The present invention relates to peptides which inhibit the binding of proteins via helix-loop-helix/leucine zipper domains (HLH-Z) to related products, to processes for their production and their use in medicine.
Overexpression of the myc gene product has been associated with a variety of tumours including pro-myelocytic leukaemia, colon carcinomas, large and small cell lung carcinoma and breast carcinoma. The Myc gene product is known to be involved in driving the cell cycle but the mechanism of action has not yet been elucidated. The Myc gene product is also known to contain a HLH-Z motif and to have the ability to bind to DNA. The sequence specific binding of the Myc gene product to DNA was first noted by Blackwell, T.K. et a_l. in 1990 (Science, 2_5_0, 1149-1151) who postulated that this binding might be responsible for some of the biological functions of the Myc gene product.
Leucine zippers and helix-loop-helix domains are two classes of peptide sequences each of which permit hetero- or homodimerisation of polypeptide chains by specific interactions, such as that of GCN4 (a leucine zipper protein) , the crystal structure of which has been described by O'Shea, E.K. et al [Science, 254, 539-544 (1991) ] and Myo D, a helix- loop-helix protein (Tapscott, S.J., Science, 24.2, 405-411 (1988)) . In Myc and Max gene products, adjacent helix-loop- helix and leucine zipper domains form a contiguous structure (HLH-Z) and a short basic region (B) , located just upstream of the HLH-Z domain is involved in DNA binding. More recently it has been found that the Myc gene product can bind, via the HLH-Z domain to the product of the Max gene [Blackwood, E.M. et al., Science, 251. 1211-1217(1991)]. The present inventors have demonstrated that Max:Max homodimerisation is possible whereas Myc:Myc homodimerisation is not possible under physiological conditions.
The present inventors have surprisingly established that it is Myc:Max heterodimerisation which leads to the oncogenic effect of Myc and that this interaction can be disrupted by peptides containing an appropriate HLH-Z domain thereby reducing or abolishing the transformation of cells into tumour cells normally associated with overexpression of Myc.
The present invention therefore provides polypeptides which specifically bind to the HLH-Z domain of either the Myc gene product or the Max gene product so as to prevent Myc: ax heterodimerisation under physiological conditions.
In a particular aspect, the polypeptides of the invention contain a HLH-Z domain and are capable of specifically binding the HLH-Z domain of either the Myc or the Max gene product. Certain of the polypeptides of the invention will be capable of specifically binding to the HLH-Z domain of the Max gene product. Others will specifically bind to HLH-Z domain of the Myc gene product; these are preferred embodiments of the invention. The sequence of the HLH-Z domains of the Myc and Max gene products have been published by Blackwood et aJL. loc. cit. Binding of a polypeptide to either of these motifs in order to inhibit the Myc:Max heterodimerisation can be tested essentially by the techniques used in Examples 1 and 2 below. The specificity of the binding between the polypeptide and the target HLH-Z domain can be tested by the same technique but with extraneous HLH-Z-containing proteins or random polypeptides also present. Preferably the polypeptide of the invention will bind only to the HLH-Z domain of the Myc or the Max gene product an will not bind to any other HLH-Z domain under physiological conditions. More preferably the polypeptide of the invention will also be incapable of homodimerisation under physiological conditions.
The polypeptides of the invention must be sufficiently long that they can adopt a suitable helical conformation for binding to the target sequence. The polypeptides must contain a sufficient number of amino acid residues to permit stable binding to the target sequence.
The polypeptide may contain a sequence capable of bindin to the helix-loop-helix domain or to the leucine zipper domain of the Myc or Max gene product. Preferably the polypeptides o the invention contain a sequence capable of binding at least the leucine zipper domain of the Myc or Max gene product and more preferably they contain a sequence capable of binding the entire HLH-Z domain of the Myc or Max gene product. In addition to binding either or both of the helix-loop-helix and leucine zipper domains, the polypeptides of the invention may advantageously contain a sequence capable of binding the B- region of the Myc or Max gene products.
