in the modified two-hybrid system of Yavuzer et al (8), in which the activator gene and bait gene encode the two proteins being investigated

Figure 2 illustrates a preferred way of carrying out the invention, using activator and bait vectors isolated from a two-hybrid system This scheme provides a practical approach to link ex vivo gene discovery by the yeast two-hybrid system to drug development assays by SPA

Figure 3 illustrates a scheme for identifying interacting proteins in a library versus library screen, using an automated system

Figures 4, 5 and 6 show results of the method according to the invention

One particular embodiment of the method according to the invention involves in vitro transcription and translation of both the first and second proteins The first protein is radiolabelled during in vitro translation with eg ³⁵S methionine and the second protein is labelled with biotin by including eg tRNA-biotinylated lysine in the in vitro translation reaction Automated processes are particularly desirable, and in vitro transcription/translation is one process which is capable of being automated. If large numbers of protein-protein interactions, which have been identified in vivo eg. by the two-hybrid method, are to be analysed in vitro eg. by SPA, the genes which encode the interacting proteins can be rescued from the ex vivo system eg by traditional methods such as plasmid rescue or by PCR Then, if an appropriate promoter has been placed in front of the genes, both the first protein and the second protein can be rapidly produced by in vitro transcription and translation, labelled and assayed by a scintillation proximity assay After the ex vivo test_, the remainder of the method according to the invention can be carried out by an automated process, for example on a Molecular Biology Robot available from Amersham (Vistra ™) A possible scheme for automation is shown in Figure 3 This will enable quantitative study of interactions between large numbers of different proteins in vitro

The method according to the invention thus makes it feasible to study large numbers of proteins and will be useful for rapid setting up of drug assays Furthermore, and in particular since automation is possible, the system described herein can be used to screen cDNA libraries against known proteins or against further cDNA libraries A library versus library screen using a two-hybrid assay is outlined in Figure 3 Library versus library two-hybrid screens involve setting up an interaction assay, using a cDNA library cloned into the activator plasmid and the same cDNA library cloned into the bait plasmid PCR is performed with oligonucleotides specific for the bait and activator plasmids to isolate sequences encoding potentially interacting proteins from the in vivo assay The 5' oligonucleotides contain a promoter, eg the T7 promoter These sequences are transcribed and translated in vitro with biotin and radioactive labelling Biotinylated proteins that were formerly fused to the activator domain in the in vivo interaction assay are attached to streptavidin coated SPA beads An in vitro interaction assay is carried out with the radiolabelled bait protein and the proteins attached to the beads This second assay eliminates false positives, thus determining genes encoding interacting proteins By performing the method described herein to carry out a library versus library screen, large numbers of interacting proteins will be identified Combined with sequence information from the human genome project, this type of screening method will give information about the function of unknown gene products

The solid phase for use in the invention may be chosen to provide powerful screening techniques For example, a slide or other solid support, coated with streptavidin or other suitable ligand, may be used to immobilise multiple second proteins simultaneously at spaced locations on the surface To this "array" a supply of first protein would be applied, and any interactions would be detected after washing, for example by means of a fluorescent label or a radiolabel Such an array method may also be employed for a set of different first proteins and a single second protein, or for a set of different first proteins and a set of different second proteins

It will be understood that the word "protein" as used herein in relation to the first and second proteins is considered to include peptides, polypeptides and related molecules One or both of the first and second proteins may consist of a fragment of a particular protein of interest

EXAMPLES

Example 1

The Pho4p and PhoδOp interaction

The yeast proteins Pho4p and Pho80p have previously been shown to interact in a two-hybrid system (7)

The plasmid pDM22 is a vector containing an ADH1 promoter, a Gal4 transcriptional activation domain followed by a T7 promoter, a (Hιstιdιne)₆ tag, an epitope tag, a polyhnker and an ADH1 terminator The Gal4 activation domain, the T7 promoter, tags and any gene cloned into the polyhnker are expressed from a S cerevisiae ADH1 promoter Alternatively, if the plasmid is introduced into E coli containing a T7 RNA polymerase gene or transcribed and translated in vitro, the tags and the gene will be expressed The plasmid contains markers and origins of replication for replication in both S cerevisiae and E coli It is a plasmid, which is commonly called the activator plasmid in the yeast two-hybrid system The plasmid pDM26 is a vector containing a ADH1 promoter, the LexA DNA-binding domain (10), followed by a T7 promoter, an S tag, a polyhnker and the ADH terminator The LexA DNA-binding domain, the S tag and any gene cloned into the polyhnker will be expressed in S cerevisiae from the ADH1 promoter Alternatively, if the plasmid is introduced into £ coli containing a T7 RNA polymerase gene or transcribed and translated in vitro, the tags and the gene will be expressed

