NO328908B1 - In vitro use of a polypeptide capable of inhibiting the interaction between presenilines and APP. - Google Patents

Abstract

Novel polypeptides capable of at least partially inhibiting the interaction between presenilin 1 or presenilin 2 on the one hand and the precursor of amyloid peptide and / or β-amyloid peptide on the other hand. In vitro tests for the detection of molecules small molecules capable of inhibiting the interaction. The application of the invention in the pharmaceutical field is described.

Description

The present invention relates to the in vitro use of a more precisely defined polypeptide to inhibit the interaction between presenilins and amyloid precursor peptide (APP).

The amyloid peptide Aβ of 37 to 42 amino acids is the main protein component of the senile plaque characteristic of Alzheimer's disease. This peptide is produced by cleavage of its precursor, amyloid precursor peptide (APP). Mutations in the gene for APP are responsible for certain early-onset familial forms of Alzheimer's disease. However, the majority of these forms are associated with the presence of mutations in two genes, presenilin PSI (initially named Sl82) and PS2 (initially STM2), recently identified by positional cloning (Hardy, 1997). These forms are dominant and all of their anomalies produce missense mutations except for one which causes the deletion of an exon. The presenilins are hydrophobic membrane proteins with a molecular mass of around 45-50 kDa and show 67% identity between them. They are homologous to two proteins of C. elegans, SPE4 and Sel-12, which are indirectly involved respectively in the intracellular transport and in the signaling of Notch receptors. However, the physiological function of the presenilins is still unknown. The association of two such related proteins with Alzheimer's disease seems to indicate that the presenilins act on an essential physiological reaction pathway in the etiology of this pathology.

The protein PSI comprises 467 amino acids and PS2 comprises 448. Both have the structure of a membrane protein with 6 to 8 potential transmembrane domains. Each of the presenilins undergoes in vivo a precise proteolytic cleavage resulting in two fragments, generally called the N(amino)- and C(carboxy)-terminal fragments (Thinakaran et al, 1996). This cleavage has been mapped between residues 291 and 299 of PSI (Podlisny et al., 1997) and a homologous region of PS2. The term "N-terminal (N-ter) fragment" generally refers to the fragment from position 1 to about 291 of PSI, and by the C-terminal fragment is generally meant the remainder. Although the exact topology of the presenilins in the lipid membranes has not been clearly established, it has been proposed that their N- and C-terminal [RCl] ends as well as the large hydrophilic loop are present in the cytosolic compartment (Doan et al., 1995 , see schematic diagram 1).

It has now been demonstrated that mutated forms of presenilins induce an increase in the production of the long amyloid peptide Ap 1-42 in relation to Ap 1-40 as much in carrier patients (Scheuner et al., 1996) as in transfected cells (Borchelt et al., 1996) or transgenic mice (Duff et al., 1996). The amyloid peptide AP that forms the senile plaque, a lesion characteristic of the pathology and its various forms, is derived from the catabolism of the amyloid precursor protein, APP. In particular, two essential forms of the amyloid peptide have been described, one with 40 residues, A(340), and the other with two additional residues at the carboxyl terminus, A042. In vitro, the peptide Aβ exhibits strong aggregation properties which are increased for the form Aβ42 and the latter seems effective to form the first detectable aggregates in the pathology. Furthermore, the form AP42 is specifically produced after cranial trauma in humans, which constitutes one of the best established environmental risks for Alzheimer's disease. Furthermore, the early genetic forms of the disease associated with mutations both on APP (of which the are six) and now on the presenilins 1 and 2, all contribute to an increase in the Ap42:ap40 ratio. All these factors seem to designate Ap42 as the key link for the pathology both in the genetic forms and the sporadic forms of the disease and a clarification of their formation mechanism considered a fundamental question.

In this regard, formation of a complex, in the same cellular envelope, between the precursor of the β-amyloid peptide and PSI or PS2 has been reported (Weidemann et al., 1997, Xia et al., 1997), but it is not known exact nature of the events responsible for the production of the P-amyloid peptide and no connection has so far been able to establish between the possible role of these complexes and the production of the P-amyloid peptide AP42. However, it is nevertheless worth noting that the peptide AP42, but not AP40, appears to be localized in the endoplasmic reticulum of neuronal cells (Hartmann et al., 1997).

Corresponding to what was reported by Weidemann et al, it is in the Journal of Molecular Neuroscience, 1997, vol. 9, pages 49-54 performed a study where different fragments of PSI and PS2 were coexpressed with APP in a yeast two-hybrid system to show a direct interaction between the presenilins and APP.

The present invention is the result of the identification and characterization of special regions of presenilin 1 (PSI) and presenilin 2 (PS2) as well as special regions in the precursor of the p-amyloid peptide (APP) which are involved in the formation of the complexes APP/PS1 and APP/PS2.

The present invention is particularly the result of the demonstration of the ability of the N-terminal hydrophilic region (amino acids 1-87) of PS2 to recognize different domains of APP. It further derives from the demonstration of similar properties for the N-terminal region of PSI (fragment 1-213). It is further a result of the demonstration of the ability of polypeptides derived from the above-defined regions of the presenilins to inhibit the formation of complexes between APP and the presenilins.

The present invention is further the result of the detection of the unexpected special cellular localization of the interaction regions in relation to the lipid membrane. The invention is further based on the fact that these interactions can take place not only in the membrane, but also in endoplasmic reticulum spaces and in the extracellular space. This is unexpected since the N-terminal region of PS (implicated in the interaction) is generally considered to be located in the cytoplasm under standard conditions.

A first aspect of the invention thus relates to the in vitro use of a polypeptide to inhibit the interaction between presenilins and APP, where said peptide consists of all or part of the sequence 1-87 of PS2 and possibly a signal sequence.

A second aspect of the invention relates to the in vitro use of a polypeptide to inhibit the interaction between presenilins and APP, where said peptide consists of all or part of the sequence 1-213 of PSI and possibly a signal sequence.

It is shown in the examples in the present application that the inhibition of this interaction with the aforementioned polypeptides leads to a reduction of the production of intracellular amyloid peptide ApÅ inhibiting this interaction and thus inhibiting the production of Ap 1.42, as a result, represents a therapeutic target in terms of diseases which implicates this form of amyloid peptide.

In a first embodiment according to the invention, the polypeptide contains a signal sequence. Said signal sequence is preferably selected from the signal peptide sequence of IgkB, the signal peptide of APP, the signal peptides of the muscular and central acetylcholine nicotinic receptor subunits.

The characterization of the interaction domains of APP and the presenilins and the detection of the different cellular localizations for these interactions allow aiming at the production of new polypeptides that can be used pharmaceuticals.

A third aspect of the present invention relates to the in vitro use of a polypeptide, as defined in said first or second aspect of the invention, in a method for detecting or isolating compounds that can inhibit the interaction between either i) peptide Ap 1.42 and the N-terminal end of presenilins; or ii) APP and the N-terminal end of presenilins, where said method comprises at least one step of detecting inhibition of the interaction either between peptide Ap 1.42 or the P-amyloid precursor peptide and the N-terminal end of a presenilin. In a first embodiment according to the third aspect of the invention, said method further comprises at least one step of labeling the presenilins and/or APP or the amyloid-P 1-42 peptide.

In a second embodiment according to the third aspect of the invention, the method is characterized by the fact that the following steps are carried out: - the peptide Ap 1.42 is absorbed at the front of a nitrocellulose membrane during incubation; - a bacterial extract containing the N-terminal end of a presenilin is then added for incubation with the molecule or a mixture containing different molecules to be tested; - loss of interaction between presenilin and the peptide Ap 1.42 on the nitrocellulose filter is detected using presenilin marker proteins.

The marker proteins used are preferably

a) the fixation protein of S-tag, linked to alkaline phosphatase or to a fluorescent chromophore or

b) an anti-PSNT antibody, ie directed against the N-terminal end of a presenilin.

In a third embodiment according to the third aspect of the invention comprises

the procedure following steps:

- the peptide Ap 1 ^2 is absorbed at the front of a nitrocellulose membrane during incubation; - a bacterial extract containing the N-terminal end of a presenilin is then added for incubation with the molecule or a mixture containing different molecules to be tested; - loss of interaction between the presenilin and the peptide Ap 1.42 on the nitrocellulose filter is detected using presenilin marker proteins, in which one of the marker proteins is the fixation protein of S-tag linked to alkaline phosphatase.

In a fourth embodiment according to the third aspect of the invention, the method comprises the following steps: - the peptide AP M2 is incubated in advance on a plate; - the N-terminal end of a purified recombinant presenilin is then added with the molecule or a mixture of different molecules to be tested, for incubation; - after washing, detection of the interaction of presenilin with peptide Ap \ ai on the plate using presenilin marker proteins, where loss of interaction between presenilins and the precursor of the p-amyloid peptide and/or the p-amyloid peptide is detected after visualization with a colorimetric substrate, by spectrophotometry.

In the special case where the fixation protein of S-tag linked to alkaline phosphatase is used as a marker protein, the signal is detected after detection with a colorimetric substrate, at 450 nm.