In practice the polypeptides are therefore likely to contain not less than 7 amino acid residues corresponding with a complete heptad repeat of the leucine zipper domain of the Myc or Max gene product. There is no specific upper limit on the length of the polypeptides though economy and other considerations, such as the requirements imposed by likely treatment regimes, will place practical constraints on the length of the polypeptide. It is considered unlikely that the polypeptide would contain more than 500 amino acid residues and the preferred polypeptide will usually be less than 250, for instance less than 100 residues in length. The HLH-Z domains of the Myc and Max gene products are respectively 67 and 66 amino acid residues in length and particularly preferred polypeptides will therefore have a length of up to 90, for instance up to 70, 50, 40, 30 or 20 amino acid residues.
The sequence of the polypeptide may be tailored to meet requirements other than specific binding to the HLH-Z domain of the Myc or Max gene products, for instance to satisfy the needs for pharmaceutical acceptability, .in vivo half life and biodegradability. The polypeptides of the invention may also contain sequences unrelated to the function of binding to the HLH-Z domains of the Myc or Max gene products.. Preferred polypeptides intended to bind specifically to the HLH-Z domain of the Myc gene product will include the HLH-Z motif, or at least 7 contiguous amino acid residues thereof, of the Max gene product. Preferred polypeptides intended to bind specifically to the HLH-Z domain of the Max gene product will include the HLH-Z motif, or at least 7 contiguous amino acid residues thereof, of the Myc gene product. Other preferred polypeptides are analogues of these in which amino acid residues at particular positions are modified or replaced so as to enhance the affinity of the polypeptide for the target whilst retaining the ability specifically to bind the target HLH-Z domain. Preferably such modification or replacement will be at one or more residues which appear at the interface between the polypeptide and the target HLH-Z domain of the target protein.
Particularly preferred polypeptides of the invention are targeted at the HLH-Z domain of the Max gene product and comprise at least 7 amino acid residues of the HLH-Z domain of the Myc gene product or are analogues thereof. The polypeptides of the invention may be produced by conventional techniques either by expression of coding DNA sequences in a cell-free expression system or in host cells containing the necessary regulatory sequences suitably associated with the expressible sequences encoding the polypeptide and cultured under suitable conditions to ensure expression thereof. The methods for expression or synthesis of polypeptides are all well known to those skilled in the art and do not require further description here although reference may be made to Sambrook, J., Fritsch, E.F., Maniatis, T. (Eds) "Molecular Cloning" (2nd Edn) , Cold Spring Harbour Laboratory Press (1989) Cold Spring Harbour, N.Y. Alternatively, particularly for the smaller polypeptides, production by de novo synthesis by conventional techniques may be convenient. Solid or liquid phase synthetic techniques are well known to those skilled in the art. It is currently preferred to use solid phase methods such as described by Atherton, E and Sheppard, R.C. in "Solid Phase Peptide Synthesis: A Practical Approach" Rickwood, D. and Haines, B.D. (Eds) 1989 IRL Press, Oxford. The polypeptides of the invention may be used in treating diseases, especially the forms of cancer associated with overexpression of the Myc gene product. Treatment may be by administration of an effective non-toxic amount of the polypeptide by any standard route. Oral, topical and parenteral routes are particularly convenient and, particularly for systemic treatment with polypeptides likely to be degraded by passage through the gastro-intestinal tract or unlikely to be absorbed therefrom, parenteral administration, eg intravenous, intra- peritoneal, intramuscular, intradermal or subcutaneous injection or infusion, is preferred. Alternatively the polypeptide can be produced in situ by expression of suitable coding sequences of DNA or RNA administered for instance as attenuated viruses. For administration to a patient the polypeptide will be presented in a suitable pharmaceutical formulation comprising a pharmaceutically acceptable diluent or carrier, e.g. water for injection, and optionally containing accessory ingredients such as anti-oxidants, preservatives, antimicrobials, buffering agents, agents to adjust the tonicity, salt concentration or viscosity and other conventional excipients. The formulations may be presented as unit or multi-dose forms such as tablets, capsules, creams, lotions, pastes, powders and aqueous solutions, dispersions or suspensions. For injection or infusion the formulation is preferably presented as an injectable aqueous solution or suspension or as a lyophilised powder for reconstitution as an aqueous solution or suspension using water for injection, sterile water or pyrogen free water. In a particular aspect of the invention the polypeptides are targeted to the tumour cells to be treated. This may be achieved simply by injection into the tumour itself or by the use of suitable targeting strategies such as by coupling antibodies against tumour cell-surface antigens to the polypeptides or to materials encapsulating the polypeptide. Such techniques are well known in general to those skilled in the art and can readily be adapted to targeting the present polypeptides to tumour cells which overexpress the Myc gene product. The dose of the polypeptide to be administered will depend upon the age, weight, sex and condition of the patient, the size, nature and location(s) of the tumour(s) to be treated and the chosen route of administration. As a general guide each dose may be in the region of from 1 mg to 1 g, for instance 10 mg to 100 mg, preferably about 50 mg of polypeptide. Such doses may be repeated several times per day and for several days, weeks or even months in order to treat the tumours. Thus the daily dose for an average human adult of about 75kg would typically be in the range of from lmg to log, preferably about 50 mg to 2 g.