The Δ156 PH04 DNA (7), was amplified by PCR using suitable primers, and cloned into the EcoRI and Htnd\\\ sites of pDM22, creating the plasmid pNWO5 The PHO80 DNA was amplified by PCR using suitable primers and cloned into the EcoRI and Hιnd\\\ sites of plasmid pDM26, resulting in the plasmid pDM27. The interaction between Pho4p and PhoδOp was shown to be functional in an ex vivo two-hybrid test. The plasmids pDM27 an pNWO5 were co-transformed into the yeast strain YM4136 which has the genotype is MATa, ura3-52, his3-200_, ade2- 101 , /ys2-801 ^, fφ1-901 ,leu2-3, 112, ga/80-538, (11), along with a lexo-lacZ reporter plasmid, pDM28 and a significant increase in beta-galactosidase activity was detected when pDM27 and pNW05 were transformed compared with the controls (see Figure 4).

The Pho80p was radiolabelled, by using an in vitro T7 transcription reaction with pDM27 as the template (Amersham

International, RPN3153). The RNA transcript was then translated using a rabbit reticulocyte in vitro translation reaction (Amersham International RPN3153) and ³⁵S-labelled methionine. A radiolabelled protein of the predicted size of the epitope tagged PhoδOp was detected by SDS-PAGE. No product was detected in an in vitro transcription/translation reaction containing no template DNA (see Figure 5).

The Pho4p protein was also produced in an in vitro transcription/translation reaction. In this case pNWO5 was used as a template in the in vitro transcription reaction, producing PH04 mRNA. The Pho4p transcript was translated in an in vitro translation reaction including tRNA-biotinylated lysine (ECL in vitro translation system, RPN 2197_, Amersham). The product was detected on a Western blot using streptavidin-HRP (Amersham) and Enhanced Chemiluminescence (ECL)_. A protein of approximately 29 kD was detected on a SDS-PAGE gel. This is the predicted size of the T7 epitope tagged-Pho4p expressed from pNW05.

The 29kD protein was incubated with streptavidin-coated SPA beads (Amersham) at 4°C. The T7 epitope tagged Pho4p was bound to the SPA beads. Attachment of the 29kD protein was confirmed by precipitating the SPA beads for one minute in a microfuge_, and washing in 1 ml of PBS The SPA beads were resuspended in SDS PAGE loading buffer, and boiled for 5 minutes to re-elute the Pho4p from the SPA beads The eluted protein was then run on SDS PAGE, along with the in vitro transcribed and translated T7 epitope tagged Pho4p protein Also run was a control in which there was no DNA template in the in vitro transcription/translation reaction The gel was Western blotted, probed with an antibody to the T7 epitope and detected using ECL (Amersham) The T7 epitope tagged Pho4p was seen to be present on the SPA beads (see Figure 7) i o

Figure Legends

Figure 4 Illustrates the results of a 2 hybrid assay for the interacting proteins Pho4p and PhoδOp The following plasmids were transformed into the yeast strain YM4136 and the Beta galactosidase i s activity assayed by a method based on Harshman et al Column 1 , pNW05, pDM27 and pDM28, Column 2, pDM22, pDM27 and pDM28, Column 3, pNW05, pDM26 and pDM2δ, Column 4, pDM22, pDM26 and pDM28, Column 5, pNW05 and pDM27, Column 6, pDM28

Figure 5 shows ³⁵S methionine labelled protein produced by 0 in vitro transcription/ translation reactions (Amersham, RPN 2197) programmed with Lane 1 , no plasmid, Lane 2, pNW05, Lane 3, pDM27, Lane 4, pCITE-Beta galactosidase (positive control plasmid included in RPN 2197), Lane 5, pDM27 The reactions were separated on a 15% SDS-PAGE gel and the gel exposed to Hyperfilm MP for 3 days

25 Figure 6 shows a Western blot of biotinylated labelled protein produced by in vitro transcription and then translation reactions, in the presence of biotinylated tRNA-lysine (Amersham, RPN 2199) The transcription reactions were programmed with Lane 1 , pDM27, Lane 2, pNW05, Lane 3 no DNA The reactions were separated on a 15% SDS-

30 PAGE gel, Western blotted and probed with a 1 /1000 dilution of HRP labelled streptavidin followed by detection using ECL reagents (Amersham).