In a fifth embodiment according to the third aspect of the invention, the method comprises the following steps: - the peptide Ap 142 is incubated in advance on a plate; - the N-terminal end of a purified recombinant presenilin is then added with the molecule or a mixture of different molecules to be tested, for incubation; - after washing, detection of the interaction of presenilin with peptide Ap1.42 on the plate using presenilin marker proteins, in which one of the marker proteins is the fixation protein of S-tag linked to alkaline phosphatase and loss of interaction between presenilins and the precursor of the P-amyloid peptide and/or the P-amyloid peptide is detected after visualization with a colorimetric substrate at 450 nm, by spectrophotometry.

In a fifth embodiment according to the third aspect of the invention, the method comprises the following steps: - a molecule or a mixture containing different molecules is brought into contact with the peptide Ap 142 synthesized with a biotin and an arm of 3P-alanines or of 3 lysines at the N- terminal end; - the above reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore; - a streptavidin linked to a second fluorophore, capable of being excited at the emission wavelength of the first fluorophore, is added so that it is favored by a

transfer of fluorescence if the two fluorophores are close to each other; and

- inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore.

According to a particular embodiment, the second fluorophore XL665 which is allophyco-cyanine is chemically cross-linked to increase its fluorescence at 665 nm (CisBiointernational). The absence of the interaction is thus detected by fluorescence at the emission wavelength of the first fluorophore and by reduction of the signal of XL665 whose emission wavelength is at 665 nm.

This method is very advantageous because it allows directly detecting the interaction between the presenilins and APP and/or the AP peptide in liquid and homogeneous phase and, as a consequence, detecting the molecules that inhibit this interaction. Thus, the method is based on the transfer of the fluorescence between two fluorophores if the two chromophores are in physical proximity (as in the case of interaction with Ap^2 and the recombinant protein).

According to a preferred variant of the process, the first fluorophore Europium cryptate, carried by the labeled recombinant protein (excited at 337 nm) and which reacts with streptavidin-XL665 is fixed on peptide biot-Ap and in particular on peptide biot-Ap^ -The absence of fluorescence at 665 nm and increase of fluorescence at 620 nm which is characteristic of Europium cryptate, indicates an inhibition of the interaction between the presenilins or their N-terminal ends and APP and/or the peptide Aβ due to the searched molecules.

According to a variant of the process, the molecule or the mixture containing the different molecules can be brought into contact first with the N-terminal end of a labeled, purified presenilin by means of Europium cryptate (PSNT-K) and then with the peptide AP 1.40 or AP 1.42 which carries a biotin and an arm of 3β-alanines (or 3 lysines) at the N-terminal end. The detection of the new molecules capable of modulating or at least partially inhibiting the interaction between presenilin 1 or presenilin 2 and the precursor of the β-amyloid peptide and/or the β-amyloid peptide also occurs after association of the labeled streptavidin with XL665, by spectrofluorometry according to the process above and in particular by reading the fluorescence at 665 nm.

In a sixth embodiment according to the third aspect of the invention, the method is characterized by the following steps being carried out: - a molecule or a mixture containing different molecules is brought into contact with the peptide Ap 1.42 synthesized with a biotin and an arm of 3P-alanines or of 3 lysines at the N-terminal end; - the above-mentioned reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore, where said fluorophore is europium cryptate; - a streptavidin linked to a second fluorophore, which is XL665, which is able to be excited at the emission wavelength of the first fluorophore, is added so that it is favored by a transfer of fluorescence if the two fluorophores are close to each other; and - inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore.

In a seventh embodiment according to the third aspect of the invention, the method is characterized by the following steps being carried out: - a molecule or a mixture containing different molecules is brought into contact with the peptide APm2 synthesized with a biotin and an arm of 3p-alanines or of 3

lysines at the N-terminal end;

- the above reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore, where said fluorophore is europium

cryptate;

- a streptavidin linked to a second fluorophore, which is XL665, capable of being excited at the emission wavelength of the first fluorophore, is added so that it favors a transfer of fluorescence if the two fluorophores are close

each other; and

- inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore at 665nm.

In an eighth embodiment according to the third aspect of the invention, the method is characterized by the following steps being carried out: - a mixture of a) cell lysates containing the N-terminal end of a presenilin; b) cell lysates containing APP, lysates obtained from cells infected with virus;

and c) the molecule or mixture containing different molecules which shall

being tested; brought into contact

- the proteins that are solubilized and correspond to the presenilins or APP or the peptide

Aβ is co-immunoprecipitated using appropriate antibodies; and

- the loss of co-immunoprecipitation of the presenilins and APP is visualized by western blot with marker antibodies indicating that the tested molecules have the desired inhibitory property.

It is thus possible, based on the polypeptide parts described in the present application, to realize molecules which inhibit at least partially the interaction between presenilin 1 or presenilin 2 and the precursor of the P-amyloid peptide and/or the P-amyloid peptide, and which not exclusively is peptidic and compatible with a pharmaceutical application.

In this connection, the invention concerns the use of polypeptides as described above for the isolation of non-peptidic molecules or exclusively peptidic molecules, with pharmacological activity, by determining structural elements of these polypeptides that are important for their activity and reproduction of these elements with non-peptidic structures or not exclusively peptidic ones. The isolated molecules can be included in pharmaceutical preparations.

The polypeptides used according to the invention can be produced by growing a cell containing a nucleotide sequence that codes for the polypeptide under expression conditions for the sequence, and then recovering the produced polypeptide. In this case, the part that codes for the polypeptide is generally placed under the control of signals that allow its expression in a host cell. The choice of these signals (promoters, terminators, sequence leader for the secretion, and so on) can vary as a function of the host cell used. Furthermore, the nucleotide sequences can be part of a vector that can be replicated autonomously or integratively. More specifically, vectors with autonomous replication are prepared by using sequences for autonomous replication in the chosen host. If it concerns integrative vectors, these can be produced, for example, by using sequences that are homologous to certain areas of the genome of the host to allow integration of the vector by homologous recombination.

The host cells that can be used for the production of the peptides used in a recombinant manner can be both eukaryotic and prokaryotic hosts. Suitable eukaryotic hosts include animal cells, yeast or fungi. When it comes to yeast, yeast of the genus Saccharomyces, Kluyveromyces, Pichia, Schwanniomyces or Hansenula can be mentioned in particular. If it concerns animal cells, one can mention the cells COS, CHO, C127, human neuroblastomas, and so on. Among the fungi, one can particularly mention Aspergillus ssp. or Trichoderma ssp. The bacteria E.coli, Bacillus or Streptomyces are preferred as prokaryotic hosts.

The host cells are preferably recombinant yeast strains.

Said nucleotide sequence can be used for the construction of an expression cassette which can in turn be used in an expression vector. In particular, the expression cassette serves for the production of the polypeptide used in the present invention.

The polypeptides used in the invention can be obtained by expression in a cellular host of a nucleotide sequence as described above, optionally incorporated into recombinant

DNA, using techniques known per se, or by a combination of such techniques.

Preferably, the nucleic sequences form part of a vector that is used to induce the expression of the polypeptides in vivo, ex vivo and/or in vitro. The vector used can come from different origins, but must be able to transform animal cells and preferably human nerve cells. It can be a viral, a non-viral or a plasmid vector. Preferably, a viral vector is used which may originate from adenoviruses, retroviruses, adeno-associated viruses (AAV), herpesviruses, cytomegalovirus (CMV), vaccinia virus, and so on. Vectors derived from adenoviruses, retroviruses or AAV incorporating heterologous nucleic acid sequences are described in the literature [Akli et al., "Nature Genetics", 3 (1993) 224; Stratford-Perricaudet et al., "Human Gene Therapy", 1 (1990), 241; EP 185 573; Levrero et al., "Gene", 101 (1991) 195; Le Gal la Salle et al., Science, 259 (1993) 988; Roemer and Friedmann, "Eur. J. Biochem.", 208 (1992) 211; Dobson et al., "Neutron", 5 (1990) 353; Chiocca et al., "New Biol.", 2 (1990) 739; Miyanohara et al., "New Biol.", 4 (1992) 238; WO 91/18088].

A nucleic acid as defined above and which codes for a polypeptide used according to the invention can preferably be inserted into the genome of a recombinant virus.

Preferably, the recombinant virus is a defective virus. The term "defective virus" is intended to denote a virus capable of replicating in the target cell. In general, the genome of defective vims is deprived of at least the sequences necessary for replication of the vims in the infected cell. These regions may be eliminated (in whole or in part), rendered non-functional or replaced with other sequences. Preferably, however, the defective vims preserves the sequences in its genome that are necessary for encapsidation of the viral particles.

It is particularly advantageous to use nucleic acids that code for the proteins used according to the invention in incorporated form in an adenovirus, an AAV or a defective, recombinant retrovirus. Preferably, it is an adenovirus.

