WO2002033104A2 - Antibiotic-based gene regulation system - Google Patents

Abstract

The invention relates to a novel system for gene regulation in eukaryotic cells, and methods of using the same for protein production, tissue engineering and gene therapy. In particular, the invention provides a new system for antibiotic-regulated gene expression in eukaryotic cells based on sequences from Actinomycetes antibiotic resistance promoters, polypeptides that bind to the same in an antibiotic responsive manner, and nucleotides encoding such polypeptides. Further, the invention provides novel and sensitive methods of screening for candidate antibiotics.

Description

Antibiotic-Based Gene Regulation System

Cross Reference to Related Applications

This application is a continuation-in-part of application serial no. 09/411,687, filed

October 4, 1999, which is a continuation-in-part of application serial no. 09/298,768, filed

April 23, 1999. The contents of these applications are hereby incorporated herein by reference in their entireties.

Field of the Invention The invention relates to a novel system for gene regulation in eukaryotic cells, and methods of using the same for protein production. In particular, the invention provides a new system for antibiotic-regulated gene expression in eukaryotic cells, including plant and mammalian cells, based on sequences from Actinomycetes antibiotic resistance promoters, polypeptides that bind to the same in an antibiotic responsive manner, and nucleotides encoding such polypeptides.

Background of the Invention

Citation of any reference in this section or any section hereof is not to be construed as an admission that such reference is available as prior art to this invention.

Controlled expression of transgenes or of target endogenous genes will likely be essential for success of many genetic therapies. Constitutive expression of transgenes has resulted in down-regulation of effector systems and/or cellular toxicity in animal studies

(Efrat et al, 1995, Proc. Natl. Acad. Sci. USA 92, 3576-3580). Regulated expression in response to metabolic, hormonal, or environmental signals is the normal situation for many eukaryotic genes. In order to mimic natural physiological expression patterns with transgenes, and to minimize interactions with human gene regulation signals, binary promoter/transactivator configurations of heterologous origin which respond to heterologous stimuli have been developed in recent years. However, many exogenous stimuli which modulate these artificial mammalian regulons have proven to be incompatible with human therapeutic use due to cytotoxicity or undesired side effects (Bairn et al, 1991, Proc. Natl.

Acad. Sci. USA 88, 5072-5076; Braselmann et al, 1993, Proc. Natl. Acad. Sci. USA 90, 1657-1661; No D. et al, 1996, Proc. Natl. Acad. Sci. USA 93, 3346-3351; Rivera et al, 1996, Nat. Medicine 2, 1028-1032; Suhr et al, 1998, Proc. Natl. Acad. Sci. USA 95, 7999- 8004; Wang et al, 1994, Proc. Natl. Acad. Sci. USA 91, 8180-8184). The tetracycline- regulated mammalian expression system avoids these problems, and is described in U.S. Patent Nos.: 5,888,981; 5,866,755; 5,789,156; 5,654-168; and 5,650,298, to name just a few examples. However, the tetracycline-regulated system can fail to suppress gene expression adequately under repressed conditions.

Moreover, future gene therapy strategies will require technology which allows independent control of two different transgenes or sets of transgenes which are cotranscribed in a multicistronic configuration. For example, many tissue expansion and ex vivo gene 5 therapy scenarios will require a two-step process beginning with expression of growth- promoting genes to enable expansion of grafted tissues in culture, followed by induction of growth suppressors to prevent tumorigenic behavior of treated cells after reimplantation. Sustained proliferation control is also required for stem cell-based technologies currently evaluated for eventual cell and tissue replacement therapy, since stem cells are tumorigenic ⁰ (Rossant et al, Nat. Biotechnol. 17, 23-24; Solter et al, Science 283, 1468-1470). The second independent gene regulation system could be used in such cells for pharmacologic control of one or several secreted therapeutic proteins, such as insulin, to enable titration of circulating proteins into the therapeutic range or adapt expression to optimal daily dosing ₅ regimes. There is, therefore, a need for new mammalian gene regulation systems which employ modern, therapeutically proven antibiotics as controlling agents, and which can be used in combination with the tetracycline regulation system, with minimal interaction between tetracycline control and the new control modality.

The human oral antibiotic pristinamycin consists, like other streptogra ins, of a ^KJ mixture of two structurally dissimilar molecules, the streptogramin A component pristinamycin II (PII), a polyunsaturated macrolactone, and the streptogramin B component pristinamycin I (PI), a cyclic hexadepsipeptide. The water-soluble form of pristinamycin, Synercid , recently approved in the U.S. and Europe for use against most multiple drug- _ resistant Gram-positive bacteria (Barriuere et al, 1994, Expert Opin. Invest. Drugs 3, 115- 131; Baquero et al, 1997, J. Antimicrob. Chemother. 39, 1-6), is composed of dalfopristin, a 26-sulphonyl derivative of PII, and quinupristin, which is derived from PI by synthetic addition of a (5δ R)-[(3S)-quinuclidinyl] thiomethyl group. Nirginiamycin is another important streptogramin used as a growth promotant in livestock feed (Nagaraja et al, 1998, 5 J. Anim. Sci. 76, 287-298). Either the A or B streptogramin components are individually bacteriostatic, but streptogramins A and B together are synergistically bactericidal (up to 100 times more active), a phenomenon which lowers incidence of acquired antibiotic resistance, since high level resistance to the combined streptogramins is likely only when both type A and type B streptogramin resistance determinants are present simultaneously (Cocito et al, 1997, J. Antimicrob. Chemother. 39, 7-13). Recently, a pristinamycin resistance determinant (ptr) has been cloned from S. pristinaespiralis (Blanc et al, 1995, Mol. Microbiol. 17, 989- 999; Salah-Bey et al, 1995, Mol. Microbiol. 17, 1001-1012; Salah-Bey et al, 1995, Mol. Microbiol. 17, 1109-1119). Expression from the ptr promoter (P_PTR) is induced by _j pristinamycin, particularly by PI. A protein called Pip (pristinamycin-induced protein) has recently been purified based on its affinity to P_PTR and its gene cloned from Streptomyces coelicolor.

Recently, the prevalence of multidrug resistant human pathogenic bacteria has increased dramatically. This increase correlates with an escalation of bacterial disease and

20 related mortality. Also, antibiotic chemotherapy is becoming more difficult as the percentage of elderly and immunocompromised patients grows. The European Commission has already reacted to this situation by banning the use of certain antibiotics as a growth promoter in livestock feed, among them the streptogramin virginiamycin, so as to limit the environmental

₂.- spread of antibiotics (thought to be a major driving force for selection of multidrug resistant pathogenic bacteria) thereby preserving the use of antibiotics for human therapy.

Summary of the Invention

The present invention is directed to a new system for antibiotic-regulated gene on

^Jυ expression in eukaryotic cells based on sequences from Actinomycetes antibiotic resistance promoters, and polypeptides that bind to the same in an antibiotic responsive manner. In one aspect, the invention is directed to an isolated nucleic acid encoding a polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic. The P_ptr sequence is derived

„_. from an actinomycetes bacterium, in particular an antibiotic responsive operon from an actinomycetes bacterium. In another aspect, the invention encompasses an isolated nucleic acid encoding a fusion protein which regulates transcription, the fusion protein having a first polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic, operatively linked to a second polypeptide which activates or represses transcription in eukaryotic cells. In another aspect, the invention is directed to the proteins encoded by such nucleic acids. These proteins can be used to activate or repress transcription from a desired nucleotide sequence to be transcribed that is operatively linked to a P_ptr sequence.

The invention also includes host cells that contain nucleic acids encoding the proteins and fusion proteins of the invention. Such host cells optionally contain a desired nucleotide sequence to be transcribed operatively linked to a P_ptr sequence. The nucleotide sequence to be transcribed can be endogenous or exogenous to the host cell. Suitable host cells include, for example, mammalian cells such as CHO-Kl, BHK-21, HeLa, COS-7, HEK 293, HT1080, PC12, MDCK, C2C12, Jurkat, NTH3T3, K-562, TF-1 and P19, or plant cells such as those derived from barley, wheat, rice, soybean, potatoe, arabidopsis and tobacco (e.g. Nicotiana tabacum SRI) etc. Suitable hosts also include plant-derived hairy root cultures such as, for example, hairy root cultures derived from Artemisia, Atropa, Beta, Brugmansia, and others such as those described in Shanks and Morgan, 1999, Curr. Opin. Biotechnol. 10:151-155. Other suitable cell lines of mammalian and plant origin are well known to those of skill in the art and include, for example, those described in ATCC Cell Lines and Hybridomas 8^th Edition, 1994, American Type Culture Collection, Rockville, MD. The present invention also provides transgenic animals comprising the nucleic acids of the invention. Preferred transgenic animals are transgenic mice. Yet another aspect of the invention provides an isolated nucleic acid having a P_ptr sequence operatively linked to a first eukaryotic promoter. The first eukaryotic promoter can also be operatively linked to a first coding sequence. Optionally, the nucleic acid can also contain at least one tet operator sequence operatively linked to a second eukaryotic promoter, which in turn can be operatively linked to a second coding sequence. Either coding sequence can encode any protein of interest for which production is desired. Further, one of the coding sequences can encode a tumor suppressor gene product or a gene product which activates cell proliferation. Still further, at least one of the promoters can be operatively linked to more than one coding sequence through the use of, for example, an interanl ribosomal entry site. Host cells genetically engineered to contain these nucleic acids are also provided by the invention. Another aspect of the invention provides vectors for P_ptr- regulated expression of a gene in a eukaryotic cell. Suitable vectors for P_ptr- regulated expression include mammalian expression vectors, plant expression vectors, retroviral expression vectors, lentiviral expression vectors and other vectors known to those of skill in the art.

In another aspect, the invention provides a method for regulating expression of a P_ptr -linked gene in a eukaryotic cell. The method entails introducing into the cell a nucleic acid molecule encoding a PIT. The PIT can be a P_ptr -binding protein, or can comprise a P_ptr - binding protein operably linked to a polypeptide that activates or represses transcription in eucaryotic cells, thereby rendering the P_ptr -linked gene capable of regulation by an antibiotic that binds to the P_ptr -binding protein in the cell. One can then modulate the level of the antibiotic in the cell to regulate expression of the P_ptr -linked gene

Still another aspect of the invention is a process for producing a protein by culturing a eukaryotic cell containing a P_ptr -linked gene that encodes the protein and a nucleic acid molecule encoding a PIT. The PIT can be a P_ptr -binding protein, or can comprise a P_ptr - binding protein operably linked to a polypeptide that activates or represses transcription in eucaryotic cells. Expression of the P_ptr -linked gene is then regulated by modulating the level of an antibiotic that binds to the P_ptr -binding protein in the cell. Optionally, the process 0 entails the step of collecting the protein produced by the cell.

Another aspect of the invention is a method of screening for candidate antibiotics.

The method entails incubating the host cells of the invention, the host cells containing a P_ptr - linked reporter gene and a sequence encoding a P_ptr -binding protein, in the presence of a test 5 compound, wherein a change in the transcription of the reporter gene indicates that the test compound is a candidate antibiotic.

Brief Description of the Figures

Figure 1 : Gel retardation assay showing Pip/P_PTR interaction. P_PTR of S. ^υ pristinaespiralis contains 3 dyad symmetrical binding sites¹⁶ (lane A; P_PTR in the absence of Pip) all of which are saturated at relatively large amounts of purified Pip (0.4 pmol; lane A) resulting in a single shifted band. Smaller amounts of Pip (0.02 pmol) presented an intermediate situation in which all possible permutations of site occupancy are represented in _. three distinct retarded bands (lane C). Pip (0.2 pmol lanes D-M) was titrated with a range of PI from 50 pmol to 0.01 pmol (lanes D-H; D: 50 pmol; E: 10 pmol; F: 1 pmol; G: 0.1 pmol; H: 0.01 pmol). Under identical conditions the semisynthetic PI derivative quinupristin was also titrated (lanes I-M; I: 50 pmol; J: 10 pmol; K: 1 pmol; L: 0.1 pmol; M: 0.01 pmol). Figure 2: Dose-response curve for Pi-dependent gene expression. The cell line CHO- 5 PITl-SEAP was grown for 48 h at different PI concentrations (ng/ml). The relative SEAP production is shown over 5 orders of magnitude of PI concentration.

Figure 3: Graphical representation of pMF188, a vector that combines tetracycline- responsive SEAP (human secreted alkaline phosphatase) and pristinamycin-responsive GFP expression in a single double regulated expression unit: PhCMN*-l-SEAP-pAi-PpiR-GFP-

Figure 4: Graphical representation of pMF201, a vector that contains tet-regulated CFP (cyan fluorescent protein) and pristinamycin-regulated YFP (yellowish fluorescent protein) in a single expression unit: PhCMN*-l-CFP-pAj8 -PpiR-YFP-pAπ. 15 Figure 5: Graphical representation of pMF196, a vector that is isogenic to pMF188 and pMF201 but contains the following double regulated expression unit: PpiR-YFP-pAj-pA

Iis28 -CFP-PhCMN*-l-

Figure 6: Graphical representation of vector pDuoRexl. pDuoRexl contains two 2_Q expression units which allow independent adjustable gene expression in mammalian cells using tetracycline and pristinamycin as regulating agents. The tetracycline-responsive expression unit is driven by P_hCMV*.ι followed by an internal ribosomal entry site (IRES) flanked by two multiple cloning sites. The expression unit contains a terminal artificial polyadenylation sequence (pAl). The IRES element enables dicistronic P_hCMV*_ι -driven expression. The second regulatable cistron is driven by P_PIR followed by a multiple cloning site and terminated by a second polyadenylation sequence.

Figure 7: Graphical representation of vector pMF195. PMF195 is a pDuoRexl derivative containing the tumor suppressor gene p27 under control of the tetracycline-

,, _n responsive promoter P_hCMV*-ι and the Mekl gene under control of P_Pj_R. Mekl is a central kinase of the main growth factor signalling pathway and was particularly mutated (MeklDD) to be constitutively active even in the absence of exogenous growth factors. Mekl is an example of a class of proteins which promote cell proliferation.