In other aspects the present invention provides the following:
(a) Nucleic acids (DNA or RNA, single or double stranded) having sequences encoding a polypeptide of the invention. Such materials may contain additional coding and/or non- coding sequences, regulatory sequences necessary to secure expression of the coding sequence, markers, ligation and splicing sites, restriction endonuclease cutting and/or recognition sites, and may be circular or linear and in the latter case may have sticky or blunt ends.
(b) Expression or cloning vectors comprising a nucleic acid as described in (a) . Such vectors may be plasmids, cosmids, viral genomic nucleic acids or yeast artificial chromosomes or other vectors known per s_e.
(c) Transformed or transfected cells containing heterologous nucleic acid according to (a) or vectors according to (b) .
(d) Processes for producing polypeptides of the invention by synthesis from precursor amino acid derivatives or by expression of a nucleic acid according to (a) .
(e) Polypeptides of the invention for use in a method of treatment practised on the human or animal body.
(f) Use of a polypeptide of the invention in the production of a medicament for use in a method of treatment practised on the human or animal body, particularly in the treatment of a tumour associated with overexpression of the Myc gene product.
(g) A pharmaceutical composition comprising a polypeptide according to the invention and a diluent or carrier therefor.
(h) A method of treatment comprising administering a polypeptide of the invention or a pharmaceutical composition comprising a polypeptide of the invention and a diluent or carrier therefor to a human or animal having a tumour associated with overexpression of the Myc gene product.
A process for screening candidate therapeutic agents intended to interfere with Myc: Max heterodimerisation, which process comprises contacting cells with the candidate therapeutic agents or, when the candidate therapeutic agent is a peptide, optionally expressing the peptide in the cells, which cells contain a reporter gene the expression of which is enhanced or reduced by agents which interfere with Myc:Max heterodimerisation, and observing the level of expression of the reporter gene by the cells.
In one aspect of this method peptides which are candidate therapeutic agents are expressed in cells whic also express (a) either (but not both) of Myc or a Max construct also containing a transcription activation domain and (b) a reporter gene regulated by the transcription activation domain of Myc or the Max construct. Preferably a library of constructs encoding candidate peptides is expressed. Candidates for further evaluation are selected on the basis of good levels of expression of the reporter gene resulting from heterodimerisation of the peptide with the expressed Myc or Max construct. Suitably the cells used do not expres Myc or Max from homologous DNA.
In another aspect of this method candidate therapeutic agents are screened for their ability to abolish expression of a reporter gene regulated by Myc: Max heterodimerisation in cells which, preferably, express Myc or Max only from heterologous DNA. This technique may be used for screening peptides expressed in the cells from heterologous DNA or for screening any candidate agent, whether or not a peptide, for instance a synthetic peptide analogue which can be administered to the cells.
Preferably the cells used in this screening technique are yeast cells and the screening is conducted by conventional methods such as those of Examples 1 and 2. The present invention will now be illustrated by reference to the figures of the drawings in which
Fig. 1 shows in diagrammatic form the various domains of the Myc and Max gene products and the Myc 92 polypeptide of the invention.
Fig. 2 shows the nucleotide sequence of double stranded DNA encoding the Myc 92 polypeptide of the invention and the amino acid residue sequence of Myc 92 polypeptide.