There are different serotypes of adenoviruses whose structure and properties vary somewhat. Among these serotypes, it is preferred to use the human adenoviruses of type 2 or 5 (Ad2 or Ad5) or adenoviruses of animal origin, see WO 94/26914. Among the adenoviruses of animal origin that can be used, mention may be made of adenoviruses of canine, bovine, murine (for example Mavl, see Beard et al., "Virology", 75 (1990) 81), ovine, porcine, avian or also simian (for example SAV) origin. Preferably, the adenovirus of animal origin is a canine adenovirus and most preferably an adenovirus CAV2 (eg strain manhattan or A26/61 (ATCC VR-800)). Adenoviruses of human or canine origin or of mixed origin are preferably used. Preferably, at least the region E1 is non-functional in the genome of the adenovirus. The viral gene in question can be made non-functional in any known way and in particular by total suppression, substitution, partial deletion or addition of one or more bases in the gene or genes in question. Other areas can likewise be modified and in particular this area applies to E2 (WO 95/02697), E2 (WO 94/28938), E4 (WO 94/28152, WO 94/12649, WO 95/02697) and L5 (WO 95/02697 ). Preferably, the adenovirus comprises a deletion in the E1 and E4 regions. Alternatively, it comprises a deletion in the area E1 in which area E4 and the coding sequence are inserted. In the viruses, the deletion in the region E1 preferably runs from nucleotides 455 to 3329 in the sequence of adenovirus Ad5. Alternatively, the exogenous nucleic acid sequence is inserted into the deletion in the E1 region.

The defective recombinant vims can be produced by homologous recombination between a defective vims and a plasmid carrying, inter alia, the nucleotide sequence as defined above (Levrero et al., "Gene", 101 (1991) 195; Graham, "EMBO J.", 3 (12) (1984) 2917). The homologous recombination occurs after cotransfection of the vims and the plasmid into a suitable cell line. The cell line used must preferably (i) be transformable with these elements and (ii) include the sequences capable of complementing the genome part of the defective vims, preferably in integrated form to avoid the risk of recombination. As an example of a line that can be used for the production of defective, recombinant adenoviruses, one can mention the human embryonic kidney line 293 (Graham et al., "J. Gen. Virol.", 36 (1977) 59) which, in particular, is integrated in its genome, the left part of the genome contains an adenovirus Ad5 (12%). As an example of a line that can be used for the production of defective, recombinant retroviruses, the line CRIP can be mentioned in particular (Danos and Mulligan, "PNAS", 85 (1988) 6460). Finally, the vims that have been multiplied are recovered and purified according to classical molecular biology techniques.

Defective, recombinant vims can also comprise a heterologous nucleic sequence that codes for the polypeptide used in the invention.

Polyclonal or monoclonal antibodies or fragments of such antibodies can be generated in a manner known per se. In particular, the antibodies can be produced by immunizing an animal against a polypeptide whose sequence is preferably chosen from the sequences SEQ ID No. 1 or SEQ ID No. 2 or fragments called SEQ ID No. 3 with subsequent collection of blood and isolation of the antibodies. These antibodies can likewise be generated by the preparation of hybridomas according to known techniques. The antibodies or fragments of the antibodies can in particular be used to at least partially inhibit the interaction between presenilin 1 or presenilin 2 and the precursor of the P-amyloid peptide and/or the P-amyloid peptide.

The invention shall be explained in more detail by means of the attached, non-limiting examples with simultaneous reference to the figures, where:

Figure 1 shows

A) Schematic of the truncated PS2 constructs

B) Expression in COSI cells.

Figure 2 shows the interaction of the truncated forms of PS2 with APP.

Figure 3 shows

A) The interaction of the N terminus secreted by PS2 (SecPS2NT), but not the cytoplasmic form (myc-PS2NT) and the extracellular APP. B) Detection of the interaction SecPS2NT/APP, but not of myc-PS2NT/APP in extracellular medium.

Figure 4 shows the interaction of PS2 with truncated forms of APP:

A) The interaction of PS2 with the SPA4CT form of APP, but not with the cytoplasmic domain of APP (MC45F).

B) The interaction of PS2 with the formula Cl00 of APP.

Figure 5 shows the interaction of PS1DC2 with APP and its short form:

A) The interaction of PSI with the full form of APP and the truncated form

SPA4CT.

B) Interaction of the truncated form of PSI (PS1AC2) with APP.

Figure 6 shows the interaction of PSI and of APP in insect cells:

A) Anti-histidine immunoprecipitate (label on PS 1), detection of APP.

B) Anti-PS 1 immunoprecipitate, detection APP.

C) Inverse anti-APP immunoprecipitates, detection PSI.

Figure 7 shows the interaction of the secreted form of PS2NT with the Aβ peptide in the extracellular medium. Figure 8 shows reconstitution of the interaction Ap/PS2Nt in vitro on nitrocellulose filter:

A) Dose dependence as a function of the concentration of PS2Nt.

B) Dose dependence as a function of the concentration of Ap.

Figure 9 shows the interaction in vitro of the complete forms of PSI and APP:

A) Anti-histidine immunoprecipitates.

B) Anti-P S1 immunoprecipitates.

Figure 10 shows the displacement of the N-terminal end of PS2 which is preceded by a signal peptide (SecPS2NT) for interaction between PSI and the fragment SPA4CT. Figure 11 shows the interaction of PS2NT with the peptide Ap 1-42 in vitro, detected at 96

wells plate test (ELISA).

Figure 12 shows the interaction of PS2NT with the peptide AP142, detected by the HTRF fluorescence transfer test (homogeneous time-resolved fluorescence). Figure 13 shows the blocking of the interaction APP/PS1 with SecPS2NT carried out to inhibit the production of the intracellular APM2 amyloid peptide: A) Demonstration that the blocking of the interaction APP/PS1 with SecPS2NT leads to the inhibition of the production of the intracellular amyloid APi42 peptide. B) Control of the expression of SecPS2NT does not influence the expression of different transgenes. Figure 14: Detection of the interaction of PS2 with endogenous APP of COS cells by pharmacologically treated with lactacystin. Figure 15: Interaction of PS2 and PS2NT with another region of APP, different from

the peptide Ap.

Material and methods

A. MATERIALS

1. Constructs expressing the presenilins

The acquisition of the expression vector (host vector, pcDNA3, InVitrogen) in mammalian cells of the human proteins PSI and PS2 has been described previously (Pradier et al., 1996). Several successive deletions at the C-term end of PS2 have been generated (see Figure IA). The numbering takes place from the initiation codon of PS2 as position 1.

The restriction fragment HindIII of the vector PS2 (from the 5' non-coding position -55 to the internal site at position 1080) has been purified and ligated with the vector pcDNA3, linearized with HindIII and treated with alkaline phosphatase. The truncated PS2 produced in this way (PS2AC1) runs from the N-terminal end to residue 361 plus 7 residues added at the 3' end, thus comprising the first six transmembrane domains and a large part of the hydrophilic loop (Figure IA ).

PS2AC2 has been constructed by digesting plasmid PS2 with Pstl (internal site at position 679) and ligation with the corresponding fragment Pstl to the non-coding 3' part of the weight PS2. The truncated protein PS2AC2 runs from position 1 to residue 228 of PS2 plus 18 additional residues provided by the 3' end. It includes the first four transmembrane domains of PS2.