Figure 8: Graphical representation of vectors pMF189 andpMF229.

3 Figure 9: Graphical representation of multicistronic vectors. Figure 9A illustrates vectors pTRTDEΝT9 and pTRTDENTlO, while Figure 9B illustrates pTRTDENTl 1 and pTRTDENT12. The pTRIDENT-derived mammalian expression vectors contain a tricistronic expression unit which is either driven by the pristinamycin-repressible (P_pir) or the pristinamycin-inducible (P_pirON) promoter. Whereas the first cistron of the tricistronic expression unit is transcribed in a cap-dependent manner, the subsequent genes rely on cap- independent translation initiation based on internal ribosomal entry sites of polyio viral (IRES) origin or of encephalomyocarditis virus. The multicistronic expression unit is terminated by a polyadenylation site (pA) of the SN40 virus.

Figure 10 A: Graphical representation of the PLPpOFF expression vector pMF276. pMF276 harbors the PIT expression unit which is driven by the constitutive promoter of the cauliflower mosaic virus gene 35S (P_CaMV3ss) ^m& terminated by the polyadenylation site derived from the octopine synthase gene (pA_0CS). PIT is a fusion protein of the Streptomyces coelicolor Pip protein and the NP16 transactivation domain of the Herpes simplex virus. Figure 10B: Graphical representation of the PIPpOFF expression vectors pMF279/281. pMF279 and pMF281 are β-glucuronidase (GUS) expression vectors which are driven by the pristinamycin-repressible promoters which contain an artificial PIR3 binding site fused to a P_CaMV35s minimal promoter. P_pPIR] (pMF279) harbors the PIR3 Pip- binding module in sense orientation and P_pPIR2 (pMF281) in antisense orientation.

Figure 11 : Graphical representation of PIPpOΝ expression vectors pMF273, pMF275 and pMF265/266. pMF273 and pMF275 are Pip expression vectors encoding Pip alone

(pMF275) or as fusion to the nuclear localization signal (ΝLS) derived from the plant transcription factor TGAlb (pMF273). The Pip or Pip-ΝLS expression units are driven by Pc_aMV35s ^^ terminated by the polyadenylation site derived from the octopine synthase gene

(pA_0CS). The pristinamycin-inducible promoters comprise the PIR3 element containing 9 Pip binding sites placed in front of P_CaMV35S. In pMF265 (P_pPIR0N.) ^tne PDR-3 module is cloned in antisense and in pMF266 (P_PPIRON2) ^m sense orientation. Both promoters drive the β- glucuronidase reporter-encoding expression unit which is terminated by a polyadenylation site derived from the octopine synthase gene (pA_0CS).

Figure 12: Regulation performance of the pristinamycin-repressible gene regulation system (PIPpOFF) in Nicotiana tabacum suspension cultures. The PIPpOFF system is repressed by the streptogramin antibiotic pristinamycin (Pyostacin^®). Three suspension cultures of the tobacco cell line SRI were cotransfected (see Materials and Methods) with an equimolar ratio of pMF276 (providing PIT) and pMF279 (providing

(GUS)) or pMF281 (providing P_pPIR2-GUS). All cultures were grown for 48 h in fresh medium in the presence (+PI; 50 μg/ml) and absence (-PI) of pristinamycin and then assayed for GUS activity (pmoles 4-MU min * mg^"1 protein^"1). As negative control for GUS expression untransfected SRI cells were used (cntrl).

Figure 13: Regulation performance of the pristinamycin-inducible plant gene regulation system (PIPpON) in Nicotiana tabacum suspension cultures. The PIPpON system is induced by addition of the streptogramin antibiotic pristinamycin (50 μg/ml Pyostacin^®). Three suspension cultures of the tobacco cell line SRI were cotransfected with all combinations of transrepressor-encoding plasmids pMF273 (P_CaMV35S-Pip-NLS-pA_nos) or pMF275 (P_CaMV35s-Pip-pA„_0S) and P_pPIR0N-driven GUS expression vectors pMF265 (P_pPIR0N1) and pMF266 (P_pPIRo_N2)- In addition, pMF265, pMF266 and the isogenic control plasmid pMF255 (Pc._MV35g-GUS-pA_j.-_s) have been transfected alone to determine the maximal expression levels of the P_pPIRON promoters and the native P_CaMV35s- AH cultures were grown for 48 h in fresh medium in the presence (+PI; 50 μg/ml) and absence of pristinamycin (-PI) and then assayed for GUS (y#-glucuronidase) activity (pmoles 4-MU min ^λ mg^"1 protein ^l).

Figure 14: Graphical representation of pMF311 and pMF312. pMF311 and pMF312 are multicistronic retroviral expression vectors which form replication-incompetent retroviruses upon transfection into appropriate retroviral packaging cell lines. The multicistronic expression units is driven by the 5' LTR and terminated by 3' LTR. A partial packaging signal (Psi) is located 3' of 5' LTR. The tricistronic expression unit encodes the pristinamycin-dependent transactivator (PIT), the tetracycline-dependent (tTA; pMF311) or reverse tetracycline-dependent (rtTA; pMF312) transactivators and the zeocin resistance determinant (zeo). Whereas the first cistron (PIT) is translated in a classical cap-dependent manner, translation-initation of tTA/rtTA is mediated by an IRES (internal ribosomal entry site) element of polio viral origin and zeo translation is initiated by a CITE* element (cap- independent translation enhancer) derived from encephalomyocarditis virus.

Detailed Description of the Invention

Systems to control expression of foreign genes will be essential tools for future cell- and gene-based human therapies. The present invention is directed to a new system for antibiotic-regulated gene expression in eukaryotic cells based on sequences from

Actinomycetes antibiotic resistance promoters, and polypeptides that bind to the same in an antibiotic responsive manner. For purposes of the invention, a P_ptr sequence is a sequence from an Actinomycetes 5 antibiotic resistance promoter that binds a particular polypeptide, PIP (also referred to herein as "the first polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic"), in an antibiotic dependent way. The antibiotic resistance promoter can be derived from a naturally-occurring episome, but is preferably a chromosomal promoter. Preferably, PIP binds to the P_ptr sequence in the absence of its cognate antibiotic, and is released from the P_ptr sequence when antibiotic is present, although the reverse situation is also within the scope of the invention. Accordingly, in the presence of the PIP, expression from the antibiotic resistance promoter containing the P_ptr sequence is regulated by the presence or absence of antibiotic. Thus, for purposes of the invention, the term "cognate 5 antibiotic" means the antibiotic which when bound to the PIP results in the release of the protein from its P_PIR binding site.

PIPs are, for the purposes of the invention, derived from or related to PIP proteins produced by Actinomycetes. By "derived from" PIP proteins produced by Actinomycetes, is meant, in this context, that the amino acid sequence is identical to a naturally occurring PIP,

20 or contains only conservative amino substitutions and but remains at least 70%, preferably

80%, and more preferably 90% identical at the amino acid level. By "related to" PIP proteins produced by Actinomycetes is meant, for purposes of the invention, that the polynucleotide sequence that encodes the amino acid sequence hybridizes to a naturally occurring PIP

2₅ produced by Actinomycetes under at least low stringency conditions, more preferably moderate stringency conditions, and most preferably high stringency conditions, and binds to a P_ptr recognition sequence. Conservative substitutions known in the art and described by

Dayhof, M.D., 1978, Nat. Biomed. Res. Found., Washington, D.C., Vol. 5, Sup. 3, among others. Genetically encoded amino acids are generally divided into four groups: (1) acidic =

3 ^J0^υ aspartate, glutamate; (2) basic = lysme, arginine, histidine; (3) non-polar = alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, tryptophan; and (4) uncharged polar = glycine, asparagine, glutamine, cysteine, serine, threonine, tyrosine. Phenylalanine, tryptophan and tyrosine are also jointly classified as aromatic amino acids. A substitution of one amino acid in a particular group with another amino acid in the same group is generally regarded as a conservative substitution.

The Actinomycetes include Gram-positive bacteria from the following taxonomic groups: Actinomycineae; Corynebacterineae; Frankineae; Glycomycineae; Kineococcus group; Micrococcineae; Micromonosporineae; Propionibacterineae; Pseudonocardineae; Streptomycineae; Streptosporangineae; and unclassified Actinomycetales (Stackebrandt et al, 1997, Int. J. Syst. Bacteriol. 47 (2): 479-491). Additional information regarding Actinomycetes, and the members of each group, can be found in McNeil and Brown, 1994, Clinical Microbiol. Review 7:357-417, and the references cited therein. Further information can be found in Wheeler et al, 2000, Nucleic Acids Res. 28:10-14 and in Benson et al, 2000, Nucleic Acids Res. 28:15-18, and the references cited in those publications. Preferred actinomycetes are streptomycetes, and are thus from the group Streptomycineae.

The invention is illustrated below in one embodiment using a particularly preferred antibiotic system employing streptogramins. Protein database searches have identified Streptomyces and Amycolatopsis genes within antibiotic biosynthetic clusters bearing sequence similarity to pip that may respond to other antibiotics as described in the following Table 1. These actinomycetes can contain P_ptr sequences and PIPs that can be used in the compositions and methods of the invention.

Table 1: Sreptomyces and Amycolatopsis antibiotic biosynthetic gene clusters.

References for TABLE 1 are as follows: 5 1. August et al, 1998, Biosynthesis of the ansamycin antibiotic rifamycin: deductions from the molecular analysis of the rif biosynthetic gene cluster of Amycolatopsis mediterranei S699. Chem. Biol. 5:69-79.

2. Fernandez et al, 1991, The act cluster contains regulatory and antibiotic export genes, direct targets for translational control by the bldA tRNA gene of Streptomyces. Cell 66:769-

¹⁰ 780.

3. Guilfoile et al, 1992, Sequence and rranscriptional analysis of the Streptomyces glaucescens tcniAR tetracenomycin C resistance and repressor gene loci. J Bacteriol 174:3651-8.

25 4. Otten et al, 1995, Regulation of daunorubicin production in Streptomyces peucetius by the dnrR2 locus. J. Bacteriol. 177:1216-1224.

5. Schwecke et al., 1995, The biosynthetic gene cluster for the polyketide immunosuppressant rapamycin. Proc Natl Acad Sci U S A 92:7839-7843.

6. Westrich et al, 1999, Cloning and characterization of a gene cluster from Streptomyces

20 cyanogenus S 136 probably involved in landomycin biosynthesis. FEMS Microbiol Lett

170:381-7.

Actinomycetes antibiotic-resistant promoters that contain P_ptr sequences can also be identified by generating Actinomycetes gene libraries in a heterologous host, and growing the

2 heterologous host under conditions containing selective antibiotics. Emerging resistant clones containing antibiotic resistant determinants can be identified by their characteristic protein binding motifs following sequence analysis.

Polynucleotide sequences encoding the first polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic can be used to clone homologous PIPs in

³⁰ other actinomycetes. Such homologous PIPs can also be used as the first polypeptide of the fusion protein, and their P_ptr recognition sequence operably linked to the eukaryotic promoter of interest. Thus, the invention also is directed to to nucleic acids hybridizable to or complementary to PIPs described herein. A preferred PIP is the repressor of a streptogramin resistance operon of S. coelicolor described below. Such PIPs are at least 50%, preferably 60%, more preferably 70%, even more preferably 80%, yet more preferably 90%, and most preferably 95% homologous at the amino acid sequence level to a PIP described herein Homology can be calculated using, for example, the BLAST computer program (Altschul et al., 1997, Nucleic Acids Res. 25:3389-402). Typical parameters for determining the similarity of two sequences using BLAST 2.0 are a reward for match of 1, penalty for mismatch of -2, open gap and extension gap penalties of 5 and 2, respectively, a gap dropoff of 50, and a word size of 11.

In a specific embodiment, a nucleic acid which is hybridizable to a PIP nucleic acid, or to a nucleic acid encoding a PIP amino acid under conditions of low stringency is provided. By way of example and not limitation, procedures using such conditions of low stringency are as follows (see also Shilo and Weinberg, 1981, Proc. Natl. Acad. Sci. USA 78:6789-6792): Filters containing DNA are pretreated for 6 h at 40°C in a solution containing 35% formamide, 5X SSC, 50 mM Tris-HCl (pH 7.5), 5 mM EDTA, 0.1% PVP, 0.1%.Ficoll, 1% BSA, and 500 μg/ml denatured salmon sperm DNA. Hybridizations are carried out in the same solution with the following modifications: 0.02% PNP, 0.02% Ficoll, 0.2% BSA, 100 μg/ml salmon sperm DΝA, 10% (wt/vol) dextran sulfate, and 5-20 X 10⁶ cpm ³²P-labeled probe is used. Filters are incubated in hybridization mixture for 18-20 h at

40°C, and then washed for 1.5 h at 55 °C in a solution containing 2X SSC, 25 mM Tris-HCl

(pH 7.4), 5 mM EDTA, and 0.1% SDS. The wash solution is replaced with fresh solution and incubated an additional 1.5 h at 60°C. Filters are blotted dry and exposed for autoradiography. If necessary, filters are washed for a third time at 65-68 °C and reexposed to film. Other conditions of low stringency which can be used are well known in the art (e.g., as employed for cross-species hybridizations).