Fig. 3 is a schematic representation of human c-Myc, Max and Max deletion derivatives and of the CACGTG-CYCl-LacZ reporter gene used in Example 2. Fig. 4 shows levels of activation by Myc and Max in yeast cells.
Fig. 5 gives a schematic representation of certain fusion proteins and shows levels of activation in yeast cells.
Fig. 6 is a diagram illustrating Myc and Max interactions and function.
In Fig. 1 the basic (B) and helix-loop-helix (HLH) and leucine zipper (LZ) domains of the Myc and Max gene products are shown together with amino acid residue numbers for the boundaries between the domains. The Myc and Max gene products are shown aligned with the Myc92 polypeptide which inhibits Myc:Max heterodimerisation.
In Fig. 2 the amino acid residue sequence of the Myc 92 polypeptide is shown with residues numbered according to the corresponding positions of the amino acid residues in the full length Myc protein. A double stranded DNA sequence encoding the Myc 92 polypeptide is also shown. The internationally recognised 1-letter codes are used for both the nucleotide bases and the amino acid residues.
The invention is further illustrated by the following Examples, which should not be taken as limiting the invention in any way. Example 1
A polypeptide, Myc 92 (see Fig.l), containing a sequence corresponding to the B-HLH-Z domain of the Myc gene product wa co-expressed with the Myc and Ras oncoproteins in rat embryo cells (REC) by the same methodology as described for co- expression of the Myc and Ras oncoproteins [Land, H. et al. , Nature, 304, 596-602(1983)]. Whereas co-expression of only the Myc and Ras oncoproteins in REC leads to malignant transformation of the cells, this activity is suppressed when Myc 92 is also expressed. This occurs through competition with Myc for the interaction with cellular Max protein. EXAMPLE 2
The C-Myc protein (Myc) contains an amino-terminal transcriptional activation domain1 and a carboxyl-terminal basic/helix-loop-helix/leucine zipper (bHLH-Z) domain2"5 which directs dimerisation of Myc with its partner Max and DNA binding to sites containing a CACGTG core consensus6'9. Despite these characteristics, and the notion that Myc can modulate gene expression"0-10, a direct role for Myc or Max as transcription factors has never been demonstrated. Using Saccharo yces cerevisiae as an in vivo model system we show that the Myc protein is a sequence-specific transcriptional activator whose DNA_ binding is strictly dependent on dimerisation with Max. Transactivation is mediated by the amino-terminal domain of Myc. Max homodimers bind to the same DNA sequence as Myc/Max but fail to transactivate in our assays and thus can antagonise Myc/Max function. We also show that the Max HLH-Z domain has a higher affinity for the Myc HLH-Z domain than for itself and suggest that the heterodimeric Myc/Max activator forms preferentially at equilibrium.
Analysis of transcriptional regulation by Myc and Max proteins in mammalian cells is confused by endogenous Myc and Max. Since there is no evidence for Myc and Max homologues in ■S. cerevisiae , this organism appears as an excellent system fo such a study. Co-expression of human Myc with Maxl, Max2 or several Max deletion derivatives (Fig.3) in yeast cells leads to transcriptional activation from a CACGTG-CYCl-LacZ reporter whereas no activation is observed with either Myc or Max alone (Fig 4.a, lanes b) . To address whether the failure of Myc or Max alone to transactivate was due to an inability to bind DNA we tested their function when fused to the VP16 transactivatio domain (Fig. 4b) . As seen with Myc, VPl6-MycΔN (retaining Myc res. 180-439) transactivates only when co-expressed with Max. This is consistent with the idea that DNA binding and therefor transactivation by Myc requires association with Max (either Maxl or Max2) or at least the Max bHLH-Z domain (Max85, Figs. and 4a) , although an involvement of related yeast proteins11 in the observed effects cannot be ruled out. Unlike VP16-MycΔN, Maxl03-VP16 by itself transactivates the CACGTG-CYC1 promoter (Fig. 4b) . Thus, Max can bind DNA in the absence of Myc but does not significantly activate transcription in our system unless tagged with a heterologous transactivation domain. Accordingly, all Max proteins tested (see Fig. 3) bind DNA as homo-oligomers in yeast extracts (ref. 9 and data not shown) , as recently reported for bacterially expressed Max12,13. Transactivation of the CACGTG-CYCl promoter is sequence- specific, since much less or no transactivation by Myc/Max (Fig. 4a, lanes a) , VP16-MycΔN/Ma or Maxl03-VP16 (Fig. 4b) is observed from a control reporter lacking the CACGTG binding site (see also Fig. 4a, legend) .