Figure 7 shows a Western blot of biotinylated labelled protein produced by in vitro transcription/translation reactions programmed with: Lanes 1 and 2, no DNA; Lanes 3 and 4, pNW05. The reactions in Lanes 2 and 4 were first bound to streptavidin coated SPA beads before being eluted from the SPA beads by boiling in SDS-PAGE loading buffer for 5 minutes prior to loading onto the gel. The reactions were separated on a 15% SDS-PAGE gel, Western blotted and the blot probed with a 1/1000 dilution of HRP labelled strepatavidin, followed by detection using ECL reagents (Amersham).

Example 2

A new aliquot of 35S labelled Pho80p and biotinylated Pho4p prepared as described in Example 1 was used, and the Pho4p was attached to the SPA beads, as already described. The radiolabelled PhoδOp was added, and incubated for 1 hour at 4°C, on a rotary shaker_. The beads were collected by centrifugation for 1 minute, washed in 0.8ml of PBS and finally resuspended in 0.8ml of PBS. The beads were then counted in a scintillation counter. The control reaction was carried out as in Example 1 , except replacing the PhoδOp in vitro transcribed/translated protein with an in vitro transcription and translation reaction without plasmid DNA. In the control reaction, no Pho80p protein was produced. The experiment was run in triplicate and the results are shown in Table 1_.

Example 3

A protein protein interaction assayed using the ex vivo/in vitro assay system

The interaction between the two proteins cJun and ATF2 was investigated using the ex vivo/in vitro assay system This interaction has previously been reported to occur using coimmunoprecipitation (14) In this example, the cJun/ATF2 interaction was assayed ex vivo using the two hybrid system The two proteins were then each labelled with labels capable of participating in an in vitro SPA assay, the proteins were contacted in vitro and an interaction detected

Differential expression vectors were used (WO 96/30507) to allow expression of Gal4p activation domam-cJun and LexA DNA binding domaιn-ATF2 fusion proteins ex vivo, and the expression of cJun and ATF2 in vitro Constructions of the plasmids

The construction of the plasmids was by standard procedures (15) Purification of DNA was carried out using Qiagen columns Enzymes were obtained from Amersham Oligonucleotides were synthesised by Genosys (Cambridge, UK) The bacterial host used was E coli strain pMosS/ue (Amersham) (endA1 hsdR1 ^' 7(r_M2m_M2 ^*)supE44thι-1 recA1 gyrA96 relA1 /ac(7^r' pro/r^β+/ac/ ^qZΔM 15 Tn10(Tc^R) or NM554 (F araD139 Δ(ara- leu)7696 ga!E15 galKW Δ(lac)X74 rpsL (StηhsdR2 (r _knV) mcrA mcrB1 recA13

Construction of the bait plasmids

ATF2 was cloned into the bait plasmid pDM26 The plasmid pDM26 has been described previously (WO 96/30507) and expresses in yeast a LexA DNA binding domain A kanamycin resistance gene was cloned into pDM26 by excising the kanamycin resistance gene from pUC4K (Pharmacia) with Pst\ The fragment was inserted between the βgll and Aatll sites of pDM26 The resulting plasmid was called pDM41 The ATF2 gene (a plasmid containing ATF2 was a gift from Nic Jones, ICRF, Lincolns Inn Field, London, UK) was then cloned into pDM41 (-), at a position downstream of the LexA, in a two step process Firstly, oligonucleotides (sequence AATTCCAGCTGG and GATCCCAGCTGG. the PvuW site is underlined) were synthesized, annealed and cloned between the EcoRI and SamHI sites of pDM41 (-) to create pSD7 The A TF2 gene was excised from the ATF2 plasmid using Sa/I and C/al The fragment was treated with T4 polymerase to create blunt ends and inserted into the PvuW site of pSD7, to produce the plasmid pSD12 The plasmid pSD12 therefore contained the first chimeric gene encoding, in this example, the LexA DNA binding domain fused to ATF2 The construction of the plasmid pDM27 which is a differential expression vector containing the PH04NΔ156 inserted into the multicloning site has been described previously (WO 96/30507).