A similar construction has been carried out for PS2 carrying the mutation N141I, PS2AC2<*>. The restriction fragment HindIII (-55)/MscI(590) of the contraction PS2AC2 has then been cloned into the vector pcDNA3 treated with HindIII and thus the ApaI end has been given a fresh tip. This construct PS2AC3 extends from the N-term end of residue 198 plus 2 further residues and thus comprises the first three transmembrane domains of PS2. ;The restriction fragment of PS2AC2, HindIII (-55/NcoI(504) made fresh at the NcoI end by treatment with the Klenow fragment of DNA polymerase is recloned into the same vector pcDNA3 HindIII//Apal with fresh tip) to construct PS2AC4 extending all the way to residue 158 of PS2 plus 3 further residues. ;The construction of the hydrophilic, N-terminal end of PS2 has been achieved by amplifying the sequence PS2 with oligonucleotides ext5': (which includes the original ATG in bold and introduces an underlined restriction site EcoRI), and ext3': ;(which introduces a stop codon after residue 90 of PS2 as well as an underlined restriction site Xhol). After cloning in weight pCRII by the TA cloning method (InVitrogen), the conformity of the fragment is verified by PCR by sequencing. This fragment (EcoRI/XhoI) is then introduced into a vector pcDNA3 in phase with a sequence corresponding to the epitope very much at its N-terminal end: mycPS2Nter. ;In order not to prejudice the topology of PS2, the same fragment Nter is recloned in the vector pSectagB in phase with the sequence of the peptide signal of IgkB to direct the secretion of the protein pS2Nter: SecPS2-Nter. The C-terminus of PS2 is similarly constructed using the restriction fragment HindIII (1080)/PstI (at the non-coding 3'), recloned into the vector pSecTagB HindIII/Pstl, SecPS2Cter running from residue 361 to the C-terminus end, or in the vector pcDNA3myc in phase with the myc epitope. Furthermore, a truncated construction of PSI has been achieved. The vector pcDNA3-PS1 is digested with Pflml (site at position 636 of the nucleic sequence encoding PSI) and XhoI in the non-coding sequence 3' of PSI. These sites have been transformed into pointed ends by treatment with polymerase DNA T4. The vector fragment, purified on agarose gel, was religated onto itself to give an expression vector for a truncated PSI extending from the N-ter of PSI to residue Ile213 (after the 5th transmembrane domain) plus 12 additional residues. This construction corresponds to the germ AC2 and is called PSI AC2. ;2. Construction expressing APP ;2.1. APP constructs; The various complete APP constructs (isoform 695) and SPA4CT (the last 100 residues of APP (amino acid 597-695) preceded by a signal peptide for insertion into membranes) have previously been mentioned (Dyrks et al., 1993) . The vectors, for expression of Cl00 and the cytoplasmic domain of APP, have been obtained in the following way: cDNAs have been obtained by enzymatic amplification of DNA (PCT (polymerase chain reaction)) using the following oligonucleotides as synthesis probes: for C100: the oligonucleotides 8172 to 8181; for the cytoplasmic region of APP: oligonucleotides 8171 to 8181. ;contains ;- a recognition site for the restriction enzyme Bglll (underlined) ;- the sequence coding for amino acids 597-602 of APP (in bold) [numbering of APP of 695 amino acids]. ;contains ;- a recognition site for the restriction enzyme Noti (underlined) ;- the complementary sequence to the sequence coding for amino acids 691-695 of ;APP (in bold) [APP numbering of 695 amino acids]. ;- the sequence complementary to the sequence Asp-Tyr-Asp-Asp-Asp-Asp-Lys corresponding to the epitope FLAG (in italics). ;contains ;- a recognition site for the restriction enzyme Bglll (underlined) ;- the sequence coding for amino acids 650-655 of APP (in bold) [APP numbering of 695 amino acids]. The products from the enzymatic amplification of DNA are cloned into the vector pCRII. The nucleotide sequence is verified by the method with specific terminators of DNA. The cDNAs are then introduced by ligation into the expression plasmid derived from plasmid pSV2 and, in the same reading frame, containing a MYC epitope. ;2.2. Construction of soluble forms of APP: α-sAPP and fi-sAPP cDNAs corresponding to the secreted forms of APP ending at the α- and β-cleavage sites have been obtained by PCR. ;Oligonucleotide 1: ;;- a stop codon (underlined inverse complementary sequence) after position 1788 of ;APP corresponding to the p-cleavage site, and ;- a restriction site ClaI. ;Oligonucleotide 2: ;;- a stop codon (underlined inverse co-complementary sequence) after position 1836 of ;APP corresponding to the α-cleavage site, and ;- a restriction site ClaI. ;Oligonucleotide 3: ;;common to the two forms, corresponds to the region 1583 to 1600 of APP including the internal restriction site of APP SacI (underlined). The cDNA of APP was amplified by PCR using the pairs oligo-3-oligo-1 and oligo-3-oligo-2 for β-sAPP and α-sAPP, respectively. The amplified products were subcloned as before into pCRII and the sequences verified by sequencing. For each, the restriction fragment Sac1-Cl1 was purified and recloned into the expression vector APP (see above) which is also controlled by Sac1-Cl1 to replace the C-terminal portion of APP with the C-terminal fragments of P-sAPP and α-sAPP respectively and reconstitute the complete the proteins. ; 3. Baculovirus constructs The production of a baculovirus transfer vector encoding a human PSI has been achieved from a mammalian cell expression vector (Pradier et al., 1996). The cDNA encoding the protein PSI was extracted by digestion with the restriction enzymes XhoI and Noti and then cloned into the transfer plasmid pAcHTLB (6 histidine fusion protein) and pAcSG2 (native protein). The production of recombinant baculoviruses occurs according to the supplier's protocol (Pharmingen) and consists of cotransfecting 2 x 10<6 >insect cells (sf9) with 1 µg of the transfer plasmid containing the gene of interest and 0.5 µg of viral DNA (Baculogold). After 5 days at 27°C, the cells are scraped and then centrifuged and the supernatant is used as viral stock for amplification and determination of the viral titer, the expression of the protein being visualized by western blot on the cellular pad. ;The production of baculovirus expressing human APP (695) has been described previously (Essalmani et al., 1996). ;For the study of the expression of PSI and APP, sf9 cells are co-infected to an M.O.I. of 2 with baculoviruses expressing human APP (695), the human protein presenilin 1 or PSI with a 6 histidine tag at the N-terminus (6HisPSl) or the control protein of Pseudomonas putrida XylE with a 6 histidine tag at the N-terminus (6HisYylE ), and then solubilized with a buffer Tris 10 mM, NaCl 130 mM, Triton X100 1%, NP 40 1%, pH 7.5. The solubilized proteins are immunoprecipitated with an anti-histidine antibody (A), antiPS1 (1805) (B), or antiAPP (22C11) (C). The presence of APP or PSI is detected by western blot with antiAPP aCT43 antibody (A), 22C11 (B) or antiPS1 95/23 antibody; (C). The solubilized fractions containing APP, PSI or 6HisPS1 are mixed and then immunoprecipitated in the presence of anti-histidines antibody or antiPS1 antibody overnight at 4°C. Co-immunoprecipitation of APP is detected by western blot with the antibodies aCT43 or 22C11. ; 4. The plasmids The plasmids used according to the invention are the following: - pcDNA is a commercial plasmid (InVitrogen) which is used for cloning and expression of mammalian cells with the sequences PSI and PS2 and their truncated forms. - pCRII is a commercial plasmid (InVitrogen) which is used for cloning PCR fragments. - pSecTagB is a commercial plasmid (InVitrogen) which is used for cloning and expression in mammalian cells of cDNA to which a secretion signal (signal peptide lg k) has been added. - pSV2 is a commercial plasmid from Pharmacia that is used for cloning and expression in mammalian cells of cDNA. - pAcHTLB is a commercial plasmid from Pharmingen for introducing an epitope (His)6 into cDNAs and homologous recombination with baculoviruses. - pAcSG2 is a commercial plasmid from Pharmingen for homologous recombination with baculoviruses. - pET29a is a commercial plasmid from Novagen for expression of cDNA in bacteria. B. METHODS ;1. Transfection of cells ; The method established for COSI or CHO cells consists of using a lipofectant in a ratio of 1:8 weight:weight in relation to DNA and a synthetic peptide H1 (sequence: KTPKKAKKPKTPKKAKKP) in the same ratio for to optimize the compaction of DNA and transfection efficiency. This method is particularly based on the neutralization of the charges in the phosphates in DNA with the positive charges of the lipofectant. ;The COSI cells are grown in an incubator at 37°C, 95% humidity and 5% C02 in DMEM medium (Dulbecco's Modified Eagle's Medium) containing 4.5 g/l glucose (Gibco-BRL) supplemented with 3% L-glutamine, 1% penicillin-streptomycin and 10% fetal calf serum. ;The day before the transfection, the cells are seeded at a density of 2.5 x 106 cells per bowl of 100 mm. On the day of transfection, the cells are rinsed twice with PBS (phosphate buffer saline solution) and once with OptiMEM (patented mixture, Gibco-BRL) for a habituation of at least 15 minutes in an incubator. ;Pr. equivalent dish of 100 mm, a total of 8 µg of plasmid DNA is added to 300 µl of OptiMEM and 64 µg of peptide H1. After vortexing for 10 seconds, wait for 5 minutes and then add lipofectamine (32 µl, i.e. 64 µl), diluted in 300 µl OptiMEM, to the mixture obtained. The whole thing is once again vigorously vortexed and left for 30 minutes. 5 ml OptiMEM is added per tube and the vortexed mixture is placed on the cells (where the medium is pre-aerated). The cells are then placed in an incubator for 4 hours during which the mixture is replaced with complete medium. ;2. Lysing of the cells and determination of the proteins; The cells are usually lysed 48 hours after the transfection (usual maximum of expression). Lysis buffer containing 10 mM Tris pH 7.5, 1 mM EDTA, 1% Triton XI00, 1% NP40 and a protease inhibitor cocktail (Complete©, Boehringer-Mannheim). For each plate, after a rinse with PBS, 800 ul of cold buffer is added. The lysates are then subjected to sonication followed by stirring with a magnetic bar at 4°C overnight. A centrifugation for 30 minutes at 15,000 rpm. separate the cake from the supernatant. The soluble proteins are then determined using a BCA kit (Pierce) to normalize the subsequent experiments. ; 3. Immunoprecipitations ;Antibodies, directed against the peptide of the N-term of PS2, 95041, (Blanchard et al., 1997) and ;against the first 20 amino acids of PSI (Duff et al., 1996) are obtained from rabbits by immunization with synthetic peptides. For immunoprecipitation, 100 µg proteins are diluted in 400 µl modified RIPA (150 mM NaCl, 50 mM Tris, pH 8.0, 1% Triton X100 v/v, 1% NP40 v/v). 