Because of their apparent lack of transcriptional activity, we expected Max/Max dimers to antagonise transactivation by Myc/Max through competition for the same DNA target sites. Consistent with this idea, introducing an additional Max plasmid into the cells leads to a reduction of transactivation by Myc/Max or Myc/Max85 (Fig. 4c) . On the other hand, introduction of an additional Myc plasmid enhances transactivation levels in the presence of Max or Max85 (Fig. 4c) , supporting the conclusion that Myc provides the activation domain. Thus, the activities of different Myc/Max dimers primarily reflect the equilibrium between Myc/Max and Max/Max complexes rather than their absolute efficiencies (Fig. 4, see also legend) . For example, MaxΔC and Max85 proteins seem to allow better Myc/Max function than full-length Max (Fig. 4a) , while all Max proteins were expressed at similar levels (not shown) . Both truncated proteins lack the carboxyl-terminal nuclear import sequence12-14 and may therefore be less effective as homodimeric competitors, while the respective heterodimers with Myc are efficiently transported into the nucleus14.
Given that Max can interact with both itself and Myc, we next investigated the relative affinities of the Myc and Max HLH-Z domains for themselves and each-other using a yeast assay that monitors protein-protein interactions in vivo 5Λ~ (Fig. 5a) . We observed efficient Myc/Max but no Myc/Myc and weak Max/Max HLH-Z interactions (Fig. 5b) . Our results in yeast are paralleled by similar observations in mammalian cells, although no Max/Max interactions were detected12. The relative dimerisation affinities of Myc and Max HLH-Z domains are reminiscent of the interactions between Jun and Fos leucine zippers (reviewed in ref. 17) . We suggest that, as for Jun and Fos1!U9, Myc and Max preferentially form heterodimeric complexes at equilibrium.
The functional domains of Myc were analysed by measuring the dimerisation of various Myc mutants with SRF-Max72 and their DNA binding properties together with Max (Fig. 5c) . Mutants in the basic region (360N/P and 364,6,7R/A) are unable to transactivate the CACGTG reporter but retain dimerisation activity. Thus, similar to other bHLH proteins20-21, the basic region of Myc is essential for DNA binding but dispensable for dimerisation which is mediated by the HLH-Z domain alone (Fig. 5b) . Consistent with this idea, deletions of either HLH or LZ domains of Myc did not show any activity, although no positive controls are available for these mutants (not shown) . The transactivation domain of Myc maps to the 177 amino-terminal residues, consistent with the mapping of this domain in GAL4- Myc chimaeras in mammalian cells1. The deletion mutant MycΔN (retaining residues 178 to 439) fails to activate both in dimerisation and DNA binding assays (Fig. 5c) , but efficiently enhances transactivation by Max-VP16 (Fig. 3c) . This shows that MycΔN can dimerise and bind DNA but does not transactivate. The Myc transactivation domain also functions in yeast when fused to heterologous DNA binding domains (LexA- Myc22 and l-235Myc-SRF, not shown) .
The observations discussed in this paper are summarised in Fig. 6. In yeast, DNA binding and therefore transactivatio by Myc are dependent on dimerisation with Max. As both the transactivation and bHLH-Z domains are essential for Myc function (refs. 23-25, and Fig. 5c) our results further sugges that the growth-regulatory and oncogenic activities of Myc may depend on sequence-specific transactivation by Myc/Max dimers. Max homo-oligomers can also bind to the same DNA sequence yet do not detectably transactivate and can antagonise the function of Myc/Max through occlusion of DNA binding-sites. These findings are consistent with the lack of transactivation by GAL4-Max fusions in mammalian cells'2. However, a transactivation function for Max cannot be entirely ruled out. Myc and Max proteins are expected to preferentially form heterodimers. Unlike Myc, Max is a stable protein that is expressed in resting as well as growing cells (ref. 26; T.D.L. , D. Hancock and G.I.E., unpublished) . Therefore, we suggest that mitogenic induction of Myc expression27 leads to a shift in the equilibrium from Max/Max dimers to Myc/Max heterodimers. This transition may represent an important growth-regulatory step, since Myc activation is sufficient to commit cells into cell cycle10 or the apoptotic pathway25, although the activities of Myc and Max may also be regulated by phosphorylation5-13-26.