Construction of the activator plasmid

The plasmid containing cJun was cloned from a human brain cDNA library which had been inserted into the activator plasmid pNW17_.

The activator plasmid is based on pDM22 (WO 96/30507). The plasmid pDM22 expresses the Gal4p transcriptional activation domain. This plasmid was modified by the addition of an F1 origin of replication. The Hind\\\ sites at the junction of the ADH1 promoter with the GAL4 activation domain and at the 5' end of the ADH1 terminator were removed by excising the Hind\\\ fragment, treating the fragment with T4 DNA polymerase and recloning the blunt-ended Hind\\\ fragment back into the Hind\\\ site of the activator plasmid to create pNW17. This plasmid therefore contained a unique Hind\\\ site in the polyhnker pNW17. A cDNA library was cloned into pNW17 as described below. The cDNA library (constructed using the Novagen directional cDNA construction kit) was cloned between the EcoRI and HindlW sites of pNW17. The two hybrid screen was carried out using a bait plasmid which expresses a LexA DNA binding domain-ATF2 hybrid fusion protein. The reporters used were a lexAo-HIS3 and a lexo-lacZ. Extraction of activator plasmids from His+ lacZ+ yeast resulted in the identification of an activator plasmid which contained the complete cJun gene. This plasmid therefore expresses in yeast a Gal4 activation domain-cJun fusion protein, and was named pACT#jun. The construction of the plasmid pNW05 which expresses a Gal4p activation domaιn-Pho4NΔl56 protein has been described elsewhere (WO 96/30507) The Pho4NΔl 56 protein is not expected to interact with ATF2 protein

Assay ex vivo - the detection of the cJun ATF2 protein protein interaction a two hybrid assay

Standard methods for the growth of Saccharomyces cerevisiae was used (16) Yeast was transformed by the lithium acetate method (17)

The plasmids pACT#jun, pSD12 and pDM28 (contains a /exo-lacZ), were cotransformed into the yeast strain YM4136 (11) Beta-galactosidase assays were carried out as described previously (WO 96/30507) The beta-galactosidase activity was significantly higher in the pACT#jun, pSD12 and pDM28 transformant compared to the yeast strains containing pNW05 (PH04Δ156 containing activator), pSD12 and pDM28 or ρACT#jun, pDM41(-) (the empty bait) and pDM2δ (see table 2) Beta galactosidase units were calculated using the following formula

Beta-galactosidase units = 1000 x OD at 420/(t x V x OD at 600)

where t is the elapsed time in minutes of the incubation, V is 0 1 ml x concentration factor, OD at 600 is the absorbance at 600nm of the culture and OD at 420 is the absorbance at 420nm of the test assay This result suggests that the cJun/ATF2 interaction was detected in an ex vivo system

In vitro labelling of cJun and ATF2 proteins

Labels were incorporated into both cJun and ATF2 during an in vitro transcription/translation reaction T7 linked in vitro transcription/translation reactions (Amersham linked transcription/translation kit, Amersham) were programmed with either pACT#jun, or pSD12 (test) or pDM27 (control) In the cJun translation reaction radiolabelled methionine (Amersham) was included to radiolabel the cJun as per the manufacturer's instructions The pSD12 and pDM27 translation reactions included 2 μl of biotinylated tRNA^lγsιne (Amersham) and cold methionine as instructed in the kit protocol, thereby labelling the ATF2 and PhoδO proteins with biotin

Assay in vitro - detection of the cJun ATF2 protein protein interaction in an SPA assay

8 μl of biotinylated ATF2 and Pho80 translation mix was incubated with 25 μl streptavidin coated SPA beads (20mg/ml, obtained from Amersham) at 4° C for 1 hour in 500 μl of binding buffer (2 4 ml 5M NaCI, 0 8 ml 1 M Tπs pH 8, 0 18 g bovine serum album in 80 ml) The beads were pelleted by centrifugation in a microfuge for 2 mins and resuspended in 250μl of binding buffer. Between 0 μl and 10 μl of the radiolabelled cJun translation was added and the mix incubated at 4° C for 1 hour in a rotary mixer. The beads were harvested as described above, washed three times with 0.7 ml of binding buffer before being resuspended in 0.7 ml of binding buffer. The suspension was then counted in a scintillation counter. The results are shown in figure 8.