30 ul suspension of protein A Sepharose (0.1% w/v in solution in PBS) and 3 ul antibody are added. The suspension is mixed gently in a rotary stirrer at 4°C overnight. The protein A Sepharose complex is washed three times with 0.5 ml of modified RIPA and once with 0.5 ml of wash buffer "Wash C" (10 mM Tris, pH 7.5). ; 4. Immunoblotting ; The samples (lysates of cells) are denatured in an equal volume of storage buffer (125 mM Tris pH 6.8, 4% w/v SDS, 20% glycerol, 0.02% bromophenol blue, 50 mM dithiothreitol) at 95°C in 5 minutes. For analysis of the expression of the presenilins, the samples are denatured in the presence of 8M urea and at 37°C to avoid self-aggregation of the presenilins at 95°C. ;The samples are deposited on tris-glycine gels (Novex) with a varying percentage of acrylamide according to the molecular weight to be discriminated. A molecular weight marker is also deposited (Broad Range, BioRad). The migration takes place for around 2 hours at 100 volts constant in a buffer SDS IX final (Novex). The gel is then transferred onto a membrane of nitrocellulose or PVDF (transfer buffer IX final (Novex) with 10% methanol) for 2 hours at 150 mA constant. ;After transfer, the membrane is blocked for 2 hours at ambient temperature in 50 ml PBS-T (PBS with 0.5% Tween) containing 2% skimmed milk (Merck). The primary antibody (diluted to an optimal concentration of the order of 1:1000e to 1:5000e in PBS-T with or without 2% skimmed milk) is left overnight at 4°C. After a short rinse with PBS-T, the membrane is incubated for 45 minutes in the presence of the second antibody (anti-mouse or anti-rabbit IgG as appropriate, linked to Raifort peroxidase) diluted to 1:5000e in a buffer called "ECL " (Tris 20 mM, NaCl 150 mM, Tween 0.1%). ;The membrane is then rinsed for 4 x 15 minutes in the "ECL" buffer. It can then be determined by ECL reagent (Amersham) consisting of two buffers that are mixed immediately before use in equal volumes. Various exposures with photographic film (Hyperfilm ECL, Amersham) are carried out followed by development. ;5. Fixation in vitro of PS2NT with amyloid peptide Ap ; 5.1. Production of the recombinant protein PS2NT in bacteria ;To obtain a bacterial expression vector of PS2NT (amino acids 1 to 87), the cDNA of PS2 is amplified by PCR with the oligonucleotides: ;The resulting ;fragment is cloned into pCRII and the sequence is confirmed. This fragment is then subcloned into the vector pET29a (Novagen) in phase with the sequence of the label S-tag. The protein is produced in the bacterium BL21. After induction with IPTG for 5 hours, the bacteria are recovered by centrifugation (10 minutes at 6000 rpm) and the cell clump is dissolved in RIPA buffer (calculated volume by multiplying the OD of the culture after induction by the volume of the culture divided by 23). The bacteria are lysed by sonication and the lysate is centrifuged at 13,000 rpm. for 20 minutes at 4°C. The supernatant (total extract) is used for the fixation studies. ;The recombinant protein PS2NT was also purified by total extraction on a Nickel column (poly-His tag carried in the vector pET29a) as described by the supplier (No vågen). ;5.2. Test of fixation PS2NT/ A/ M2 on the nitrocellulose membrane; Synthetic AP peptide (in solution) is deposited on the nitrocellulose membrane (Schleicher and Schuell) by using a dot-blot apparatus of 96 wells. After deposition, the filter is blocked (against non-specific sites for fixation of proteins) with a gelatin blocking reagent (Novagen) diluted 10 in TBST. After blocking, the filter is replaced with the dot-blot apparatus and the bacterial extract PS2NT is added to the wells for an incubation of 2 hours at ambient temperature. As a control, a bacterial extract containing the empty plasmid pET29 is used in the wells in duplicate. The filter is then washed once with RIPA buffer, then withdrawn from the apparatus and washed three times with PBST (15 minutes for each wash). The detection of the S-tag label then takes place as described by the supplier (Novagen) with a colorimetric substrate. Quantification of the colorimetric reaction (precipitate) is carried out by optical scanning of the filter and quantification of the intensity in each well with the program Tina 2.1 (Raytest). ;5.3. Interaction test PS2NT/ A042 in ELISA format; Synthetic peptide Ap (1-40 and 1-42) (100 ul, 2 ug/ml) is incubated overnight in 96-well plates for fixation on the plastic. The plates are rinsed twice with PBS and non-specific fixation sites are saturated by incubation with 5% (w/v) calf serum albumin of PBS. The purified recombinant protein (according to the protocol described under 5.1), PS2NT, diluted with buffer (25 mM Tris/HCl, pH 7.5, 0.5% Triton X-100, 0.5% NP40) is added and incubated for 4 hours at ambient temperature. After two rinses with PBS-Tween 0.5%, the protein PS2NT retained on the plate (in interaction with the peptide AP) is determined by incubation with fixing protein of S-Tag coupled to alkaline phosphatase as above. The detection of the signal is carried out in a spectrophotometer at 450 nm. ;5.4. Interaction test PSNR/ A/ 3J- 42 in HTRF (Homogeneous Time-Resolved Fluorescence) format; The purified protein PS2NT (product according to the protocol as described under 5.1) is labeled using the fluorophore Europium cryptate (PS2NT-K). The peptides APi_4o and are synthesized with a biotin and a spacer of 3 β-alanines (or 3 lysines) from the N-terminal end (upstream position 1 of the AP peptides) and two arginines (R) at the C-terminal end to facilitate the synthesis, the peptides biot-3K-ApORR and biot-3K-Ap42RR. ;The interaction reaction with PS2NT-K and biot-Ap40 or biot-Ap42 is carried out in a buffer 10 mM HEPES, pH 7.2, containing 150 mM NaCl, 3.4 mM EDTA and 3 mM CHAPS (detergent). The labeled protein PS2NT (final concentration 6 nM, i.e. 40 µl initial solution with 15 nM) is incubated with peptide biot-AP40 or biot-AP42 (final concentration 2 µM, i.e. 40 µl initial solution of 5 µM and 20 µl buffer) for 10 minutes, followed by the addition of XL665-labeled streptavidin (XL665 is a cross-linked allo-phycocyanin from CisBio International) to a concentration of 8 µg/ml (that is, 100 µl stock solution of 16 µg/ml) in a HEPES buffer 100 mM pH 7.0, containing 400 mM KF, 133 mM EDTA and 1 g/l BSA. The reaction is incubated either for 4 hours at ambient temperature or for 24 hours at 4°C and the plates are read under a Packard Discovery counter which measures on the one hand the emission of Europium cryptate at 620 nm after excitation at 337 nm and on the other hand the emission of XL665 at 665 nm after transfer of fluorescence at 620 nm from Europium cryptate on XL665. The formation of the complex XL665-streptavidin/biot-Ap42/PS2NT-cryptate leads to a fluorescence transfer from cryptate to XL665 which is measured by the counter at 665 nm. In the absence of the formation of an XL665-streptavidin/biot-Ap42/PS2NT cryptate complex, Europium cryptate fluoresces at 620 nm. ;EXAMPLES ;Example 1. Interaction between APP and PS2 and cartographer! of the interaction zone ;on PS2 ;This example aims to determine the interaction zone on PS2 by detecting an interaction between the area and APP. ;The interaction between the proteins APP and PS2 in mammalian cells is exemplified in Figure 2. The lysate of COS cells transfected with PS2 and APP is subjected to an immunoprecipitation with an antibody directed against the N-term of PS2 (95041, Blanchard et al., 1997). The immunoprecipitate is then analyzed by immunoblotting with an antibody against APP. APP is clearly detected in the immunoprecipitates of cells cotransfected with APP and PS2, but not in the absence of PS2 (Figure 2, lane 6 relative to lane 7) as described previously (Weidemann et al., 1997). To map the interaction between these two proteins, several truncated forms of PS2 are constructed. In order to preserve the membrane topology of PS2, which is generally determined by the N-term part of the membrane proteins, progressive truncations are produced of the C-term end of PS2 and which end after different transmembrane regions TM6 (PS2AC1), TM4 (PS2AC2), TM3 (PS2A C3) and TM2 (PS2AC4), schematic figure IA. The hydrophilic N-term end (87 residues) of PS2 is also constructed in cytoplasmic form (native sequence) or in secreted form upon introduction of the signal peptide of the Igk chain. The expression of these different forms is exemplified in Figure 1B, elicited by means of the antibody anti-PS2 (95041). The constructions that have hydrophobic areas also show bands corresponding to monomeric forms with expected molecular weights (here during the formation of close to doublets), dimeric forms and aggregates with characteristically high molecular weights of PS2 (figure IB, tracks 3-5). Especially for complete PS2, only these aggregates are detectable in this figure, while the monomer form is not detectable (track 6). The two constructs of the hydrophilic N-term of PS2: mycPS2Nt and SecPS2Nt give rise to bands with expected molecular weights (Figure 1B, lanes 1 and 2). The construct SecPS2Nt is likewise secreted into the extracellular medium (Figure 3B, lane 2), while the construct mycPS2Nt is cytoplasmic. ;The constructs are individually cotransfected with APP. The detergent-soluble fraction of the cell lysates is immunoprecipitated with the antibody directed against the N-term of PS2 and these immunoprecipitates are analyzed by immunoblot. As with complete PS2, APP is detectable in the immunoprecipitates with all the truncated forms of PS2, PS2 AC2 to PS2 AC4 (Figure 2, lanes 3-5), demonstrating interaction between APP and the forms containing the N-terminus of PS2. This interaction with APP is preserved with the construct N-term of PS2 in secreted form (Figure 2, lane 2), which shows that the change of PS2Nt in the lipidic membrane is not necessary for this interaction. In contrast, the cytoplasmic form mycPS2NT does not interact with APP (Figure 2, lane 1). ;Experiments with immunoprecipitation with an antibody anti-APP and detection with antibody N-term of PS2 allow to confirm the interaction between APP and SecPS2Nt under different experimental conditions. ;In the culture medium of cells cotransfected with APP and SecPS2Nt, an interaction between these two proteins is also shown by co-immunoprecipitation (Figure 3 A, lane 4 and Figure 3B, lane 3). The mycPS2NT form does not interact with APP in the medium (Figure 3B, lane 2). The presence of this interaction in the medium shows that complete APP/PS2Nt is relatively stable during the secretion process. ;Example 2. Interaction between APP and PS2 and mapping of the interaction zone ;on APP ;The purpose of this example is to determine the interaction zone with regard to APP and to demonstrate an interaction between the area and presenilin 2. ;For this purpose, truncated forms of APP are used for to limit the interaction zone on APP. A construct comprising the last 100 residues of APP under the possible control of a secretory peptide (SPA4CT and Cl00, Dyrks et al., 1993) and a construct comprising only the cytoplasmic domain (the last 45 residues of APP) are used and their presence detected using an antibody directed against the cytoplasmic domain of APP (aCT43, Stephens and Austen, 1996). In the cells cotransfected with PS2, an interaction of SPA4CT, but not of the cytoplasmic domain of APP with PS2, has been demonstrated (Figure 4A, compare lanes 4 and 5). The interaction of PS2 with the construct Cl00 (without secretion signal) is also demonstrated (Figure 4B, lane 7). Even by combining the cytoplasmic domain of APP with the membrane in a chimeric construct with the receptor α of IL2, no interaction with PS2 can be observed. This example shows that there is an interaction with SPA4CT (residues 597 to 695 of APP), but not the cytoplasmic domain (residues 651 to 695), thus suggesting that on APP the region Aβ (residues 597 to 637) and the