REFERENCES FOR EXAMPLE 2
1. Kato, G. . Barret, J. , Villa, G.M. & Dang, C.V. Mol Cell Biol 10, 5914-5920 (1990)
2. Murre, C. , McCaw, P.S. & Baltimore, D. Cell 56, 777-83
3. Landschutz, W.H. Johnson, P.F. & McKnight, S.L. Science 240, 1759-1764 (1988) . 4. Penn, L.J.Z., Laufer, E.M. & Land, H. in Seminars in
Cancer Biology (eds. Jones, N.) Vol 1, 69-80 (Saunders, London, 1990) . 5. Luscher, B. & Eisenman, R.N. Genes Dev 4, 2025-2035 (1990) . 6. Blackwell, T.K., Kretzner, L. , Blackwood, E.M. , Eisenman, R.N. & Weintraub, H. Science 250, 1149-1151 (1990) .
7. Blackwood, E.M. & Eisenman, R.N. Science 251, 1211-7 (1991) .
5 8. Prendergast, G.C., Lawe, D. & Ziff, E.B. Cell 65, 395-407 (1991) .
9. Littlewood, T.D., A ati, B. , Land, H. & Evan, G.I. Oncogene 7, 1783-1792 (1992) .
10. Eilers, M. , Schirm, S. & Bishop, J.M. Embo J 10, 133-141 0 (1991) .
11. Fisher, F. , Jayaraman, P.S. & Godling, CR. Oncogene 6, 1099-1104 (1991) .
12. Kato, G.J. Lee, W.M.F., Chen, L. & Dang, C.V. Genes Dev 6, 81-92 (1992) . 5 13. Berberich, S. & Cole, M.D. Genes Dev 6, 166-176 (1992).
14. Makela, T.P., Koskinen, P.J. , Vastrik, I. & Alitalo, K. Science 256, 373-377 (1992) .
15. Dalton, S. & Treisman, T. Cell 68, 597-612 (1992) .
16. Fields, S. & Song, 0. Nature 340, 245-246 (1989) .
20 17. Ransone, L.J. & Verna, I.M. Annu Rev Cell Biol 6, 539-557 (1990) .
18. O'Shea, E.K., Rutkowski, R. , Stafford, W.F. III . , & Kim, P.S. Science 245, 646-8 (1989) .
19. Smeal, T., Angel P., Meek J. & Karin, M. Genes Dev 3, 25 2091-2100 (1989) .
20. Davis, R.L., Cheng, P.F., Lassar, A.B. & Weintraub, H. Cell 60, 733-746 (1990) .
21. Voronova, A. & Baltimore, D. Proc . Natl . Acad Sci . USA 87, 4277-4726 (1990) . 22. Lech, K. , Anderson, K. & Brent, R. Cell 52, 179-184 (1988) .
23. Stone, J. , et al . Mol Cell Biol 7, 1697-1709 (1987)
24. Penn, L.J.Z., et al . Mol Cell Biol 10, 4961-4966 (1990) .
25. Evan, G.I., et al . Cell 69, 1-20 (1992).
26. Blackwood, E.M. , Luscher, B. & Eisenman, R.N. Genes Dev 6, 71-80 (1992) .
27. Kelly, K. & Siebenlist, U. Annu Rev Immunol 4, 317-338 (1986) .
28. Guarente, L. Sc Mason, T. Cell 32, 1279-1286 (1983).
29. Gregor, P.D., Sawadogo, M. & Roeder, R.G. Genes Dev 4, 1730-1740 (1990) .
30. Smith, M., et al . Cell 16, 753-761 (1979) . 31. Harshman, K.D., Moye, R.W. & Parker, C.S. Cell 53, 321- 330 (1988) .