In another specific embodiment, a nucleic acid, which is hybridizable to a PIP nucleic acid under conditions of moderate stringency is provided. For example, but not limited to, procedures using such conditions of moderate stringency are as follows: Filters containing

DΝA are pretreated for 6 h at 55 °C in a solution containing 6X SSC, 5X Denhart's solution,

0.5% SDS and 100 μg/ml denatured salmon sperm DΝA. Hybridizations are carried out in the same solution and 5-20 X 10⁶ cpm ³²P-labeled probe is used. Filters are incubated in hybridization mixture for 18-20 h at 55 °C, and then washed twice for 30 minutes at 60 °C in a solution containing IX SSC and 0.1% SDS. Filters are blotted dry and exposed for autoradiography. Other conditions of moderate stringency which can be used are well-known in the art. Washing of filters is done at 37°C for 1 h in a solution containing 2X SSC, 0.1%

SDS. In another specific embodiment, a nucleic acid which is hybridizable to a PIP nucleic acid under conditions of high stringency is provided. By way of example and not limitation, procedures using such conditions of high stringency are as follows: Prehybridization of filters containing DNA is carried out for 8 h to overnight at 65 °C in buffer composed of 6X SSC, 50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 0.02% PNP, 0.02% Ficoll, 0.02% BSA, and

500 μg/ml denatured salmon sperm DΝA. Filters are hybridized for 48 h at 65 °C in _Λ prehybridization mixture containing 100 μg/ml denatured salmon sperm DΝA and 5-20 X 10 cpm of ³²P-labeled probe. Washing of filters is done at 37 °C for 1 h in a solution containing

2X SSC, 0.01% PNP, 0.01% Ficoll, and 0.01% BSA. This is followed by a wash in 0.1X

SSC at 50 °C for 45 min before autoradiography. Other conditions of high stringency which can be used are well known in the art.

New PIP proteins can also be isolated by binding to P_ptr sequences. For example, the polynucleotide with the P_ptr sequence is immobilized on a matrix and ideally packed in a column. Bacterial or eukaryotic extracts are applied to the column under conditions which allow PIP homologs to bind to the immobilized target sequence. Following appropriate washing steps, the PIP homolog is eluted by suitable conditions (e.g., addition of antibiotic) and the sequence of the purified protein determined and the corresponding gene cloned.

PIP proteins of the invention include polypeptides that bind to P_ptr sequences, as well as fusion proteins containing a first polypeptide that binds to P_ptr sequences operatively linked to a second polypeptide which activates or represses transcription in eukaryotic cells. In this context, operatively linked means that the two proteins are covalently or non-covalently bound to one another in such a manner that they retain their functional activities of binding to

P_ptr sequence (first polypeptide) and activating or repressing transcription (second polypeptide).

The fusion proteins of the invention contain both the first polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic, and a second polypeptide which activates or represses transcription in eukaryotic cells. By activating transcription is meant that the rate of transcriptional initiation is increased from the nucleotide sequence to be transcribed that is operatively linked to a P_ptr sequence when the fusion protein that activates transcription is bound to the P_p(r sequence, as opposed to when it is not bound. Similarly, by repression of transcription is meant that the rate of transcriptional initiation is decreased from the nucleotide sequence to be transcribed that is operatively linked to a P_ptr sequence when the fusion protein that represses transcription is bound to the P_ptr sequence, as opposed to when it it not bound.

Accordingly, in one aspect, the first polypeptide of the fusion protein that activates transcription is operatively linked to a second polypeptide which directly or indirectly activates transcription in eukaryotic cells. To operatively link the first and second polypeptides, typically nucleotide sequences encoding the first and second polypeptides are 0 ligated to each other in-frame to create a chimeric gene encoding a fusion protein, although the first and second polypeptides can be operatively linked by other means that preserve the function of each polypeptide (e.g., chemically crosslinked). The first and second polypeptides can be in any order. The second polypeptide of the transactivator can itself 5 possess transcriptional activation activity (i.e., the second polypeptide directly activates transcription) or it activates transcription by an indirect mechanism, through recruitment of a transcriptional activation protein to interact with the fusion protein. Accordingly, the term "a polypeptide which activates transcription in eukaryotic cells" includes polypeptides which either directly or indirectly activate transcription. 0 Polypeptides which can function to activate transcription in eukaryotic cells are well known in the art and are described, for example, in U.S. Patent No. 5,654,168. Such polypeptides include the herpes simplex virus virion protein 16 (VP16, the amino acid sequence of which is disclosed in Triezenberg, S. J. et al, 1988, Genes Dev. 2:718-729), particularly the 127 amino acid C-terminus or the 11 amino acid C-terminus. Suitable C-terminal peptide portions of VP16 are described in Seipel, K. et al (EMBO J., 1992, 13:4961-4968).

Acidic transcription activation domains, proline-rich transcription activation domains, serine/threonine-rich transcription activation domains and glutamine-rich transcription ^υ activation domains can all be used in the compositions and methods of the invention. NP16 polypeptides and amino acid residues 753-881 of GAL4 are acidic activating domains. Another polypeptide that activates transcription is the p65 domain of ΝF-κB (Schmitz and Baeuerle, 1991, EMBO J. 10:3805-3817). Examples of proline-rich activation domains ₅ include amino acid residues 399-499 of CTF/NFl and amino acid residues 31-76 of AP2. Examples of serine/threonine-rich transcription activation domains include amino acid residues 1-427 of ITF1 and amino acid residues 2-451 of ITF2. Examples of glutamine-rich activation domains include amino acid residues 175-269 of Octl and amino acid residues 132-243 of Spl. The amino acid sequences of each of the above described regions, and of 5 other useful transcriptional activation domains, are disclosed in Seipel, K. et al. (EMBO J., 1992, 13:4961-4968). This reference also describes methods of identifying new transcriptional activation domains which are within the scope of the invention.

In another embodiment, the second polypeptide of the fusion protein indirectly activates transcription by forming a non-covalent association with a transcriptional activator.

10

For example, a PIP of the invention can be fused to a polypeptide domain (e.g., a dimerization domain) capable of mediating a protein-protein interaction with a transcriptional activator protein, such as an endogenous activator present in a host cell. Non-covalent interactions between DNA binding domains and transactivation domains are known in the art

25 (see e.g., Fields and Song, 1989, Nature 340:245-247; Chien et al, 1991, Proc. Natl. Acad Sci. USA 88:9578-9582; Gyuris et al.,1993, Cell 75:791-803; and Zervos, A. S., 1993, Cell 72:223-232). Examples of suitable interaction (or dimerization) domains include leucine zippers (Landschulz et al, 1989, Science 243:1681-1688), helix-loop-helix domains (Murre, C. et al, 1989, Cell 58:537-544) and zinc finger domains (Frankel, A. D. et al, 1988, Science

²⁰ 240:70-73).

In another aspect, the polypeptide that binds P_ptr can be used by itself to repress transcription in the absence of antibiotic. In this manner, the polypeptide that binds P_ptr prevents transcription when bound to the P_ptr sequence, presumably by interfering with

2 binding of activating transcription factors. In the absence of antibiotic, transcription from the P_ptr linked promoter is absent or minimal. When antibiotic is added, however, the polypeptide that binds P_ptr is released, thereby allowing transcription to occur.

In an alternative embodiment, the fusion protein that binds P_ptr can be used to repress transcription. In one aspect, the first polypeptide is operatively linked, as described above, to

30 a second polypeptide which directly or indirectly represses transcription in eukaryotic cells.

Proteins and polypeptide domains within proteins which can function to repress transcription in eukaryotic cells have been described in the art (for reviews see, e.g., Renkawitz, R., 1990,

Trends in Genetics 6:192-197; and Herschbach, B. M. and Johnson, A. D., 1993, Annu. Rev.

~_. Cell. Biol. 9:479-509). Such domain can have a direct inhibitory effect on the transcriptional machinery or can repress transcription indirectly by inhibiting the activity of activator proteins. Accordingly, the term "a polypeptide that represses transcription in eukaryotic cells" as used herein is intended to include polypeptides which act either directly or indirectly to repress transcription. As used herein, "repression" of transcription is intended to mean a diminution in the level or amount of transcription of a target gene compared to the level or amount of transcription prior to regulation by the transcriptional inhibitor protein. Transcriptional inhibition may be partial or complete.

A transcriptional "repressor" or "silencer" domain as described herein is a polypeptide domain that retains its ability to repress transcription when the domain is transferred to a heterologous protein. Proteins which have been demonstrated to have repressor domains that can function when transferred to a heterologous protein include the v-erbA oncogene product

(Baniahmad, A. et al, 1992, EMBO J. 11:1015-1023) (e.g., approximately amino acid residues 362-632 of the native v-erbA oncogene product), the thyroid hormone receptor (Baniahmad, supra), the retinoic acid receptor (Baniahmad, supra), the Drosophila Krueppel

(Kr) protein (Licht, J. D. et al, 1990, Nature 346:76-79; Sauer, F. and Jackie, H., 1991,

Nature 353:563-566; Licht, J. D. et al, 1994, Mol. Cell. Biol. 14:4057-4066) (such as

C64KR, which is amino acids 403-466 of the native protein, or amino acids 26-110 of Kr), and the KRAB domain of the koxl gene family (Deuschle et al, 1995, Mol. Cell. Biol.

15:1907-1914). Other proteins which have transcri •pti •onal repressor activity in eukaryotic cells include the Drosophila homeodomain protein even-skipped (eve) (Han and Manley,

1993, Genes & Dev. 7: 491-503), the S. cerevisiae Ssn6/Tupl protein complex (Herschbach and Johnson, supra), the yeast SIRI protein (see Chien et al, 1993, Cell 75:531-541), NePl (see Kohne et al, 1993, J Mol. Biol. 232:747-755), the Drosophila dorsal protein (see Kirov et al, 1994, Mol. Cell. Biol. 14:713-722; Jiang, et al, 1993, EMBO J. 12:3201-3209), TSF3 (see Chen, et al, 1993, Mol. Cell. Biol. 13:831-840), SF1 (see Targa, et al, 1992, Biochem. Biophys. Res. Comm.. 188:416-423), the Drosophila hunchback protein (see Zhang, et al, 1992, Proc. Natl. Acad. Sci. USA 89:7511-7515), the Drosophila knirps protein (see Gerwin, et al, 1994, Mol. Cell. Biol. 14:7899-7908), the WT1 protein (Wilm's tumor gene product) (see Anant, et al, 1994, Oncogene 9:3113-3126; Madden et al„ 1993, Oncogene 8:1713-1720), Oct-2.1 (see Lillycrop, et al, 1994, Mol. Cell. Biol. 14:7633-7642), the Drosophila engrailed protein (see Badiani, et al, 1994, Genes Dev. 8:770-782; Han and Manley,, 1993, EMBO J. 12:2723-2733), E4BP4 (see Cowell and Hurst,, 1994, Nucleic Acids Res. 22:59-65) and ZF5 (see Numoto, et al, 1993, Nucleic Acids Res. 21:3767-3775). Non-limiting examples of polypeptide domains that can be used as silencing domains include: amino acid residues 120-410 of the thyroid hormone receptor alpha (THRalpha.), amino acid residues 143-403 of the retinoic acid receptor alpha (RARalpha.), amino acid 5 residues 186-232 of knirps, the N-terminal region of WT 1 (see Anant, supra), the N-terminal region of Oct-2.1 (see Lillycrop, supra), a 65 amino acid domain of E4BP4 (see Cowell and Hurst, supra) and the N-terminal zinc finger domain of ZF5 (see Numoto, supra). Moreover, shorter or longer polypeptide fragments encompassing these regions that still retain full or partial repression activity are also contemplated.

10 . .

In addition to previously described transcriptional repressor domains, novel transcriptional repressor domains, which can be identified by standard techniques (e.g., reporter gene constructs), are within the scope of the invention.

Construction of the nucleic acids of the invention can be accomplished by those of

25 skill in the art using standard molecular biology techniques (see, for example, Maniatis et al. ,

1989, Molecular Cloning A Laboratory Manual, Cold Spring Harbor Laboratory, N.Y. and

Ausubel et al, 1989, Current Protocols in Molecular Biology, Greene Publishing Associates and Wiley Interscience, N.Y.). Typically, manipulation and generation of nucleic acids is performed using prokaryotic host cells, but the invention also encompasses chemical

20 synthetic methods of nucleic acid generation and manipulation.

The host cells for use in the methods and compositions of the invention can also be any eukaryotic cell such as mammalian cells, fungal cells, plant cells and microbial parasites.

The invention is illustrated below by way of non-limiting examples using a variety of

25 mammalian cells. However, other types of eukaryotic cells can also be used. For example, the tetracycline-regulatable promoter system has been adapted for use in yeast (Gari et al, 1997, Yeast 13:837-848); similarly, the methods and compositions of the invention can also be used in such cells. Suitable host cells include, for example, mammalian cells such as CHO-Kl, BHK-21, HeLa , COS-7, HEK 293, HT1080, PC12, MDCK, C2C12, Jurkat,

³⁰ NIH3T3, K-562, TF-1, P19, or plant cells such as those derived from barley, wheat, rice, soybean, potatoe, tobacco (e.g. Nicotiana tabacum SRI) and arabidopsis. Suitable hosts also include plant-derived hairy roots such as those derived from Artemisia, Atropa, Beta, Brugmansia and others such as those described in Shanks and Morgan, 1999, Curr. Opin.

- _. Biotechnol. 10:151-155. Other suitable cell lines of mammalian and plant origin are well known to those of skill in the art and include, for example, those described in ATCC Cell Lines and Hybridomas 8^th Edition, 1994, American Type Culture Collection, Rockville, MD. Methods of genetically engineering a host cell to contain the nucleic acids of the 5 invention are well known to those of skill in the art and include transformation, transfection, and electroporation. The nucleic acids can be carried extracliromasomally or on the chromosome. Integration can be random, homologous, or site-specific recombination. Culturing a host cell is understood to include both in vitro culture and in vivo culture (for example, growing eukaryotic cells in tissue culture, growing cells in a host organism such as by implantation in a body cavity or graft, removing cells from a particular individual and replacing them after genetically engineering the cells to contain the nucleic acids of the invention, etc.).