rest of the transmembrane segment (up to residue 650) sufficient for interaction with PS2. ;Example 3. Interaction between PSI and APP and initial mapping ;The purpose of this example is to determine the interaction zone on PSI and to validate the interaction between this area and APP. ;In analogy to the results obtained for PS2 (examples 1 and 2), the study of the interaction between PSI and APP is realized in the same COSI cell system. After co-immunoprecipitation with an antibody directed against the last 20 amino acids from PSI (Duff et al., 1996), SPA4CT, the C-terminal fragment of APP, could be detected in the precipitates (Figure 5A, lane 4). APP also interacts with PSI. Similarly, the truncated form of PSI, PSI AC2(1-213), interacts with APP (Figure 5B, lane 4). These first values allow to aim that the areas of interaction between APP and PSI must be neighbors to those exemplified previously with PS2. To verify the validity and generality of this interaction PSI/APP, another cell system has been used in which the insect cells are infected with recombinant baculoviruses that express PSI with or without the tag His6 and APP (see Materials and Methods). A study of the cell lysates has allowed the detection of APP in the antiHis6 immunoprecipitates (for PSl-His6, Figure 6A, lane 4) or antiPSl (for PSI with or without His6, Figure 6B, lanes 4 and 5) when the cells are co-infected with the two types recombinant viruses, but not when a single protein is expressed (the corresponding lanes 1,2 and 3). In contrast, the proteins PS1-His6 and PSI are detectable (Figure 6C, lanes 4 and 5) in the anti-APP immunoprecipitates for dual infections. This experiment allows confirming the interaction in reverse with different antibodies. ;Example 4. Interaction between Aβ and PS2 in cells ;Given that the interaction region between APP and PS2 on APP implies a region that includes the amyloid peptide (from 595 to 635) and on PS2, its hydrophilic N-term domain, in this experiment shown that the latter interacts directly with the amyloid peptide AP produced by the cells. The expression of SPA4CT (corresponding to the last 100 residues of APP preceded by a signal peptide) in COS cells leads to a strong production of the amyloid peptide, partly because SPA4CT is considered the biological precursor of Ap. The COS cells are transfected with SPA4CT alone, SPA4CT and SecPS2Nt or with SecPS2Nt alone. The corresponding extracellular media are immunoprecipitated with the antibody antiPS2 and the peptide Aβ is detected with the help of the specific antibody WO2 (Nida et al. (1996), "J. Biol. Chem.", 271, 22908-914) (Figure 7). The peptide Aβ is identified only for cells cotransfected with SecPS2Nt and SPA4CT as a band of low intensity (Figure 7, lane 1), but not with the individual controls (lanes 2 and 3). Furthermore, a supplementary band of approx. 40 kDa also detected in a specific manner for the doubly transfected cells. After washing the filter and detection with the antibody PS2, it appears that only a band with the same molecular weight is likewise PS2-immunoreactive (Figure 7, lane 4). This band is also present for cells transfected with SecPS2Nt as expected. In the cells that are doubly transfected, this band thus represents an SDS complex that is stable against SecPS2Nt and Ap and can confirm the interaction between these two entities. The low difference in the mass brought by the peptide Aβ (4 kDa) explains that there is no detectable size difference between the transfected cells and only SecPS2Nt. The results of these experiments allow to draw the conclusion that the secreted form of PS2 (secPS2Nt) interacts in vitro with the peptide Aβ (residues 597-637 of APP695). ;Example 5. Reconstitution of the interaction of AP and PS2Nt in test in vitro on ;nitrocellulose membranes ;This example aims to show the reconstitution of the interaction PS2Nt-APP(AP) in vitro. ;In order to confirm the interaction between the peptide Ap and PS2Nt (N-terminal end of PS2), an in vitro fixation test was developed. A fusion protein PS2Nt with a peptide label/marker S (S-tag) at the N-terminal end is constructed and expressed in bacteria. The peptides Ap mo and Ap i-42 are deposited on nitrocellulose membranes which are incubated in the presence of a bacterial extract expressing the protein PS2Nt. The S-tag is then detected by the fixation protein of S-tag linked to alkaline phosphatase and by colorimetric reaction. The S-tag-PS2Nt protein binds well to Aβ peptides in the in vitro assay (Figure 8A). As controls, duplicates were incubated in the presence of a bacterial extract expressing nothing but the peptide S used as a non-specific fixation level of the peptide AP (forms 1-40 and 1-42). Serial dilutions of the bacterial extract allow establishing that this fixation is dose-dependent and saturable. In this experiment, the fixation seems to be more significant on APm2 than on Ap mo with, however, some variability. For example, the fixation of PS2Nt is dependent on the dose of Ap deposited on the membrane, while Ap mo and APm2 show equivalent fixation values. This example thus provides a demonstration of the reconstitution of the interaction PS2Nt-APP(AP) in vitro between synthetic Ap and PS2NT of bacterial origin. Provided that the pathological mutations of PS2 lead to an increase in the ratio APm2: AP mo in the numerous systems and that there is further physical interaction between PS2 and APP, it seems that this physical interaction may be implicated in the production of the peptide APm2-Thus the inhibition of this interaction is considered an extremely original therapeutic approach for Alzheimer's disease. ;Example 6. Interaction test Ap42/PS2NT in 96-well format (of the ;ELISA type). Figure 11; Example 5 gives results that directly show the interaction between the peptide Ap and the protein PS2NT on nitrocellulose membrane. This example is intended to confirm the information in example 5 and to describe the detection of the interaction in a 96 well format. The peptide Ap (AP40 or Ap42) is fixed by incubation on 96-well plastic plates. The plates are then incubated with recombinant protein PS2NT. After rinsing, the detection of the interaction (Ap/PS2NT) takes place with fixation of the fixation protein of S-tag in the wells and colorimetric detection. PS2NT binds in a dose-dependent manner to the peptide A|3m2 (figure 11), but not to the peptide APmo or to the inverse sequence peptide Ap40-1 nor to another amyloidogenic peptide, amylin. The amounts of the peptides APm2 or Ap mo fixed on the plates are identical as verified by immunodetection. The fixation constant of PS2NT on Ap42 is 0.18 µM. The specificity of PS2NT for the Ap42 form of the peptide relative to Ap40 is not indicated in Example 5. This example establishes that this specificity is absolutely reproducible in the present test. ;This interaction test format for AP42/PS2NT thus allows aiming at an easy test for screening molecules that inhibit this interaction. ;Example 7. Test of the interaction AP42/PS2NT in HTRF format ;In examples 5 and 6 the direct interaction between the peptide Ap42 and the protein PS2NT is detected on one side on the nitrocellulose membrane and on the other side on 96-well plates (ELISA type ). This example aims to confirm this information and describes the detection of the liquid/homogeneous phase interaction using the fluorescence transfer technique. ;The main principle of the test is described in Materials and Methods (5.4) and is based on fluorescence transfer. The recombinant protein PS2NT, labeled with Europium cryptate (PS2NT-K) interacts with the peptide biot-Ap42 as shown in figure 12 (column no. 3). This signal reduces and thus shifts the PS2NT-K/biot-Ap42 interaction by an excess of unlabeled PS2NT (Cl40 = 400 nM). The detected fluorescence signal is stable during the time period 4 hours at ambient temperature to 24 hours at 4°C (according to the chosen conditions, described in Materials and Methods). ;This interaction is dose-dependent for the peptide biot-Ap42 (linearity zone for the signal from 0 to 2.5 µM) and for PS2NT-K (linearity zone for the signal from 0 to 7.5 nM). As indicated in Figure 12, bar #5, there is no interaction of PS2NT-K with the peptide biot-AP40 carrying an additional specificity element. The specificity of PS2NT for the form Ap42 of the peptide in relation to AP40 which has not been mentioned in example 5, is confirmed in example 6 and in the present example. ;This HTRF assay for interaction between PS2NT and Ap42 (which reflects the APP/PS interaction in cells) thus allows the identification of chemical molecules that inhibit this interaction, by a high-flux screening. Example 8. The reconstitution of the interaction APP/PS1 in vitro This example aims to show that the interaction between the complete proteins APP and PSI can be recreated from different cell lysates, only mixed to demonstrate the interaction. The baculovirus expression system allows expression in large quantities of recombinant proteins and lysates of cells infected individually with three viruses are used as a source for the proteins APP, PSI and PSl-His6. The solubilized fractions containing APP, PSI or 6HisPS1 are mixed and then immunoprecipitated in the presence of an anti-histidine antibody or antiPS1 antibody overnight at 40°C. APP is clearly detected in the immunoprecipitates (Figure 9A, lane 3 and Figure 9B, lane 3), which shows that the interaction APP with PSI-His or PSI is reconstituted in vitro by incubation of the two proteins. APP alone appears to precipitate little with the antibody anti-PS1 (Figure 9B, lane 1), but not with the antibody anti-His, confirming the specificity of the interaction in the latter. These results allow the application of an interaction test in vitro for the two complete proteins APP and PSI. Example 9. SecPS2Nt blocks the interaction of APP and PSI in transfected cells It has been shown in the preceding examples that APP interacts with PSI in a manner similar to PS2 and that for this latter construct SecPS2Nt is sufficient for interaction with APP. The purpose of this example is to assess whether the fixation of SecPS2Nt on APP can block the interaction with PSI in an increased way (heterologous). In the COSI system, SPA4CT (corresponding to the last 100 residues of APP, preceded by a peptide signal) can be detected in anti-PS1 immunoprecipitates from cells expressing SPA4CT and PSlwt or PSI mutant, PSI<*> (Figure 10A, lanes 1 and 2) . When SecPS2NT is also cotransfected, the SPA4CT signal virtually disappears in the anti-PSAl immunoprecipitates (Figure 10A, lanes 3 and 4). After anti-PS1 immunoprecipitation, the supernatants (fraction not bound to protein A Sepharose) are subjected to a second immunoprecipitation with anti-PS2 antibody. SPA4CT is clearly detected in cells cotransfected with PSI and SecPS2Nt (Figure 10B, lanes 3 and 4), which shows that in these cells SecPS2Nt by its fixation displaces the fixation of SPA4CT on PSI. This experiment thus allows us to draw the conclusion that SecPS2Nt is a molecule capable not only of attaching to APP, but also of displacing the fixation of APP on PSI and probably PS2. SecPS2Nt can thus, in cells, serve to block the interaction APP with the two presenilins PSI and PS2.1 reality, the results of the mapping of the interaction PS1/APP confirm that the interaction zones that come into consideration correspond to those of PS2.