FIGURE LEGENDS Fig. 3. Schematic representation of human c-Myc, Max, and Max deletion derivatives, and of the CACGTG-CYCl-LacZ reporter gene used in this study. The bHLH-Z domains of Myc and Max are aligned. Maxl and Max2 are the natural Max variants without and with the 9 amino acid insert, respectively7. The MaxΔC and Max85 mutants are truncated at the position in Max equivalent to the Myc carboxyl-terminus.
METHODS. Max2, MaxlΔC, Max2ΔC, and Max85 were generated by PCR with appropriate primers from a Maxl cDNA template (a gift from D. Gillespie) . All the coding regions were subcloned into galactose-inducible CEN-ARS plasmids of the pSD series of yeast expression vectors15, with either TRP1, LEU2 or HIS3 as selective markers. Due to usage of the CYC1 ATG initiation codon the following amino-terminal extensions precede the proteins: MTGFPGLQEFELAPTM (Myc), MTGFELE (Maxl and Maxl03-VP16 see Fig. 4) , MTGFT (Max 2, MaxlΔC, Max2ΔC) , and MTGFTMG (Max 85) , and MTGPFG (MycΔN see Fig. 4) . Additional experiments with proteins initiated at their own start codons indicated that the extensions do not significantly alter the behaviour of Myc, Maxl, Max2 , MaxlΔC and Max2ΔC (not shown) . All chimaeric fusions and PCR-generated inserts were verified by DNA sequencing. As a reporter gene we used a 2μ-CYCl promoter-LacZ plasmid derived from pLGΔ312 (ref. 28) containing the adenovirus major late promoter (MLP) element CTAGGCCACGTGACCGGGTGT (ref. 29) inserted between the Sma 1 and Xho 1 sites. Two other Myc binding sequences (CM1, ref. 6 and the PH04 promoter") gave similar results.
Fig. 4. Transcriptional activation by Myc and Max in yeast. a. jS-galactosidase units (U) in cells containing the indicated proteins and reporters. Residual transactivation of the control reporter may be due to two CACATG boxes in the CYC1 promoter30. CACATG is the only half-site change permitting DNA binding by Myc/Max, Maxl or Max2 , albeit with decreased affinity (ref. 12, D. Solomon, B.A. and H.L. , unpublished data) . Since all Max proteins were expressed at similar levels, the relatively higher efficiencies of Myc with Max2 , Max2ΔC or Max85, compared with Maxl or MaxlΔC, may in part be due to differences in affinity for DNA (ref. 9 and data not shown) . b . Transactivation by VPl6-MycΔN and Maxl03-VP16 together with Max or MycΔN. Note that VP16-MycΔN and Maxl03-VP16 activities cannot be directly compared, since their relative expression levels are unknown. c. Effect of a third Myc or Max plasmid on transactivation of the CACGTG reporter. Myc plasmids with different genetic markers than in (a) were used for Myc+Max and Myc+Max85 (LEU2 in a, HIS3 in 3 ) .
METHODS. Maxl03-VP16 was constructed by replacing Max sequences 3' of codon 103 with a VP16 fragment (residues 410- 490) from pSD.06a (ref. 15) . VP16-MycΔN contains the VP16 fragment upstream of Myc codons 180 to 439. MycΔN retains codons 178 to 439. Myc and Max plasmids were transformed into the ho ura3 his 3 trpl ade2 leu2 canl-100 yeast strain W303-1B (MATα) , and reporter plasmids into the isogenic strain W303-1A (MATa) . Protein/reporter combinations were generated by crossing transformants. Relative β-galactosidase units (U) in cultures induced for 12 hours with galactose were measured as previously described31 and normalised to cell numbers as U=1000A420/ (CVt) , where A420 is the absorbance at 420 nm, C is the density of the cell suspension (in A600/ml) , V is the vol. of cell suspension (ml) , and t is the total incubation time (min) .