Furthermore, the present invention provides non-human transgenic animals having 25 cells comprising nucleic acids encoding the proteins and fusion proteins of the invention. Such host cells optionally contain a desired nucleotide sequence to be transcribed operatively linked to a P_ptr sequence. The non-human transgenic animals contemplated by the present invention generally include any vertebrates, and preferably mammals. Such nonhuman transgenic animals may include, for example, transgenic pigs, transgenic rats, transgenic

20 rabbits, transgenic cattle, transgenic goats, and other transgenic animal species, particularly mammalian species, known in the art. Additionally, bovine, ovine, and porcine species, other members of the rodent family, e.g. rat, as well as rabbit and guinea pig, and non-human primates, such as chimpanzee, may be used to practice the present invention. Particularly 2 preferred animals are rats, rabbits, guinea pigs, and most preferably mice.

Detailed methods for generating non-human transgenic animals are known to those of skill in the art and include, for example, those described in Hogan et al, 1994, Manipulating the mouse embryo, 2nd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor,

New York. ^3υ In an exemplary embodiment, the "transgenic non-human animals" of the invention are produced by introducing transgenes into the germline of the non-human animal.

Embryonal target cells at various developmental stages can be used to introduce transgenes.

Different methods are used depending on the stage of development of the embryonal target „_. cell. The specific line(s) of any animal used to practice this invention are selected for general good health, good embryo yields, good pronuclear visibility in the embryo, and good reproductive fitness. In addition, the haplotype is a significant factor. Introduction of a transgene into an embryo can be accomplished by any means known 5 in the art such as, for example, micromjection, electroporation, or lipofection. For example, a construct can be introduced into a mammal by micromjection of the construct into a mammalian oocyte to cause one or more copies of the construct to be retained in the cells of the developing mammal. Following introduction of the transgene construct into the oocyte, the oocyte may be incubated in vitro for varying amounts of time, or implanted into a

10 surrogate host, or both. In vitro incubation to maturity is within the scope of this invention.

One common method is to incubate embryos in vitro for about 1-7 days, depending on the species, and then reimplant them into a surrogate host.

The progeny of transgenically manipulated embryos can be tested for the presence of 5 the construct by any means known to those of skill in the art including, for example, Southern blot analysis or PCR analysis of a segment of tissue. If one or more copies of the exogenous cloned construct is stably integrated into the genome of such transgenic embryos, it is possible to establish permanent transgenic mammallian lines carrying the transgenically added construct.

20 The litters of transgenically altered mammals can be assayed after birth for the incorporation of the construct into the genome of the offspring. Preferably, this assay is accomplished by PCR analysis or by hybridizing a probe corresponding to the DNA sequence coding for the desired recombinant protein product or a segment thereof onto chromosomal

25 material from the progeny. Those mammalian progeny found to contain at least one copy of the construct in their genome are grown to maturity.

The number of copies of the transgene constructs which are added to the oocyte is dependent upon the total amount of exogenous genetic material added and will be the amount which enables the genetic transformation to occur. Theoretically only one copy is required;

3 ^D0^J however, generally, numerous copies are utilized, for example, 1,000-20,000 copies of the transgene construct, in order to insure that one copy is functional.

Any technique which allows for the addition of the exogenous genetic material into nucleic genetic material can be utilized so long as it is not destructive to the cell, nuclear

_ _ membrane or other existing cellular or genetic structures. The exogenous genetic material is preferentially inserted into the nucleic genetic material by micromjection. Micromjection of cells and cellular structures is known and is used in the art. Reimplantation is accomplished using standard methods. Usually, the surrogate host 5 is anesthetized, and the embryos are inserted into the oviduct. The number of embryos implanted into a particular host will vary by species, but will usually be comparable to the number of off spring the species naturally produces.

Transgenic offspring of the surrogate host may be screened for the presence and/or expression of the transgene by any suitable method. Screening is often accomplished by Southern blot or Northern blot analysis, using a probe that is complementary to at least a portion of the transgene. Western blot analysis using an antibody against the protein encoded by the transgene may be employed as an alternative or additional method for screening for the presence of the transgene product. Typically, DNA is prepared from tail tissue and analyzed 25 by Southern analysis or PCR for the transgene. Alternatively, the tissues or cells believed to express the transgene at the highest levels are tested for the presence and expression of the transgene using Southern analysis or PCR, although any tissues or cell types may be used for this analysis.

Alternative or additional methods for evaluating the presence of the transgene include,

20 without limitation, suitable biochemical assays such as enzyme and/or immunological assays, histological stains for particular marker or enzyme activities, flow cytometric analysis, and the like. Analysis of the blood may also be useful to detect the presence of the transgene product in the blood, as well as to evaluate the effect of the transgene on the levels of various

2 types of blood cells and other blood constituents.

Progeny of the transgenic animals may be obtained by mating the transgenic animal with a suitable partner, or by in vitro fertilization of eggs and/or sperm obtained from the transgenic animal. Where mating with a partner is to be performed, the partner may or may not be transgenic and/or a knockout; where it is transgenic, it may contain the same or a

^3υ different transgene, or both. Alternatively, the partner may be a parental line. Where in vitro fertilization is used, the fertilized embryo may be implanted into a surrogate host or incubated in vitro, or both. Using either method, the progeny may be evaluated for the presence of the transgene using methods described above, or other appropriate methods.

„ A variety of vectors may be used to engineer host cells and transgenic animals to contain the nucleic acids of the invention. Suitable mammalian expression vectors include pSG5, pCMV-Script, pcDNA3.1, pcDNA4 series, pEFl, pBK-CMV, pBK-RSV, pSBC-1, pSBC-2. Suitable viral vectors include adenoviral vectors, adeno-associated viral vectors, retroviral vectors, lentiviral vectors and other viral vectors known to those of skill in the art. Suitable retroviral and/or lentiviral vectors include, for example, pLAPSN, pLHCX, pLIB, pLNCX2, pLNHX, pLBCX, pLXRN, pLXSN, pMSCVneo, pSIR and lentiviral vectors such as those described by Naldini and Nerma, 1999, 47-60 (and references therein) in Friedman (ed.) The development of human gene therapy. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York). Suitable plant expression vectors include, for example, pBI, pRT99, pGPTV series, pGSN4, pBI 121, pBI221, pBHOl, pKINI105, pBIΝ19, pPZP121, and the pCAMBIA family). Additional suitable vectors will be apparent to those of skill in the art and inlcude, for example, those described in ATCC Catalogue of Recombinant DNA Materials 3^rd Edition, 1993, American Type Culture Collection, Rockville, MD. A P_ptr -linked gene is defined herein as a promoter which directs the expression of a coding sequence, wherein the promoter and coding sequence are operatively linked to a P_ptr sequence. Preferably, the promoter is a eukaryotic promoter. By operatively linked in this context is meant that the P_ptr sequence is placed proximal to the promoter, either 5' to, or 3' to, or within the sequence of the promoter, such that when a fusion protein that modulates transcription is bound to the P_ptr sequence, initiation of transcription at the promoter is affected.

The coding sequence operatively linked to the promoter can encode for any gene product for which regulated expression is desired, and can be exogenous to the host cell or endogenous. By endogenous coding sequence is meant coding sequence that is naturally present in the host cell and not introduced into the host cell via transformation techniques. Exogenous coding sequence is not endogenous coding sequence. Coding sequences include not only sequences encoding proteins, but also other coding sequences, e.g., encoding antisense gene products, ribozymes, etc. Further, the coding sequences can be multicistronic (see, for example, U.S. Application No. 08/948,381, filed October 9, 1997).

Production of any gene product can be regulated using the compositions and methods of the present invention. For example, production of a marker gene product, such as green fluorescent protein, or of a model secreted gene product, such as SEAP, may be regulated. Naturally, the invention finds particular use in the production of products of industrial or pharmaceutical interest such as industrial enzymes (e.g. proteases, cellulases, glycosidases, or ligninases), interferons (e.g. β-INF, α-INF, γ-INF), hGH, insulin, erythropoietin, tissue plasminogen activator (tPA), DNAse, monoclonal antibodies, Factor VIII, Factor Nil, Factor IX, HSA, IL-2, glucagon, EGF, GCSF, GMCSF, thrombopoietin, gpl60, HbSAg, and other viral antigenic proteins and peptides (rotavirus, HIN, p53ras).

Another aspect of the invention is the use of tumor suppressor gene products to regulate proliferation of the host cells of the invention. Regulated expression of tumor suppressor gene products are particularly useful for a variety of applications. For example, one may want the host cells to undergo a rapid proliferation phase followed by a production phase where cellular energies are devoted to protein production, or a rapid proliferation phase in vitro followed by regulated growth in vivo (see, for example, U.S. Application No. 08/948,381, filed October 9, 1997, the disclosure of which is incorporated by reference). For purposes of the invention, tumor suppressor gene products are intracellular proteins that block the cell cycle at a cell cycle checkpoint by interaction with cyclins, Cdks or cyclin-Cdk complexes, or by induction of proteins that do so. Thus, these tumor suppressor gene products inhibit the cyclin-dependent progression of the cell cycle. Particularly preferred tumor suppressor gene products act on the Gl-S transition of the cell cycle. The invention encompasses the use of any tumor suppressor gene product which performs this function, whether known or yet to be discovered. Examples of tumor suppressor genes include p21, p27, p53 (and particularly, the p53175P mutant allele), p57, pl5, pl6, pl8, pl9, p73,

GADD45 and APCl.

Optionally, one can also use the methods and compositions of the invention to express survival factors in the host cells. Survival factors are intracellular proteins that prevent apoptosis such as bcl-2, bcl-x_L, E1B-19K, mcl-1, crmA, abl, p35, bag-1, A20, LMP-1, Tax,

Ras, Rel and NF-κB-like factors. Additionally, all known survival factors, as well as survival factors yet to be discovered, are useful in the methods and compositions of the invention. In yet another embodiment, the tumor suppressor gene(s) is expressed concomitantly with a factor that stabilizes the tumor suppressor gene product in the cell. Examples of stabilizing factors are members of the CAAT enhancer binding protein family. For example, p21 protein activity is stabilized when coexpressed with C/EBPα. Additionally, C/EBPα specifically induces transcription of the endogenous p21 gene. Thus, C/EBPα functions as both a stabilizing factor and as a specific inducer of p21.

Still another aspect of the invention is the use of the nucleotides and methods of the invention to express a gene product that activates cell proliferation. For example, a protein that activates cell proliferation is Mekl, a central protein kinase in the conserved mammalian Ras-MAP signal transduction pathway responding to growth-promoting signals such as 5 cytokines. A particularly preferred version of Mekl is the MeklDD mutant (Gruelich and Erikson, 1998, J. Biol. Chem. 273: 13280-13288) described more fully below. Other genetic determinants exerting positive control of mammalian cell cycle that can be used as a protein that activates cell proliferation are cyclins (e.g., cyclin E), Ras, Raf, the MAP kinase family (e.g., MAP, Erk, Sap) E2F, Src, Jak, Jun, Fos, pRB, Mek2, EGF, TGF, PDGF, and a polynucleotide that is antisense to a tumor suppressor gene (e.g., p27 anti-sense expression has been shown to stimulate proliferation of quiescent fibroblasts and enable growth in serum-free medium (Rivard et al , 1996, J. Biol. Chem. 271: 18337-18341.) and nedd5 which is known as positive growth controlling gene (Kinoshita et al , 1997, Genes Dev. 25 11: 1535-1547).

One aspect of the present invention is illustrated below by the development of a new system for antibiotic-regulated gene expression in eukaryotic cells based on the repressor of a streptogramin resistance operon of S. coelicolor (a Pip). A chimeric protein (PIT) comprised of Pip fused to a eukaryotic transactivator was able to control expression of a synthetic

20 eukaryotic promoter (P_PIR) contaimng the Pip binding site. Genes placed under the control of this PιT/P_prR system were responsive to clinically approved therapeutic compounds belonging to the streptogramin group (pristinamycin, virginiamycin and Synercid ) in a variety of mammalian cell lines (CHO-Kl, BHK-21 and HeLa) and plant cells (Nicotania tabacum SR-

25 1). This novel system exhibited superior inducibility and background expression properties compared to the well-established tetracycline-based system in CHO cells engineered to provide both streptogramin and tetracycline regulation. In these cells, streptogramin does not affect gene expression from the tetracycline-responsive promoter, and therapeutically relevant tetracycline concentrations have only minor effect on expression from the streptogramin-

3 ^D0 responsive PIT/P_PIR system. Therefore, these two different systems can be used together in advanced future therapies requiring independent regulation of different transgenes. In addition, responsiveness of the PIT/P_PIR system to all type B streptogramins tested indicates that reporter gene expression from P_PIR can be used as an efficient high-throughput assay for ~_<- discovery of new streptogramins. The same concept applies to other eukaryotic antibiotic- responsive transcription regulation systems which can be linked to a reporter gene for the discovery of new antibiotics. Another aspect of the invention is the development of a double-regulation system 5 particularly advantageous when independent control of at least two different transgenes or sets of transgenes is desired. For example, this aspect of the invention is illustrated by way of a non-limiting example using the combined installation of tetracycline-dependent and pristinamycin-responsive gene regulation in the same cells. In the examples below, a double regulation vector pDuoRexl was constructed which contains both tetracycline-dependent and

10 pπstmamycm-responsive expression units. These units were used, for example, to express the tumor suppressor gene p27 (under the control of the tetracycline-responsive promoter) and a gene product stimulating cell proliferation such as Mekl (here under the control of P_PIR)

(pMF195). For advanced gene therapy and tissue engineering strategies, induction of Mekl

25 (while repressing p27) can activate proliferation of grafted cells and facilitate cell expression while induction of p27 (while repressing Mekl) will arrest cell proliferation which is a desired trait for reimplantation of engineered cells and tissues. Other tumor suppressor genes and other genes stimulating cell proliferation can be substituted.