Example 10. Blocking of the interaction APP/PS1 leading to inhibition of

the production of intracellular peptide amyloid A(342.

The example above shows that the expression of SecPS2NT can block the interaction between APP and PSI or PSI mutant. This example analyzes the consequence of this inhibition on the production of the amyloid peptide, particularly the two forms Ap40 and AP42. Thus, it has previously been described in the literature that the pathological mutations of the presenilins (PSI or PS2) lead to an increase in the ratio between the long form of the peptide, AP, the form Ap42, on the form Ap40, the ratio Ap42:Ap40 (Borchelt et al., 1996 for the journal "Hardy", 1997).

SPA4CT is coexpressed with PSI wt (Figure 13A, lanes 1 and 3) or with PSI mutant (Figure 13A, lanes 2 and 4), either in the absence (lanes 1 and 2) or in the presence of SecPS2Nt (lanes 3 and 4) . The cell lysates and the conditioned cell media are analyzed with regard to the production of amyloid peptide. The forms Ap40 and AP42 are analyzed by immunoprecipitation with antibodies that specifically recognize the terminal ends of Ap40 (FCA3340) or Ap42 (FCA3542, Barlelli et al., 1997), and the immunoprecipitates are analyzed by immunoblot with an antibody that recognizes the two forms.

In the cell lysates, the expression of the PSI mutant leads to an increase in the production of Ap42 (1.5 to 2-fold) and of its multimeric forms compared to PSlwt and little variation in the amount of Ap40 (compare Figure 13 A, lanes 1 and 2, plates AP42 and AP40 cell lysates) as expected. In the presence of SecPS2NT, the level of AP42 (and multimers) is significantly reduced (Figure 13A, lanes 3 and 4), as well as with PSlwt as a PSI mutant. In the extracellular medium, the level of AP42 also appears to be reduced, but in a less significant way. There are no variations in the amount of amyloid peptide AP40 between the different conditions, which shows that the effect on AP42 is specific and is not due to any total modification of the expression level. This is confirmed by analysis of the expression of different transfected genes: SPA4CT, SecPS2NT and PSI (Figure 13B). In this example, it is thus shown that inhibition of the interaction PSI/APP with genetically dominant SecPS2NT leads to a reduction of the levels of the production of intracellular AP42, as well with PSI mutant as with PSI wt. Regarding the primordial role associated with AP42 in the development of Alzheimer's disease, this example provides a demonstration that the inhibition of the APP/PS interaction thus represents a significant therapeutic opportunity both for the genetic forms and for the sporadic forms of the disease.

Example 11. Detection of the interaction of PS2 with APP, endogenous to the COS cells by means of a pharmacological treatment

The purpose of this example is to show that the results obtained in the previous examples (these results in cells corresponding to overexpression of the two interaction partners APP and PS (PSI or PS2)) are equally valuable with overexpressed or endogenous proteins. Thus, the strong overexpression of the two proteins can lead to an artificial interaction. The detection of the interaction under these conditions does not imply that the simultaneous overexpression of the two partners is sought in this example.

The COS cells express APP endogenously, albeit at low levels. The COS cells are thus transfected with only PS2. The cell lysates have been analyzed by immunoprecipitation with antibodies directed against the peptide of the N-term of PS2 and determined by immunoblotting with the anti-APP antibody, WO2. The panel above figure 14 shows that the transfection with only PS2 does not allow detecting interaction with endogenous APP by co-immunoprecipitation (figure 14, lane 1).

Because further the interaction of APP/PS leads to the production of the amyloid peptide A042 (previous example), thus to the catabolism of APP, in the endoplasmic reticulum, the proteasome, which is the system of proteolytic degradation in this cellular compartment, may be implicated. The effect of lactacystin, a selective proteasome inhibitor, has been analyzed. After incubation of cells transfected with PS2 in the presence of lactacystin, endogenous APP in COS cells can be clearly detected in the PS2 immunoprecipitates (Figure 14, lane 5, band at 110 kDa). This interaction with endogenous APP exhibits the same characteristics as before because it can be displaced by the genetically dominant SecPS2NT (Figure 14, lanes 6 and 7) with a dose-dependent effect. Thus, track 7 shows that at a moderate dose of SecPS2NT, a band with weak intensity corresponding to endogenous APP is constantly visible. At a higher dose of SecPS2NT (lane 6), the corresponding band of endogenous APP appears weaker and shows a dose-dependent character. A weak residual APP signal is always present (band of around 110 kDa) and is due to the complex between APP and SecPS2NT which is rapidly secreted (when SecPS2NT does not have transmembrane anchoring domains) and thus does not accumulate intracellularly.

Figure 14 (middle and lower panel) shows that the treatment with lactacystin does not affect the total APP levels, cellular as well as secreted. Thus, the bands corresponding to the expression level of APP are, so to speak, constant in terms of intensity. The specificity of the action of lactacystin is thus shown on the subpopulation of APP interacting with PS2.

These results have shown that the interaction APP/PS2 can be detected with endogenous APP from COS cells if the proteasome is inhibited. Furthermore, these results show that the interaction APP/PS2 can be detected under less artificial conditions. However, this interaction is very labile. In order to achieve a more marked detection, one has resorted to either a proteolytic degradation inhibitor in the present example or to overexpression of the two partners in the previous example. This example thus shows that the results obtained in the examples above with overexpression in the cell of the two interaction partners (APP and PSI or PS2) are equally valid with non-expressed or endogenous proteins.

Example 12: PS2 interacts with a second segment of APP, different from AP This example aims to show that another segment of APP interacts with PS2.

It has previously been shown that SecPS2NT interacts in the extracellular medium with secreted forms of APP (Figure 3A, lane 4). The secreted forms of APP are released after cleavage either at the P-site (position 595) corresponding to the beginning of the peptide Ap, or at the a-site (position 612) in the same peptide. These results suggest that PS2NT also interacts with an N-term region of APP distinct from Ap. The truncated forms of APP are constructed by inserting a stop codon in the sites P (P-sAPP) and a (a-sAPP) and tested. Complete PS2 and SecPS2NT efficiently interact with α-sAPP (Figure 15, lanes 3 and 5) and β-sAPP (Figure 15, lanes 4 and 6). These results establish that a segment of APP between position 1 and 595 (and thus distinct from AP) is also capable of interacting with PS2 and PS2NT. These results further allow to confirm that the interaction PS2NT/APP can take place in the absence of anchoring to the membrane in the case of the two partners and in the luminal space (or the extracellular space) of the cell.

Literature references

- Doan et al. (1996), "Protein topology of presenilin 1", "Neuron", 17,1023-1030.