Fig. 5. Myc/Max interactions in vivo . a . Assay for interactions between Myc and Max HLH-Z domains. In this assay15 activation of the SRE-LacZ reporter gene is strictly dependent on the interaction in trans between chimaeric proteins containing the SRF DNA-binding and the VP16 transactivation domains, respectively. A schematic representation of the fusion proteins and the reporter gene used is given. The SRE- CYCl-LacZ reporter gene is integrated into the genome of the indicator yeast strain S62L, (ref. 15.) b . Relative B- galactosidase units (as defined in Fig. 4) in S62L cells expressing the indicated proteins. The reciprocal combinations used provide an internal functional control for expression of the fusion proteins. Furthermore, all the SRF derivatives used here recruit equally well a SAP1-VP16 fusion protein which binds to the SRF moiety1'' (not shown) . c. Wild-type Myc and the indicated mutant derivatives were assayed for transactivation of SRE-CYCl-LacZ through dimerisation with SRF- Max72 and of CACGTG-CYC-l-LacZ through dimerisation and DNA binding with Max. The activity of each Myc protein in both assays is normalised to wild-type Myc activity (100%) . Wild- type Myc does not interact with SRF alone (not shown) . METHODS. Myc92, Myc73, and Max72-encoding DNA fragments were 'generated by PCR and subcloned in frame downstream of VP16 or SRF412 (residues 1-412 of SRF) in vectors pSD.06a and pSD.08, respectively15. SRF412 alone was expressed from a modified version of pSD.08. Myc point mutants were generated by site- directed mutagensis using standard methods. The reading frame of all fusions, the point mutations and all the PCR-generated inserts were verified by DNA sequencing.
Fig.6. Schematic summary of Myc and Max interaction and function. Myc and Max form stable heterodimers in the absence of DNA (in solution) . Weak Max/Max interactions can also be detected, but Myc/Myc homodimers do not form at physiological concentrations in vitro or in vivo . Both Myc/Max and Max/Max bind to the same DNA sequence. However, only the Myc/Max heterodimer detectably functions as a sequence-specific transcriptional activator in yeast.
1. A polypeptide which specifically binds to the helix-loop-helix/leucine zipper domain of either the Myc gene product or the Max gene product so as to prevent Myc:Max heterodimerisation under physiological conditions.
2. A polypeptide according to claim 1 which does not bind to any helix-loop-helix/leucine zipper domain other than the helix-loop-helix/leucine zipper domain of the Myc or Max gene products under physiological conditions.
3. A polypeptide according to claim 1 or claim 2 which specifically binds to the whole of the helix-loop-helix/leucine zipper domain of the Myc or Max gene product.
4. A polypeptide according to any one of claims 1 to 3 which does not form homodimers under physiological conditions.
5. A polypeptide according to any one of claims 1 to 4 which specifically binds to the helix-loop-helix/leucine zipper domain of the Max gene product.
6. A polypeptide according to any one of claims 1 to 4 which specifically binds the the helix-loop-helix/leucine zipper domain of the Myc gene product.
7. A polypeptide according to any one of claims 1 to 6 coupled to a targeting antibody. 3. A nucleic acid having a sequence encoding a polypeptide according to any one of claims 1 to 7.
9. An expression or cloning vector comprising a nucleic acid according to claim 8.
10. A transformed or transfected cell containing heterologous nucleic acid having a sequence encoding a polypeptide according to any one of claims 1 to 7 or a vector according to claim 9.
11. A process for producing a polypeptide according to any one of claims 1 to 7 comprising synthesis from amino acid precursors or expression of a nucleic acid according to claim 8.
12. A polypeptide according to any one of claims 1 to 7 for use in a method of treatment practised on the human or animal body.
13. Use of a polypeptide accordiλng to any one of claims 1 to 7 in the preparation of a medicament for use in a method of treatment practised on the human or animal body for treatment of a tumour associated with overexpression of the Myc gene product.
14. A pharmaceutical composition comprising a polypeptide according to any one of claims 1 to 7 and a diluent or carrier therefor. 15. A method of treatment comprising administering a tumour-reducing non toxic amount of polypeptide according to any one of claims 1 to 7 or a pharmaceutical composition according to claim 8 to a human or animal having a tumour associated with overexpression of the Myc gene product.
16. A process for screening candidate therapeutic agents intended to interfere with Myc: Max heterodimerisation which process comprises contacting cells with the candidate therapeutic agents or, when the candidate therapeutic agent is a peptide, optionally expressing the peptide in the cells, which cells contain a reporter gene the expression of which is enhanced or reduced by agents which interfere with Myc: Max heterodimerisation, and observing the level of expression of the reporter gene by the cells.