References

1 ) Bartel.P , Chιen,C-T , Sternglanz,R and Fιelds,S (1993)

Biotechniques 14 920-924 5 2) Fields.S and Song, O-K (1989) Nature 340 245-246

3) Fagan,R , Flint.K J and Jones, N (1994) Cell 78 799-81 1

4) Phizicy.E and Fields, S (1995) Microbiological Reviews 59 94-123

5) Kurzchaha.T V , Wiedmann.M , Breter,H , Zimmermann.W , i o Bauschke.E and Rapoport.T A (1988) Eur J Biochem 172 663-668

6) Bunter C A and Kingston, R E (1995) Nucleic Acids Res 23 269-76

7) Jayaraman,P-S , Hirst, K and Goding.C R (1994) EMBO 2192-2199

15 8) Yavuzer.U. and Goding.C R (1995) Gene 165 93-96

9) Gramatikoff.K , Georgιev,0 and Schaffner.W (1994) Nucleic Acids Res 22 3761-2

10) Gyurιs,J , Golemis.E Chertkov H and Brent.R (1993) Cell 75 791 -803 0 1 1 ) Feilotter.H E , Hannon,G J , Ruddell.C J and Beach, D

(1994) Nucleic Acids Res 22 1502-3

12) Enckson,J W and Cenone.R A (1991 ) Biochemistry 30 71 12-71 18

13) Harshman, K D , Moye,R W and Parker,C S (1988) Cell 53 25 320-330

14) Du, W , Thanos.D and Maniatis.T (1993) Cell 74 887-898

15) Sambrook, J , Fntsch, E F and Maniatis, T (1989) Molecular Cloning ed Cold Spring Harbor Laboratory Press

16) Sherman, F (1991 ) Guide to yeast genetics and molecular 30 biology Methods in Enzymology 1944 3-21 17) Gietz, D., St.Jean, A., Woods, R.A. and Schiestl, R.H. (1992).

Nucleic Acids Research 20 14-25.

Claims

">->CLAIMS

1. A method of investigating whether two proteins are capable of interacting with one another, which method comprises: a) providing nucleic acids encoding a first protein and a second protein; b) performing an ex vivo test involving producing the first and second proteins using the nucleic acids, to indicate whether the first and second proteins may be capable of interacting with one another; c) performing a further test to indicate whether the first and second proteins are capable of interacting with one another in vitro, comprising the steps of: i) producing the first and second proteins using the nucleic acids; ii) labelling the first protein from i) with a first label capable of participating in a proximity detection assay; iii) labelling the second protein from i) with a second label capable of participating in a proximity assay with the first label; iv) contacting the labelled first protein and the labelled second protein in vitro under conditions to allow protein-protein interactions; v) detecting an interaction between the labelled first protein and the labelled second protein by means of the proximity of the first label to the second label when the first and second proteins interact.

2. A method as claimed in claim 1 , wherein the second labelled protein is immobilised on a solid phase and the first labelled protein is contacted with the immobilised second protein.

3. A method as claimed in claim 2, wherein the second label is comprised in the solid phase and the second protein is labelled by virtue of being immobilised on the solid phase.

4. A method as claimed in claim 3, wherein the first label is a radiolabel suitable for use in a scintillation proximity assay, the second label is a fluorescer and an in vitro interaction between the labelled first and second proteins is detected by means of a scintillation proximity assay.

⁵- A method as claimed in claim 1 , wherein the first and second labels are different fluorescent labels and an in vitro interaction between the two labelled proteins is detected by means of fluorescence resonance energy transfer.

6- A method as claimed in any one of claims 1 to 5, wherein the ex vivo test comprises a two-hybrid assay. i o 7 ■ A method as claimed in any one of claims 1 to 6, wherein at least one of the first and second labels, or an attachment means for one of the labels, is incorporated into the protein during a translation reaction in vitro. δ. A method as claimed in any one of claims 1 to 7, wherein

15 steps i) to v) are all performed in vitro.

9- A method as claimed in any one of claims 1 to δ, wherein at least one of the nucleic acids is from a library of nucleic acids.

10- A method as claimed in any one of claims 1 to 9, wherein steps iv) and v) are performed as a homogeneous assay.