Still another aspect of the invention are multipurpose expression vectors, as well as

20 cells and methods using the same, which take advantage of the antibiotic dependent activator and repressor systems of the invention. Such vectors can be mono, di- or multicitronic. Non- limiting examples of such vectors are described below by way of working embodiments. Although the regulated gene expression invention described herein was originally 25 designed for general applications in functional genomic research, gene therapy and tissue engineering, the finding that the streptogramin system illustrated below by way of nonlimiting example responds to all commercially available streptogramins, including Synercid, Pyostacin^®, and virginiamycin, indicates its use as a powerful screening tool for the discovery of novel antibiotics. Accordingly, still another aspect of the invention is a method

3 ^J0^υ of screening for candidate antibiotics. The method entails incubating the host cells of the invention, the host cells containing a P_ptr -linked reporter gene and a sequence encoding a P_ptr -binding protein, in the presence of a test compound, wherein a change in the transcription of the reporter gene indicates that the test compound is a candidate antibiotic. ^ _ For example, detection of streptogramins is based on addition of metabolic libraries of

Streptomyces or fungal origin to cultured mammalian cells containing the pristinamycin- responsive reporter system. The presence of streptogramins will downregulate expression of the reporter protein, for example SEAP, driven by P_prR. This screening approach offers two decisive advantages over classical screening technology using indicator bacteria-based antibiogram tests: (i) Antibiotic screening is not limited by the sensitivity of indicator bacteria to a yet uncharacterized streptogramin (sensitivity of bacteria to antibiotics greatly varies between strains and even isolates), and (ii) the mammalian cell-based streptogramin detection concept shows at least one order of magnitude higher sensitivity to this class of antibiotics than antibiogram tests based on bactericidal activity. However, given the suggested interaction of Pip with other classes of antibiotics (a list is given in Salah-Bey K, Blanc, N., and Thompson C.J, 1995; Mol. Microbiol. 17, 1001-1012) the Pip/P_PIR-based mammalian gene regulation system is likely to detect antibiotic compounds in general. Also, extension of this detection concept to include antibiotic-responsive reporter gene expression using other PIP and P_ptr components from other Actinomycetes is within the scope of this invention.

The invention having been described, the following examples are offered by way of illustration and not limitation.

Example 1: Steptogramin-repressible Mammalian Gene Regulation System

After showing that the presence of PI reverses Pip association with P_PTR in vitro, this example describes the use of Pip and P_PTR to design a novel streptogramin-responsive mammalian gene regulation system with excellent regulatory properties which is functionally compatible with the tetracycline-based transcription regulation system (TET system) most widely used presently.

METHODS

Construction of the streptogramin-dependent transactivator (PIT) and the pristinamycin- regulatable promoter P_PIR PIT, the fusion protein of Pip and the NP16 C-terminal transactivation domain of

Herpes simplex (Triezenberg et al, 1988, Genes Dev. 2, 718-729) was constructed by amplifying the S. coelicolor Pip from pGemT::Epip4 with oligos:

OMF63: GTACGAATTCCCACCATGAGTCGAGGAGAGGTG (SEQ ID ΝO:l) and OMF64: GCGCGCGGCTGTACGCGGAGGCCTGTTCGACCATC (SEQ ID NO:2) followedby cloning into vector pcDNA3.1/V5/His-TOPO (Invitrogen) under control of P_CMV(pMF150).

These primers can also be used to clone the desired gene sequences from S. pristinaespiralis genomic DNA. P_PIR seqence (SEQ ID NO:3): GATCGACGTCGGAGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGTGTACGAGTA TCTTCCCGTACACCGTACAAGGAGCCTGCAGGgagtaccctcgaccgccgGAGTATAAATAG AGGCGCTTCGTCTACGGAGCGACAATTCAATTCAAACAAGCAAAGTGAACACGTCGCTAAGC GAAAGCTAAGCAAATAAACAAGCGCAGCTGAACAAGCTAAACAATCTGCAGTAAAGTGCAAG TTAAAGTGAATCAATTAAAAGTAACCAGCAACCAAGTAAATCAACTGCAACTACTGAAATCT GCCAAGAAGTAATTATTGAATACAAGAAGAGAACTCTGAATACTTTCAACAAGTTACCGAGA AAGAAGAACTCACACACAGCTAGCGTTTAAACTTAAGCTTGGTACCGAGCTCGGATCCACTA GTCCAGTGTGGTGGAATTC.

The Pip-containing SspVBssHIl fragment of pMF150 was subsequently cloned into the corresponding sites of ρSBC2-tTA (pSAM200 (SspVBssHIT) (Fussenegger et al, 1997, Biotechnol. Prog. 13, 733-740) thereby replacing the TetR Domain of tTA by Pip. The resulting plasmid has the following PIT coding sequence: PIT sequence (SEQ ID NO:4): ATGAGTCGAGGAGAGGTGCGCATGGCGAAGGCAGGGCGGGAGGGGCCGCGGGACAGCGTGTG GCTGTCGGGGGAGGGGCGGCGCGGCGGTCGCCGTGGGGGGCAGCCGTCCGGGCTCGACCGGG ACCGGATCACCGGGGTCACCGTCCGGCTGCTGGACACGGAGGGCCTGACGGGGTTCTCGATG CGCCGCCTGGCCGCCGAGCTGAACGTCACCGCGATGTCCGTGTACTGGTACGTCGACACCAA GGACCAGTTGCTCGAGCTCGCCCTGGACGCCGTCTTCGGCGAGCTGCGCCACCCGGACCCGG ACGCCGGGCTCGACTGGCGCGAGGAACTGCGGGCCCTGGCCCGGGAGAACCGGGCGCTGCTG GTGCGCCACCCCTGGTCGTCCCGGCTGGTCGGCACCTACCTCAACATCGGCCCGCACTCGCT GGCCTTCTCCCGCGCGGTGCAGAACGTCGTGCGCCGCAGCGGGCTGCCCGCGCACCGCCTGA CCGGCGCCATCTCGGCCGTCTTCCAGTTCGTCTACGGCTACGGCACCATCGAGGGCCGCTTC CTCGCCCGGGTGGCGGACACCGGGCTGAGTCCGGAGGAGTACTTCCAGGACTCGATGACCGC GGTGACCGAGGTGCCGGACACCGCGGGCGTCATCGAGGACGCGCAGGACATCATGGCGGCCC GGGGCGGCGACACCGTGGCGGAGATGCTGGACCGGGACTTCGAGTTCGCCCTCGACCTGCTC GTCGCGGGCATCGACGCGATGGTCGAACAGGCCTCCGCGTACAGCCGCGCGCGTACGAAAAA CAATTACGGGTCTACCATCGAGGGCCTGCTCGATCTCCCGGACGACGACGCCCCCGAAGAGG CGGGGCTGGCGGCTCCGCGCCTGTCCTTTCTCCCCGCGGGACACACGCGCAGACTGTCGACG GCCCCCCCGACCGATGTCAGCCTGGGGGACGAGCTCCACTTAGACGGCGAGGACGTGGCGAT GGCGCATGCCGACGCGCTAGACGATTTCGATCTGGACATGTTGGGGGACGGGGATTCCCCGG GTCCGGGATTTACCCCCCACGACTCCGCCCCCTACGGCGCTCTGGATATGGCCGACTTCGAG TTTGAGCAGATGTTTACCGATGCCCTTGGAATTGACGAGTACGGTGGGTAG.

PIT2, the fusion protein of Pip and the p65 transactivation domain of the human NF-

KB transcription activator was cloned in a three step procedure: (i) The p65 domain of the human NF- B transcription activator was amplified from pSG424-p65 (kindly provided by

H. Hauser, GBF, Braunschweig, Germany) with oligos OMF97:

GATCGCGCGCATGATGAGTTTCCCACC (SEQID NO:22) and OMF98:

GATCAAGCTTGGATCCTTAGGAGCTGATCTGACT (SEQIDNO:23) and clonedin sense orientation into pcDNA3.1/V5/His-TOPO (Invitrogen) to result in plasmid pMF197. (ii) pMF197 was restricted with R-?-?HII and H dIII contained in the oligos (underlined) and the p65 domain was ligated to the corresponding sites (R-?-?HII/Ht/-dIII) of pSBC2-tTA

(Fussenegger et al., 1997) thereby replacing the NP16 domain of the tetracycline-dependent transactivator and resulting in plasmid pMF204. (iii) The tetR domain was excised from pMF204 by Sspl/BssΗlI and replaced by Pip excised from pMF150 by SspVBssBIl to give plasmid pMF206. pMF206 contains the following PIT2 coding sequence:

ATGAGTCGAGGAGAGGTGCGCATGGCGAAGGCAGGGCGGGAGGGGCCGCGGGACAGCGTGTG

GCTGTCGGGGGAGGGGCGGCGCGGCGGTCGCCGTGGGGGGCAGCCGTCCGGGCTCGACCGGG ACCGGATCACCGGGGTCACCGTCCGGCTGCTGGACACGGAGGGCCTGACGGGGTTCTCGATG CGCCGCCTGGCCGCCGAGCTGAACGTCACCGCGATGTCCGTGTACTGGTACGTCGACACCAA GGACCAGTTGCTCGAGCTCGCCCTGGACGCCGTCTTCGGCGAGCTGCGCCACCCGGACCCGG ACGCCGGGCTCGACTGGCGCGAGGAACTGCGGGCCCTGGCCCGGGAGAACCGGGCGCTGCTG GTGCGCCACCCCTGGTCGTCCCGGCTGGTCGGCACCTACCTCAACATCGGCCCGCACTCGCT GGCCTTCTCCCGCGCGGTGCAGAACGTCGTGCGCCGCAGCGGGCTGCCCGCGCACCGCCTGA

CCGGCGCCATCTCGGCCGTCTTCCAGTTCGTCTACGGCTACGGCACCATCGAGGGCCGCTTC

CTCGCCCGGGTGGCGGACACCGGGCTGAGTCCGGAGGAGTACTTCCAGGACTCGATGACCGC

GGTGACCGAGGTGCCGGACACCGCGGGCGTCATCGAGGACGCGCAGGACATCATGGCGGCCC GGGGCGGCGACACCGTGGCGGAGATGCTGGACCGGGACTTCGAGTTCGCCCTCGACCTGCTC GTCGCGGGCATCGACGCGATGGTCGAACAGGCCTCCGCGTACAGCCGCGCGCATGATGAGTT TCCCACCATGGTGTTTCCTTCTGGGCAGATCAGCCAGGCCTCGGCCTTGGCCCCGGCCCCTC CCCAAGTCCTGCCCCAGGCTCCAGCCCCTGCCCCTGCTCCAGCCATGGTATCAGCTCTGGCC CAGGCCCCAGCCCCTGTCCCAGTCCTAGCCCCAGGCCCTCCTCAGGCTGTGGCCCCACCTGC CCCCAAGCCCACCCAGGCTGGGGAAGGAACGCTGTCAGAGGCCCTGCTGCAGCTGCAGTTTG ATGATGAAGACCTGGGGGCCTTGCTTGGCAACAGCACAGACCCAGCTGTGTTCACAGACCTG GCATCCGTCGACAACTCCGAGTTTCAGCAGCTGCTGAACCAGGGCATACCTGTGGCCCCCCA CACAACTGAGCCCATGCTGATGGAGTACCCTGAGGCTATAACTCGCCTAGTGACAGGGGCCC AGAGGCCCCCCGACCCAGCTCCTGCTCCACTGGGGGCCCCGGGGCTCCCCAATGGCCTCCTT TCAGGAGATGAAGACTTCTCCTCCATTGCGGACATGGACTTCTCAGCCCTGCTGAGTCAGAT CAGCTCCTAA (SEQ ID NO:24).

For construction of P_PΪR the minimal promoter of the Drosophila heat-shock gene hsp70 P_hsp70min; bp -40 to 241; Corces et al, 1984, J. Biol. Chem. 259, 14812-14817) contained on pTRIDENT7 (Fussenegger et al, 1998, Biotechnol. Bioeng. 57, 1-10) was amplified by PCR using oligos OMF62:CTTGTACGGTGTACGAGTATCTTCCCGTACACCGTACAAGGAGCCTGCAGGgagta ccctcgaccgccg (SEQ ID NO:5) and OMF57: GATCCATCGATTGATCAGGCGC (SEQ ID NO: 6) and cloned in anti-ZαcZ orientation into the pCR2.1-TOPO vector (Invitrogen) to give plasmid pMF153. The premature P_PIR sequence contained in pMF153 was subjected to a second round of PCR using primers OMF69:

GATCGACGTCGGAGAAATAGCGCTGTACAGCGTATGGGAATCTcttgtacggtgtacgagta tc (SEQ ID NO:7) and OMF57 described above. The resulting PCR fragment containing the complete P_PIR sequence was cloned into pCR2.1-TOPO in lacZ orientation to give pMF161.

The annealing sequence of OMF69 (small letters) is complementary to the extension of OMF62 (capital letters) and adds the remaining Pip binding site of P_PTR to P_PIR.

Construction of the pristinamycin-responsive reporter plasmids pMF 164 andpMF172 and the pTWIN vectors

For the construction of pMF164 the P_PIR-containing fragment was excised from pMF161 by Aatll andEcoRI and ligated to the corresponding sites (AatTUEcoRI) of the GFP expression vector pMF104, thereby replacing the TΕT-responsive promoter P_hCMv_*-ι

(Fussenegger et al, 1998, supra). Similarly, pMF172 was constructed for quantitative analysis of P_P]R-mediated expression, by exchanging P_hCMV*.ι of pMFlll (Fussenegger et al,

1997, Biotechnol. Bioeng. 55, 927-939) by P_PIR of pMF164 via Sspl and EcøRI. pTWIN contains both PIT and tTA in a P_sv40-driven dicistronic configuration. For construction of pTWIN, PIT was first transferred from pMF156 as an EcoRI/HindlU fragment to the corresponding sites of pSBC-1 (Ecd UHindΩl; resulting in pMF167) before the PIT expression unit could be fused to the tTA expression unit contained on pSBC2-tTA (Fussenegger et al, 1997, supra) via SspllNotl sites. Thus, pTWIN contains the dicistronic Psv₄o-PIT-IRES-tTA-pA expression cassette.