- Thinakaran et al. (1996), "Endoproteolysis of Presenilin 1 and accumulation of processed derivatives in vivo", "Neuron", 17,181-190. - Podlisny et al. (1997), "Presenilin proteins undergo heterogeneous endoproteolysis between Thr291 and Ala299 and occur as stable N- and C-terminal fragments in

normal and Alzheimer brain tissue", "Neurobiology of Disease", 3, 325-337.

- Pradier, L., Czech, C, Mercken, L., Revah, F. and Imperato, A. (1996), "Biochemical characterization of presenilins (Sl82 and STM2) proteins", "Neurobiol. Aging", 17, S137 - Scheuner et al. (1996), "Secreted amyloid b-protein similar to that in the senile plaques of Alzheimer's disease is increased in vivo by the presenilin 1 and 2 and APP mutations links to familial Alzheimer's disease", "Nature Med.", 2, 864- 870. - Dyrks, T., Dyrks, E., Monning, U., Urmoneit, B., Turner, J. and Beyreuther, K. (1993), "Generation of bA4 from the amyloid protein precursor and fragment thereof,

"FEBS Lett.", 335, 89-93.

- Weidemann, A., Paliga, K., Diirrwand, U., Czech, C, Evin, G., Masters, C. and Beyreuther, K. (1997). "Formation of stable complexes between two Alzheimer's disease gene products: Presenilin-2 and b-Amyloid precursor

protein.", "Nature Medicine", 3, 328-332.

- Blanchard, V., Czech, C, Bonici, B., Clavel, N., Gohin, M., Dalet, K., Revah, F., Pradier, L., Imperato, A. and S. Moussaoui (1997 ), "Immunohistochemical analysis of presenilin 2 expression in the mouse brain: distribution pattern and

co-localization with presenilin 1 protein", "Brain Res.", 758,209-217.

- Borchelt, D.R., Thinakaran, G., Eckman, C, Lee, M.K., Davenport, F., Ratovitsky, T., Prada, C.-M., Kim, G., Seekins, S., Yager, D. , Slunt, H.H. , Wang, R. , Seeger, M. , Levey, A.I. , Gandy, S.E. , Copeland, N.G. , Jenkin, N. , Price, D.L. , Younkin, S.G. and S. Sisodia (1996). "Familial Alzheimer's disease-linked presenilin 1 variants elevate Abl-42/1-40 ratio in vitro and in vivo", "Neuron", 17,1005-1013. - Duff, K., Eckman, C, Zehr, C, Yu, X., Prada, C.-M., Perez-tur, J., Hutton, M., Buee, L., Harigaya, Y., Yager , D., Morgan, D., Gordon, M.N., Holcomb, L., Refolo,

L., Zenk, B., Hardy, J. and S. Younkin (1996), "Increased amyloid-b42(43) in

brains of mice expressing mutant presenilin 1", "Nature", 383, 710-713.

- Hardy, J. (1997), "Amyloid, the presenilins and Alzheimer's disease", 'Trends in Neurosci.", 20,154-159. - Stephens D.J. and B.M. Austen (1996), "Metabolites of the b-amyloid precursor protein generated by b-secretase localize to the Trans-Golgi Network and late endosome

in 293 cells", "J. Neurosci. Res.", 46, 211-225.

-Essalmani, R., Guillaume, J.-M., Mercken, L., and Octave, J.-N. (1996), "Baculovirus-Infected Cells Do not Produce the Amyloid Peptide of Alzheimer's Disease from

its Precursor", "FEBS Lett.", 389,157-161.

- Barelli et al. (1997), "Characterization of new polyclonal antibodies specific for 40 and 42 amino acid-long amyloid p peptides : their use to examine the cell biology of presenilins and the immunohistochemistry of sporadic Alzheimer's disease and cerebral amyloid angiopathy cases", "Molecular Medicine" , 3, 695-707.

Claims (14)

1. In vitro use of a polypeptide to inhibit the interaction between presenilins and APP, where said peptide consists of all or part of the sequence 1-87 of PS2 and possibly a signal sequence. 2. In vitro use of a polypeptide to inhibit the interaction between presenilins and APP, where said peptide consists of all or part of the sequence 1-213 of PSI and possibly a signal sequence. 3. In vitro use of a polypeptide according to one of claims 1-2, where the polypeptide contains a signal sequence. 4. In vitro use of a polypeptide according to claim 3, where the signal sequence is selected from the signal peptide sequence of IgkB, the signal peptide of APP, the signal peptides of the muscular and central acetylcholine nicotinic receptor subunits. 5. In vitro use of a polypeptide as defined in claim 1 or 2 in a method for detecting or isolating compounds which can inhibit the interaction between either i) peptide AP M2 and the N-terminal end of presenilins; or ii) APP and the N-terminal end of presenilins, where said method comprises at least one step of detecting inhibition of the interaction either between peptide APm2 or the p-amyloid precursor peptide and the N-terminal end of a presenilin. 6. In vitro application according to claim 5, wherein said method further comprises at least one step of labeling the presenilins and/or APP or the amyloid-P 1-42 peptide. 7. In vitro application according to claim 5, wherein said method is characterized by the following steps being carried out: - the peptide Ap is absorbed at the front on a nitrocellulose membrane during incubation; - a bacterial extract containing the N-terminal end of a presenilin is then added for incubation with the molecule or a mixture containing different molecules to be tested; - loss of interaction between presenilin and the peptide AP M2 on the nitrocellulose filter is detected using presenilin marker proteins. 8. In vitro application according to claim 5, where said method comprises the following steps: - the peptide Ap\ ai is absorbed at the front of a nitrocellulose membrane during incubation; - a bacterial extract containing the N-terminal end of a presenilin is then added for incubation with the molecule or a mixture containing different molecules to be tested; - loss of interaction between presenilin and the peptide Ap m2 on the nitrocellulose filter is detected using presenilin marker proteins, in which one of the marker proteins is the fixation protein of S-tag linked to alkaline phosphatase. 9. In vitro application according to claim 5, where said method comprises the following steps: - the peptide Ap M2 is incubated at the front on a plate; - the N-terminal end of a purified recombinant presenilin is then added with the molecule or a mixture of different molecules to be tested, for incubation; - after washing, detection of the interaction of presenilin with peptide Ap \ ja on the plate using presenilin marker proteins, where loss of interaction between presenilins and the precursor of the P-amyloid peptide and/or the P-amyloid peptide is detected after visualization with a colorimetric substrate, by spectrophotometry. 10. In vitro application according to claim 5, where said method comprises the following steps: - the peptide Apink is incubated at the front of a plate; - the N-terminal end of a purified recombinant presenilin is then added with the molecule or a mixture of different molecules to be tested, for incubation; - after washing, detection of the interaction of presenilin with peptide APm2 on the plate takes place using presenilin marker proteins, in which one of the marker proteins is the fixation protein of S-tag linked to alkaline phosphatase and loss of interaction between presenilins and the precursor of the P-amyloid peptide and/ or the P-amyloid peptide is detected after visualization with a colorimetric substrate at 450 nm, by spectrophotometry. 11. In vitro application according to claim 5, in which the following steps are carried out: - a molecule or a mixture containing different molecules is brought into contact with the peptide APi^ synthesized with a biotin and an arm of 3P-alanines or of 3 lysines at the N-terminal end ; - the above reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore; - a streptavidin linked to a second fluorophore, capable of being excited at the emission wavelength of the first fluorophore, is added so that it is favored by a transfer of fluorescence if the two fluorophores are close to each other; and - inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore. 12. In vitro use according to claim 5, in which the following steps are carried out: - a molecule or a mixture containing different molecules is brought into contact with the peptide Ap^2 synthesized with a biotin and an arm of 3P-alanines or of 3 lysines at the N-terminal end; - the above-mentioned reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore, where said fluorophore is europium cryptate; - a streptavidin linked to a second fluorophore, which is XL665, which is able to be excited at the emission wavelength of the first fluorophore, is added so that it is favored by a transfer of fluorescence if the two fluorophores are close to each other; and - inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore. 13. In vitro application according to claim 5, in which the following steps are carried out: - a molecule or a mixture containing different molecules is brought into contact with the peptide Ap 1^ 2 synthesized with a biotin and an arm of 3p-alanines or of 3 lysines at the N- terminal end; - the above-mentioned reaction mixture is incubated with the N-terminal end of a purified presenilin labeled by means of a first fluorophore, where said fluorophore is europium cryptate; - a streptavidin linked to a second fluorophore, which is XL665, which is able to be excited at the emission wavelength of the first fluorophore, is added so that it is favored by a transfer of fluorescence if the two fluorophores are close to each other; and - inhibition of the interaction is detected by spectrofluorometry at the emission wavelength of the first fluorophore and/or by measuring the reduction of the signal at the emission wavelength of the second fluorophore at 665 nm. 14. In vitro use according to claim 5, in which the following steps are carried out: - a mixture of a) cell lysates containing the N-terminal end of a presenilin; b) cell lysates containing APP, lysates obtained from cells infected with virus; and c) the molecule or mixture containing different molecules to be tested; brought into contact - the proteins solubilized and corresponding to the presenilins or APP or the peptide AP are co-immunoprecipitated by means of suitable antibodies; and - the loss of co-immunoprecipitation of the presenilins and APP is visualized by western blot with marker antibodies indicating that the tested molecules have the desired inhibitory property.