Cell culture, transfection, construction of stable cell lines, SEAP activity test and gel retardation assays

Chinese hamster ovary cells (CHO-Kl, ATCC: CCL 61), Baby hamster kidney cells

(BHK-21, ATCC: CRL-8544), and HeLa cells (ATCC: CRL-7923) as well as stable cell lines CHO-PIT1, CHO-PIT2, CHO-TWINl₉₅, CHO-TWTN1₁₀₈ and CHO-PIT1-SEAP were cultured as described before (Fussenegger et al, 1998, Nat. Biotechnol. 16,468-472) in the presence of appropriate antibiotics: G418 400 μg/ml; zeocin 100 μg/ml. Optimized CaPO₄ protocols were used for high efficiency transient transfection of all cell lines (id). Transient transfected cells were routinely analyzed after 48 h for GFP or SEAP expression using fluorescence microscopy and p-nitrophenolphosphate-based light absorbance timecourse, respectively, as described before (Fussenegger et al, 1998, Nat. Biotechnol. 16, 468-472;

Berger et al, 1988, Gene 66, 1-10 b). For construction of stable CHO-PIT and CHO-TWIN cell lines, CHO-Kl was cotransfected either with pMF156 or pTWTNl and pSV27.eo carrying the G418 resistance gene. CHO-PIT 1 -SEAP was constructed by cotransfection of pMF172 and pZeoSN2 (Invitrogen). For assessment of dose-response characteristics of P_PIR-regulated gene expression, CHO-PITl-SEAP were cultured at cell densities of 150,000/ml for 48 at various PI concentrations. The mixed populations were cloned using FACS-mediated single- cell-sorting (FACStar*¹¹¹¹⁵; Beckton Dickinson). Gel retardation assay were performed as described (Salah-Bey et al, 1995, supra) using Pip purified from an overproducing strain of E. coli. Pip (0.02pmol) was titrated with a range of antibiotic concentrations in a reaction volume of 20 μl. Protein solutions were pre-incubated at room temperature for 15 min. in the absence or presence of PI or its derivative quinupristin in a buffer containing 10 mM Tris (pH 7.8), 10 mM MgCl₂, 300mM Νacl, 2mM DTT and 10% glycerol. An Ecom-Hindll P_PTO fragment (0.02pmol) containing three Pip binding sites (radiolabelled using α~³²P)dATP was added and incubated for 15 min. To minimize non-specific protein binding and increase protein stability, 2 μg of poly dldC and 20 μg BSA were included in the incubation mix. Protein-DΝA complexes were separated on 5% polyacrylamide, TBE gels and visualised by autoradiography. Regulating streptogramins Five different sources of pristinamycin were used: The two therapeutic forms Pyostacin™ and Synercid™ , pristinamycin discs, and the single compounds pristinamycin I (PI) and pristinamycin II (PII). Pyostacin™ pills (500 mg) were ground in a mortar and solubilized in DMSO or water at a stock concentration of 50 mg/ml. Synercid™ (RP 59500; Lot Nr. CB06253) was provided by Aventis in injection-ready vials containing 500 mg lyophilized antibiotic which was reconstituted in 5 ml of 5% glucose solution and frozen at -

20°C. Individual pristinamycin compounds PI and PII (identical to virginiamycin M. (VM_j);

Sigma ref. V2753) were solubilized in DMSO at a concentration of 50 mg/ml and kept at -

20°C. Antibiotic discs containing 15 μg pristinamycin or virginiamycin (bioMerieux, ref.

54552; bioMerieux, ref. 54612) were soaked in 1 ml cell culture medium at 4jC for up to 24 h. Nirginiamycin consists of virginiamycin Sj (VS_), the PI analogue, and virginiamycin M_, which is identical in structure to PII (Barri re et al, 1994, supra). 2 μg/ml of streptogramin B was routinely used for regulation studies in cell culture. The concentration of individual streptogramin components was calculated based on their fixed 70:30 ratio (w/w;

PII/NM dalfopristin : PI/NS/quinupristin). For control experiments TET and doxycycline

(DOX) were used either as discs (30 μg/disc, bioM_rieux ref. 54882 (TET) and ref. 54172

(DOX)) or as powder (Sigma refs. T3383 and D9891).

RESULTS Construction and functional studies of the streptogramin-based gene regulation system in CHO cells

Gel retardation assays using purified Pip protein and a DNA fragment containing P_PTR target sequence indicated a strong in vitro interaction between both elements in the absence of pristinamycin I. However, in the presence of PI, Pip was completely released from P_PTR (Fig. 1). P_PTR of S. pristinaespiralis contains 3 dyad symmetrical binding sites (Fig. 1; lane A; P_PTR in the absence of Pip) all of which are saturated at relatively large amounts of purified Pip (0.4 pmol; lane A) resulting in a single shifted band. Smaller amounts of Pip (0.02 pmol) presented an intermediate situation in which all possible permutations of site occupancy are

25 represented in three distinct retarded bands (lane C). Pip (0.2 pmol lanes D-M) was titrated with a range of PI from 50 pmol to 0.01 pmol (lanes D-H; D: 50 pmol; E: 10 pmol; F: 1 pmol; G: 0.1 pmol; H: 0.01 pmol). Pip was completely released from P_PTRby excess molar concentrations of PI. Under identical conditions the semisynthetic PI derivative quinupristin shows only partial release of Pip from individual P_PTR binding sites even at highest

20 concentration (lanes I-M; I: 50 pmol; J: 10 pmol; K: 1 pmol; L: 0.1 pmol; M: 0.01 pmol).

The relatively low Pip-releasing activity of quinupristin correlates with its lower efficiency in regulating the PIT P_PIR system (see Table 5, infra).

In order to analyze the potential of the Pip/P_PIR system for the design of a novel

25 mammalian gene regulation system, we adapted these Streptomyces regulatory elements for use in a eukaryotic context by construction of a set of two chimeric determinants: the PI- inhibited transactivator (PIT, Pip fused to the VP16 transactivation domain of the Herpes simplex virus; Triezenberg et al, 1988, supra) and the Pi-responsive promoter (P_PIR) (P_PT fused to the minimal insect promoter P_hsp70mi_n» Corces et al, 1984, supra). Following

3 ^J0^υ transfection of a P_PIR-driven GFP expression construct (pMF164) into CHO cells, no green fluorescence could be observed by fluorescence microscopy, indicating that no endogenous host factors activate P_PIR. GFP-expression could only be detected when pMF164 was cotransfected with PIT-encodin pMF156, showing that the chimeric PIT protein functions as

~_. a transactivator for P_PIR in mammalian cells. Transactivation of GFP expression was strictly

PIT-dependent and could not be achieved by cotransfection of pMF164 with vectors containing other transactivators such as the tetracycline- (TET-) or ecdysone-dependent transactivators (pUHD15-l encoding the tetracycline-repressible transactivator tTA; pUHD17-lneσ encoding the tetraycline-inducible reverse tTA (rtTA; Gossen et al, 1995, supra; pNgRXR encoding NgEcR and RXR which heterodimerize to form the ecdysone- dependent transactivator; No et al, 1996, supra). Also PIT did not activate GFP expression from TET- or ecdysone-dependent promoters (pMF104; pIND-GFP (Invitrogen)).

Addition of pristinamycin to the culture medium of CHO cells simultaneously transfected with PIT (pMF156) and a P_PIR-SEAP (human placental secreted alkaline phosphatase) reporter construct (pMF172) greatly decreases SEAP expression compared to a cotransfected culture without pristinamycin as shown in Table 2.

Table 2: Regulation potential of pristinamycin (Pyostacin⁰) and the pristinamycin group B (PI) and group A (PII) compounds. CHO-Kl cells were simultaneously transfected with pMF156 (PIT) and pMF172 (P_prR-SEAP) in the presence or absence of the inducer indicated. Reporter gene expression (SEAP) was assayed 48 hr later.

In addition, pMF206 (PIT2; Pip-p65) was cotransfected withpMF172 (P_PIR-SEAP) and the SEAP readout in the absence and presence of PI (2 μg/ml) was directly compared to CHO-Kl cells cotransfected with pMF156 (PIT; Pip-NP16) and pMF172 (P_PIR-SEAP). As shown in the following table, PIT2 mediates higher overall SEAP expression levels compared to the pMF156/pMF172 configuration in the absence of pristinamycin but basal expression levels in the presence of pristinamycin remain significantly higher.

Table 3 (previous page): Comparison of the regulation characteristics of PIT and PIT2. CHO-Kl cells were simultaneously cotransfected with the plasmids indicated and the SEAP readout (slope of the p-nitrophenolphosphate-based light absorbance time courses) was determined in the presence and absence of pristinamycin 48 hr post transfection.

This experiment indicates that external pristinamycin can enter mammalian cells and exert control of PIT/P_PIR-regulated gene expression there. However, pristinamycin has no influence on expression levels of P_SV40 or P_CMV-driven SEAP expression constructs (data not shown). Interestingly, only addition of the group B streptogramin PI shows effective downregulation of PIT/P_PIR-mediated SEAP expression; PII did not display any significant repression activity. Similar Pi-dependent gene expression using the PIT/P_PIR system has also been observed in BHK-21 and HeLa cells (not shown). No deleterious effects on CHO cell morphology and growth were observed at PI concentrations of 2 μg/ml found to be effective for repression of the PIT/P_PIR system. This is the same order of magnitude as tissue concentrations reached in antibiotic therapy. Only much higher PI concentrations (above 50 μg/ml) were toxic for CHO cells. Stable expression of PIT in CHO cells

Two representative clones, CHO-PIT 1 and CHO-PIT2, were chosen at random among 11 PIT-expressing CHO cell clones stably transfected with a constitutive PIT expression construct (pMF156). Both cell lines show no unusual cell morphologies and display similar growth behavior compared to wild-type CHO-Kl cells, indicating that sustained constitutive PIT expression does not have obvious deleterious physiological effects on CHO cells.

Transient transfection of CHO-PIT1 and CHO-PIT2 with a P_PIR-SEAP expression vector (pMF172) resulted in high level SEAP expression in the absence of PI and significant repression in the presence of 2 μg/ml PI as shown in Table 4.

Table 4: Pristinamycin I- (PI) dependent SEAP production of PIT- (Pi-dependent 5 transactivator) expressing stable cell lines CHO-PIT 1 and CHO-PIT2 transiently transfected with the SEAP-encoding, Pi-responsive reporter plasmid pMF172.

Induction factors (the ratio of SEAP activity without PI to SEAP activity with PI) reach 13 and 8 for CHO-PIT1 and CHO-PIT2, respectively. Pi-responsive SEAP regulation of both CHO-PIT derivatives is fully reversible following repeated cycles of addition and

10 withdrawal of this streptogramin, indicating reversible PI-PIT interaction in mammalian cells.

This characteristic is necessary to achieve fluctuating daily dosing regimes optimal for many therapeutic proteins such as insulin.

Dose-dependence of Pi-mediated gene regulation in CHO cells

25 The levels of P_PIR-regulated gene expression could be controlled by varying the concentrations of PI used for induction. This was assessed using a stable CHO cell line

(CHO-PIT1-SEAP) which expresses P_PIR-SEAP (pMF172) and contains also a constitutive

PIT expression vector (pMF156). Fig. 2 shows the dose-response curve for Pi-dependent

SEAP expression of CHO-PITl-SEAP cultures over 5 orders of magnitude of PI

20 concentration. Beyond the concentration window of 10-500 ng/ml, the PIT/P_PIR is either fully induced ( < 10 ng/ml) or repressed ( > 500 ng/ml), while gene expression can be adjusted to different levels within this window of concentrations. Increasing PI concentration above 10 ng/ml leads to a gradual decrease in SEAP expression, with lowest

2 SEAP expression levels at PI concentrations over 500 ng/ml. Besides the response plateaus of the PIT/P_PIR system at PI concentrations below 10 ng/ml and above 500 ng/ml, this system shows a broad window between these two concentrations in which the gene expression can be adjusted to intermediate levels. Such adjustable gene expression characteristics are particularly important for clinical applications which require titration of circulating proteins

^J 30^J into the therapeutic range.

Regulation efficiency of different streptogramins

Access to alternative effective streptogramins would expand the spectrum of regulating agents available for future use in human cell and gene therapy. Accordingly,

„ - several commercially available streptogramin sources were tested for their potential to regulate the PIT/P_PIR system as shown below in Table 5. Among available streptogramins, virginiamycin proved to be the most efficient regulating agent, showing an induction factor of

32, almost three times higher than pristinamycin-based regulation.

Table 5: Regulation potential of different streptogramins. CHO-PIT 1 was transiently transfected with the SEAP-encoding plasmid pMF172 and grown for 48 h in the absence or presence of different streptogramins at a concentration of 2 μg/ml of their group B component.

Nirginiamycin is very similar in structure to pristinamycin, since its group A component virginiamycin M₁ is identical to PII and its group B constituent, virginiamycin Si (NS_j), differs from PI only by the lack of a dimethylamine group.

The improved antibiotic therapeutic potential of Synercid resides in its water solubility. Although Synercid was expected to achieve better control of the PIT/P_PIR system due to its greater water solubility, its repression activity was almost 4-fold lower compared to Pyostacin^" and PI (Table 5). Apparently, the [(5δR)-[(3S)-quinuclidinyl]-thiomethyl modification of PI decreased the affinity of quinupristin to PIT, which is supported by gel retardation assays showing that quinupristin releases Pip from its target sequence P_PTR less efficiently than PI (Fig. 3).

Comparison and compatibility studies of the streptogramin- and tetracycline-based gene regulation systems in CHO cells

The regulation performance of the Pi-based regulation system was compared directly with the widely used TET-responsive expression concept. pTWTNl encoding both PIT and the tetracycline-dependent transactivator tTA in a single constitutive dicistronic expression unit was stably established in CHO-Kl cells. Two representative clones out of 20 pTWTNl - containing clones, CHO-TWTNl₉₅ and CHO-TWINli_0S, were chosen for further studies.

CHO-TWIN cells show the same cell morphology and similar growth behavior as wild-type

CHO cells.

CHO-TWINl₉₅ and CHO-TWIN1_10S were transfected under identical conditions with the isogenic plasmids containing the SEAP gene under control of the tTA-activated promoter

P_hCMV (pMFl 11) or the PIT-activated promoter P_prR (pMF172). Transfected cells were grown for 48 h in the presence or absence of either antibiotic prior to assay of SEAP activity.

Results are shown below in Table 6.

Table 6: Comparison of the regulation performance of streptogramin- (PI and virginiamycin; NIR) and tetracycline-based gene regulation systems. The two cell clones CHO-TWIΝl_9S and CHO-TWIN1₁₀₈ containing both pristinamycin- and tetracycline-dependent transactivators were transfected with either the streptogramin-responsive or the TET-responsive SEAP 5 reporter plasmids pMF172 (P_PIR-SEAP) and pMFl 11 (P_hCMV -SEAP), respectively, and the relative SEAP production was assessed. SEAP production values are normalized by SEAP activities under the antibiotic-free, active promoter conditions as follows: pMF172 in CHO- TWIN1₉₅: 1.7mU/ml; in CHO-TWTNl₁₀₈: 1.8mU/ml. pMFl ll in CHO-TWTNl₉₅: 18.6mU/ml; in CHO-TWIN1₁₀₈: 19.2mU/ml.

The maximum SEAP expression levels of the TET-system were approximately 10- fold higher compared to the P_PIR-driven SEAP expression. This is expected since the different chimeric promoters involved differ in their minimal promoter elements. While the TET- promoter P_hCMV*.ι has been optimized by heptameric tTA binding elements to reach 35-fold 25 higher expression levels than the precursor viral CMN promoter (Yin et al, 1995, Cancer Res. 55, 4922-4928), P_PIR is in its unmodified, original "wild-type" configuration prior to any such refinements to increase fully-activated transcription. To analyze the relative regulatory characteristics, we evaluated induction factors (IF; ratio of maximal expression level to antibiotic-repressed expression level). IFs have been shown to remain largely unaffected by

20 the choice of the minimal promoter or other promoter modifications aimed at increasing the maximum expression level (No et al, 1996, supra). While the PI- and the TET-system of

CHO-TWTNl₉₅ exhibit nearly identical IFs of 12 in the SEAP reporter assay, CHO-TWIN1₁₀₈ displays 13 -fold downregulation of SEAP in the presence of TET but downregulation by a

25 factor of 45 by the addition of PI. Using virginamycin as regulating agent instead of PI for

CHO-TWINl₉₅ and CHO-TWIN1₁₀₈ transfected with the P_pm-SEAP expression vector

(pMF172) IFs are even higher (27 for CHO-TWINl₉₅ and 83 for CHO-TWIN1₁₀₈) (Table 6).

Therefore, the PIT/P_PIR-based mammalian gene regulation system shows superior gene inducibility characteristics than the TET system in CHO cells engineered to provide

^D 30^U regulation of both systems.

The PI and the TET systems can be used for combined therapeutic applications requiring different control modalities for different transgenes since they are compatible and regulated by different therapeutic antibiotics for which large sets of clinical and

„ . pharmacokinetic data are available. Crossrepression of PI and TET (or the TET derivative doxycycline) on the other expression regulation system was evaluated for CHO-TWIN1₁₀₈ transfected separately with P_PIR-SEAP (ρMF172) and P_hCMV*__rSEAP (pMFl 11) expression constructs, and the results shown in Table 7.

¹⁰

Table 7: Compatibility of the streptogramin- and tetracycline-based gene regulation systems.

Crossregulation of pristinamycin (PI) and the tetracyclines doxycycline (DOX) and tetracycline (TET). The cell line CHO-TWINl ₁₀₈ which contains the tetracycline-dependent transactivator (tTA) as well as the pristinamycin-dependent transactivator (PIT) was

15 transfected with the streptogramin-responsive reporter SEAP plasmid pMF 172 (P_PIR-SE AP) or the tetracycline-responsive SEAP reporter plasmid pMFl 11 (P_hCMV*.₁-SEAP).

PI shows no effect on expression from the TET-responsive system. This experiment indicates that PI activation of SEAP expression in constructs containing PIT and P_PIR operates by specific interaction of PI and PIT, as occurs between PI and the S. coelicolor protein Pip

20 (Table 7; Fig. 5), and not via any generalized, nonspecific effect of PI on CHO cells.

Both tetracycline derivatives downregulate the PIT/P_PIR system in a dose-dependent manner. However, at tetracycline concentrations commonly used for regulation of the TET- responsive promoter (2 μg/ml), the repressive effect of TET on the PIT/P_PIR system is only 25 17%). This is sufficiently low that, for practical purposes, the TET- and Pip/P_PIR systems provide essentially independently regulatable expression.

DISCUSSION

Properties of an ideal gene regulation system for therapeutic use include (i) low baseline expression and high induction ratio, (ii) control by a readily bioavailable, small-

30 molecule drug showing no interference with host metabolism, (iii) high pharmacokinetic turnover of the regulating agent in all tissues to allow rapid reversion to the wild-type configuration, (iv) low potential for immune recognition, and (v) high modularity of regulation components to allow independent and efficient optimization as well as adaptation 35 of the system to specific therapeutic situations. The streptogramin-based mammalian gene regulation system described here ideally satisfies these requirements. In cell lines engineered to afford control of both systems, the streptogramin-based regulation system shows over 10- fold lower baseline expression levels and up to 4-fold higher induction rations than the TET- responsive system. More importantly, streptogramins and tetracycline regulate their 5 respective mammalian transcription systems essentially independently, showing that these two systems can be used together to achieve advanced therapeutic regimens in which to sets of transgenes can be regulated separately.

A whole set of streptogramin antibiotics, including the recently approved Synercid , are available with confirmed excellent bioavailability, pharmacokinetics and human compatibility. Also, streptogramin antibiotics are actively taken up by human cells, leading to regulation-effective concentrations in almost every part of the human body (Bebear, et al, 1997, J. Antimicob. Chemother. 39, 59-62). Nevertheless, streptogramin antibiotics are rapidly eliminated from the blood and most tissues, with half-lives typically not exceeding 25 one hour in humans (Bergeron et al, 1997, J. Antimicrob. Chemother. 39, 129-138), ensuring ready reversibility of possible therapeutic conditions. The brain is the only site where virtually no streptogramin could be found following oral or intravenous application, since this class of antibiotics is unable to cross the blood-brain barrier (Bergeron et al, 1997, supra).

This natural tissue-specific selectivity could allow straightforward brain-exclusive expression

20 of the PIT/P_PIR system in the presence of regulation-effective concentrations of streptogramin in the remaining body parts to enable design of therapeutic strategies for neurodegenerative diseases (Hardy et al, 1998, Science 282, 1075-1079).

A mammalian gene regulation system such as the PIT/Ppm configuration which

25 responds to a class of chemically related antibiotics offers another powerful new dimension: finding new antibiotics which are urgently needed to cope with increasing prevalence of multidrug resistant bacterial pathogens. Metabolic libraries from Streptomyces or fungi could be screened using cultured mammalian cells containing the PIT/P_PIR system linked to a reporter gene to identify novel streptogramins. This technology is at least one order of

30 ^u magnitude more sensitive than classical microbial inhibition tests and not biased by antibiotic resistance of indicator bacteria.

Additionally, the system described above can be adapted to use any of a number of different antibiotic resistance systems found in secondary metabolite/antibiotic producers

_ ,. within the Actinomycetes, particularly the Streptomyces. Examples of such systems that contain P_ptr- responsive promoters with cognate antibiotic binding proteins are the rifamycin- responsive gene cluster, the daunorubicin-responsive gene cluster, the landomycin-responsive gene cluster, the rapamycin-responsive gene cluster, and the tetracenomycin-responsive gene cluster.

Example 2: The Double Regulation System

Most of today's gene therapy and tissue engineering strategies focus on stable integration of transgenes into human somatic cells either in vivo or ex vivo. While initial success was achieved using sophisticated gene transfer technology including attenuated viruses, site-specific recombination for targeted integration and non-immunogenic selection markers, gene transfer is not the only challenge in future gene therapy and tissue engineering. However, the success and realization of this technology will largely dependent on flanking concepts which allow ex vivo expansion of grafted tissue followed by sustained growth control and reimplantation of treated cells or tissues. This concept requires two consecutive steps of opposite proliferation control which enables first expression of genes which activate proliferation for ex vivo expansion of tissue cells followed then by gene therapeutic operation and activation of proliferation control to allow reimplantation of genetically engineered ti •ssue.

Using two human-compatible gene regulation systems, the tetracycline- and the pristinamycin-responsive system we set out to construct a double regulation system to achieve completely externally controlled proliferation management of mammalian cells. MATERIALS AND METHODS

In order to construct the double regulatable mammalian expression vector pDuoRexl a PpiR-pA containing fragment was first amplified from pMF164 (PpiR-GFP expression construct) using oligonucleotides OMF87:

GATCACTAGTGATATCgctagctgtgtgtgagttc (SEQ JO NO:8) and OMF86:

GATCGGATCCATCGATAATAAAATATCTTTATTTTCATTACATCTGTGTGTTGGTTTTTTGT

GTGcgtcggagaaatagcgc (SEQ ID NO:9). OMF86 contains in its 5' extension an artificial polyadenylation site (Levitt et al, 1989, Genes Dev. 3: 1019-1025) optimized to eliminate restriction sites. Also, OMF87 and OMF86 contain in their 5' extensions Spel and Clάl sites (bold), respectively. The PpiR-pA-containing fragment was cloned into pcDNA3.1/V5/His-TOPO (Invitrogen) in anti-PcMV orientation to give plasmid pSAM224. The PpiR-pA-containing cassette was released from pSAM224 by restriction with Spel and Clal and cloned into the corresponding sites (Spell Claϊ) of pTRTDENTl to replace IRESIL to result in pDuoRexl. As shown in Fig. 6, pDuoRexl consists of the following double regulated expression unit: P CMN*-l- CSI-/RES-MCSII-pAi-/-PpiR-MCSIII-pAlI (P hCMN*-l. tetracycline-regulated promoter; MCS, multiple cloning site, IRES, internal ribosomal entry site;^4, polyadenylation site; PpiR, pristinamycin-responsive promoter). Double regulation of MeklDD, a constitutively active protein kinase of the central mammalian growth factor signaling pathway and the cyclin-dependent kinase inhibitor (CDI)

10 p27 using pDuoRexl was achieved by a 4-step cloning procedure: (1) MeklDD was amplified from (pGem-MeklDD; Greulich and Εrikson, 1998, J. Biol. Chem. 273:13280- 13288) by OMF92: GATCGATATCACTAGTCGCCACCatgcccaagaagaagcc (SΕQ ID ΝO:10) and OMF93: GATCGGATCCACGCGTtcagatgctggcagcgtg (SΕQ ID NO:ll) ^ and the 1.3 kb fragment was cloned into pcDNA3.1 /V5/His-TOPO under control of the P CMN promoter to give pMF192. (2) p27 was amplified (from pDD6; Fussenegger et al, 1998) with OMF94: GATCGAATTCAAGCTTgcggtcgtgcagacccgg (SΕQ ID ΝO:12) and OMF95: ATGCATCGATGCGGCCGCttacgtttgacgtcttctg (SΕQ ID NO:13) and the

20 600 bp was cloned into pcDNA3.1 /V5/His-TOPO under control of PCMN o give pMF 193. (3) p27 was excised from pMF193 by EcoRI and Nøtl (contained in OMF94 and OMF95, respectively (bold)) and cloned into the corresponding sites of pDuoRexl (EcoRI NotT) thereby replacing the /RES-containing fragment and placing p27 under control of PhCMV*-l to result in plasmid pMF 194. (4) pMF 192 was restricted with Spel and BamHl (contained in OMF92 and OMF93, respectively (bold)) and MeklDD was transferred to pMF194 linearized by Spel and BgHl thus placing MeklDD under control of PpiR to give plasmid pMF195. As shown in Fig. 7, pMF195 contains the following double regulated expression unit: PhCMN*- l-p27-pAι-/-PpiR-MeklDD-pAii.

30

Additional double regulation vectors analogous to pMF195 (which contained p27-

MeklDD) were also constructed in which the MeklDD gene coding sequence was replaced by other growth promoting genes. Specifically constructs containing sequences encoding the adenoviral large T antigen, the adenoviral small T antigen, and the human papillomavirus Ε7

35 protein under control of the P_PΪR promoter were constructed using well known techniques.

In a parallel cloning procedure pMF188 was cloned containing the PhCMV*-l-

SEAP-/RES-pAl-/-PpiR-GFP-pAπ. GFP was amplified from pMF164 by OMF84:GATCGCGGCCGCAATAAAATATCTTTATTTTCATTACATCTGTGTGTTGGTTTTTT GTGTGcgtcggagaaatagcgc (SEQ ID NO:14) and OMF85: GTACAGATCTtta tttgtagagctcatc (SEQ ID NO:15) and cloned into pcDNA3.1/V5/His-TOPO under control of PCMN to give plasmid pSAM223. Since OMF84 contains in its extension an artificial pA element pSAM223 contains a pA-GFP cassette which was excised by Notl and Bglϊl (contained in OMF84 and OMF85, respectively (bold)) and cloned into the corresponding sites (NotllBgRl) of ρMF124 (pTRIDEΝT3 containing SEAP in the first cistron) to give pMF188.

A further double regulated configuration was cloned for use in transgenic animals utilizing the two compatible GFP-derived fluorescent proteins, CFP (cyan fluorescent protein) and YFP (yellowish fluorescent protein) were combined in a pDuoRexl -derived expression vector by a 4-step cloning procedure: (1) CFP was amplified from pECFP-Cl (Clontech) with oligos

OMF96:GATCGAATTCCCTCAGCACCAGGTCATGCTTAAGTCGCGACATATGgatccgctag cgctaccg (SEQ ID NO: 16) and OMF89: GATCAAGCTTGCCCGGGCCACACAAA AAACCAACACACAGATGTAATGAAAATAAAGATATTTTATTTGATCAGGCGCGCCGCGGCCG CATGCttacttgtacagctcgtc (SEQ ID NO:17) and cloned into pcDNA3.1/V5/His- TOPO under control of PCMN to give plasmid pSAM228. (2) YFP was amplified from pEYFP-Cl (Clontech) with oligos

OMF90:GTACGAATTCGATATCATGCATGGCGCCGTTTAAACGCGTATTTAAATgatccgct agcgctaccg (SEQ ID NO:18) and OMF91: GATCAAGCTTGCGGCCGCGGATCCG

CCCGGGCCACACAAAAAACCAACACACAGATGTAATGAAAATAAAGATATTTTATTATCGAT

ACTAGTGCGATCGCTTAATTAATTTAAATttacttgtacagctcgtcc (SEQ ID NO:19) and cloned into pcDNA3. l/V5 His-TOPO under control of PCMN to give plasmid pSAM222.

(3). CFP was excised from pSAM228 by EcoRI and Notl (contained in OMF96 and OMF89, respectively (bold)) and cloned into the corresponding sites (EcoRI/Notl) of pDuoRexl thereby replacing the IRES element to give pMF200. (4) pMF200 was linearized with

EcoRN and Spel ligated to YFP excised from pSAM222 by EcoRN and Spel (contained in

OMF90 and OMF91, respectively (bold)) to give pMF201. As shown in Fig. 4, pMF201 contains the following double regulated expression unit: P_nCMN*-l-CFP-pAi-/-PprR-YFP- pAπ. pMF196 is a CFP/YFP-containing double regulation construct similar to pMF201 but 5 both expression units converge. Cloning of pMF196 required a 4-step cloning procedure: (1) The YFP-pA containing cassette was excised from pSAM222 by EcoRI and Notl and ligated to the corresponding sites of pMF164 (EcoRI/Notl) thereby replacing GFP of pMF164 to result in plasmid pSAM226 (2) The PprR-YFP-pA cassette was excised with Sspl/Notl from pSAM226 and ligated into the corresponding sites pTRIDΕΝT 1

10

(SspllNotl) thereby replacing PhCMN*-l and IRES I of pTRIDΕΝT 1 to give plasmid pMF190. (3) The CFP-pA cassette was excised from pSAM228 (OMF89 contains in it extension an artificial pA site) by EcoRI and Hindlϊl (contained in OMF96 and OMF89, respectively) and ligated into the corresponding sites (Eco^'RIlHindllT) of pTBCl (Dirks et al,

15 1993) to result in plasmid pSAM227. (4) The PhCMN*-l-CFP-pA-containing cassette was released from pSAM227 by digestion withN7?oI and Srtl and was ligated to pMF190 digested with SΩtl and Srtl to result in plasmid pMF196. As shown in Figure 4 pMF196 contains the following double regulated expression unit: PpiR-CFP-pAi/pAπ-YFP-PhCMN*-l-

RESULTS 20

Construction of tetracycline- and streptrogramin-responsive double regulation vectors In order to demonstrate the double regulation concept a set of three vectors were constructed (i) pMF188 combining tetracycline-responsive SEAP (human secreted alkaline phosphatase) and pristinamycin-responsive GFP expression in a single double regulated

25 expression unit: P_nCMN*-l-SEAP-pAi-PpiR-GFP-pAπ (Fig.3). (ii) Similar to pMF188, pMF201 contains tet-regulated CFP (cyan fluorescent protein) and pristinamycin-regulated YFP (yellowish fluorescent protein) in a single expression unit: PhCMN*-l^_CFP-pAτ8 -PpjR -YFP-pAπ (Fig. 4). (iii) ρMF196 is isogenic to pMF188 and pMF201 but contains the

30 following double regulated expression unit: Pp R-YFP-pAi-pAπ -CFP-PhCMN*-l (Fig. 5). When these double regulation vectors were transiently transfected into CHO-TWINl 108 cells all expression units contained on pMF188/pMF201/pMF196 showed independent on/off regulation regulation in response to the respective regulating antibiotics tetracycline and pristinamycin as monitored by SEAP activity tests and fluorescence microscopy. Based on experience with pMF188/pMF201/pMF196 we designed a multipurpose double regulation vector pDuoRexl for use in tissue engineering and gene therapy (Fig. 6). For the novel gene therapy and tissue engineering concept positive and negative proliferation control genes were carefully chosen. For negative proliferation control the use of tumor suppressor genes such as p21, p27, p53 or combinations thereof with differentiation factors such as c/ebpa and/or survival factors such as bcl-2 and bcl-xχ_ in a multicistronic configuration is well-tested and reliable (Fussenegger et al, 1998). The choice of the positive proliferation control gene is important since its effect on engineered cells has to be sufficiently strong to induce resumption of the cell cycle in terminally differentiated cells but weak enough to allow reversion of the proliferating state and reimplantation of treated cells into the in vivo context.

We chose Mekl, a central protein kinase in the conserved mammalian Ras-MAP signal fransduction pathway responding to growth-promoting signals such as cytokines. At the center of such signal fransduction Mekl not only transfers but also amplifies growth- promoting signals. Upstream kinases (Erk) phosphorylate Mekl at two positions (serine 218) and (serine 222). A recently described double mutant contains two mutations which alter these ser218 and ser22 to aspartate and render MeklDD constitutively active in the absence of upstream proliferation signals and in the absence of phosphorylation (Greulich and

Erikson, 1998, J. Biol. Chem. 273: 13280-13288). Previous results by Greulich and Erikson,

1998) have shown that regulated overexpression of MeklDD enables proliferation of NIH-

3T3 cells in low serum. A pDuoRexl -based expression vector, pMF195, was constructed which contains p27 under control of PhCMV*-l and MeklDD under control of PpiR (Fig.

7). Immunofluorescence analysis of CHO-TWINl sub 108 transiently transfected with pMF195 show clear independent regulation of p27 and MeklDD in response of the respective regulating antibiotic tetracycline and pristinamycin.

Besides MeklDD other genetic determinants exerting positive control of mammalian cell cycle such as cyclins (e.g. cyclin E) or E2F could also be envisioned. For example, double regulation vectors analogous to pMF195 (which contained p27-MeklDD) were constructed in which the MeklDD gene coding sequence was replaced by sequences encoding the adenoviral large T antigen, the adenoviral small T antigen, and the human papillomavirus E7 protein (all under control of the P_PIR promoter). One can also use a sequence encoding the human papillomavirus E6 protein. In principle, this two gene double regulation concept can also be adapted to the multicistronic level to carefully fine tune expression of opposing proliferation control genes or to combine, for example, growth- inhibiting genes with genes such as c/ebp a which induce and maintain terminal 5 differentiation. Also, monocistronic double regulation concepts could be foreseen which use sense and anti-sense expression of a single gene to induce or inhibit proliferation (or vice versa). Candidate genes include p27 anti-sense, expression of which has been shown to stimulate proliferation of quiescent fibroblasts and enable growth in serum-free medium

(Rivard et αZ.,1996, J. Biol. Chem. 271: 18337-18341.), and nedd5 which is known as

10 positive growth controlling gene (Kinoshita et al, 1997, Genes Dev. 11: 1535-1547 but anti- sense expression may inhibit growth and leads to fried egg-like cell shapes as seen for cells arrested by overexpression of p27 (our own data not shown).

15 Example 3 : Promoter modification to enhance maximal expression from the streptogramin-responsive system

In principle, every part of the PIT/PpiR system can be improved individually or exchanged by a more powerful component: (1) The Pip domain of PIT can be subjected to random mutagenesis to alter DNA binding affinity or specificity, to improve affinity to 20 pristinamycin or other streptogramins or antibiotics or to reverse streptogramin responsiveness of the PIT/PpiR system (reverse PIT/Ppπi system), for example allow induction of gene expression upon addition of pristinamycin instead of its withdrawal. (2)

The VP16 domain could be exchanged by other transactivation domains such as the p65

25 domain of human NF-κ B (Schmitz and Baeuerle, 1991, EMBO J. 10: 3805-3817) to "humanize" the regulation system and reduce immune recognition of PIT or the KRAB silencing domain of the human koxl gene (Deuschle et al, 1995, Mol. Cell Biol. 15: 1907- 1914) to construct a reverse PIT/PpiR system (these systems are under construction). (3)

_« exchange of the minimal Drosophila heatshock promoter by other minimal promoters such as the minimal CMN promoter or the minimal promoter of the adenoviral E1B gene (also under construction).

The naturally occurring Pptr sequence contains only three distinct Pip binding sites which could accommodate PIT. In order to enhance maximal expression from streptogramin-

³ ^ responsive system, we constructed Pi-responsive promoters with two (P_PJR2) and three (P_PIR3) ptr binding motifs with a total of 6 and 9 Pip binding sites, respectively. Two synthetic DΝA fragments were synthesised: PPIR2 (SEQ ID NO:20): gacgtcgatatcGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGTGTACGAGTAT CTTCCCGTACACCGTACGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGTGTACG AGTATCTTCCCGTACACCGTACcctgcaggand PPIR3(SEQJ-DNO:21): gacgtcgatatcGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGTGTACGAGTAT CTTCCCGTACACCGTACGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGTGTACG AGTATCTTCCCGTACACCGTACGAAATAGCGCTGTACAGCGTATGGGAATCTCTTGTACGGT GTACGAGTATCTTCCCGTACACCGTACcctgcagg.

Both DNA fragments were digested with-4αtll and S-?e8387I (corresponding restriction sites are shown in bold) and ligated to the corresponding sites of pMF172 (P_PIR-SEAP;

thereby replacing the single Pip binding sequence (Pptr of S. pristinaespiralis; Salah-Bey et al, 1995; Salah-Bey and Thompson, 1995) to result in plasmids pMF198 (P_Pm2-SEAP) and pMF199 (P_PIR3-SEAP).

All three PIT-dependent promoters PpiR, PpiR2 and PpiR3 were cloned in front of a seap reporter gene to measure their transcriptional activity in the absence of pristinamycin (full induction). Table 8 shows relative expression profiles of PpiR2 (pMF198; PpiR2- SEAP) and PpiR3 (pMF199; PpiR3-SEAP) compared to "wildtype" PpiR.

Table 8: Relative SEAP Expression from Modified P ^-linked Genes

Example 4: Construction of a pristinamycin-inducible expression system with enhanced regulation characteristics

The Classical PIT/P_PIR system belongs to the "OFF" family of regulation concepts since gene expression is activated upon withdrawal of the regulating antibiotic. However, in some applications such as gene therapy and tissue engineering an "ON" system which is induced upon addition of streptogramin antibiotics is more desirable. We therefore constructed a new Pip-based binary PI_0N system which consists of a P_PIRON promoter and a set of two different transrepressors. P_PIRON consists of a PIR3 module containing 3 consecutive Pip binding elements (SEQ ID NO:21; Example 3) placed in front of the strong viral SN40 promoter. Transrepressors such as PIT3, which consists of a protein fusion between Pip and the KRAB silencing domain of the human kox-1 gene (Deuschle et al.,

1995, Mol. Cell Biol. 15: 1907-1914) and PIT4, which is simply Pip expressed in a eukaryotic configuration, bind to PIR3 in front of P_sv40 and block transcription of this promoter. Besides sterical transcription blocking the silencing domain of PIT3 can additionally downregulate P_SV40 activity.

MATERIALS AND METHODS

The PIR3 sequence (SEQ ID NO:20) was excised from GeneOp™-PIR3 (the pUC- derived GeneOp expression vector is sold by Operon Technologies Inc.) by EcoRN and Smαl and ligated into the Nrul site of pSΕAP2-control (Clontech) thus resulting in plasmid pMF208 containing P_SV40-PIR3-SEAP cassette. The combination of P_SV40 and PIR3 is designated P^-OΝ.

PIT3, the Pip-KRAB fusion protein was constructed by a three step cloning procedure: (i) The KRAB domain of the human kox-l gene was amplified from pSCTENgal93Kox (Moosmann et al., 1997, Biol. Chem. 378: 669-677) with oligos

OMF99:GATCGCGCGCC (AGATCCAAAAAAGAAGAGAAAGGT) (AG ATCCAAAAAGAAGAGAAAGGT) (AGATCCAAAAAAGAAGAGAAAGGT) AATGGATGCTAAGT CACTA (SEQ ID NO:25) and OMF100: GATCAAGCTTGGATCCTTACC

AGAGATCATTCCTTGC (SEQ ID NO:26) and ligated in antisense orientation into pcDNA3.1/V5/His-TOPO (Invitrogen) to result in pMF203. OMF99 contains three consecutive nuclear localization signals (bracketed) derived from the SV40 large-T-antigen

(Kalderon et al., 1984). (ii) The KRAB domain was excised from pMF203 by BssΗIl/HindUl and ligated to the corresponding sites (R-wHII/Hmdlli) of pSBC2-tTA (Fussenegger et al,

1997) thus replacing the VP16 domain of tTA and resulting in pMF205. (iii) The tetR domain was excised of pMF205 by Sspl/BssΕDI and replaced by Pip excised from pMF150 by

SspllBssBϊl to give plasmid pMF207. The resulting plasmid has the following PIT3 coding sequence:

ATGAGTCGAGGAGAGGTGCGCATGGCGAAGGCAGGGCGGGAGGGGCCGCGGGACAGCGTGTG GCTGTCGGGGGAGGGGCGGCGCGGCGGTCGCCGTGGGGGGCAGCCGTCCGGGCTCGACCGGG ACCGGATCACCGGGGTCACCGTCCGGCTGCTGGACACGGAGGGCCTGACGGGGTTCTCGATG CGCCGCCTGGCCGCCGAGCTGAACGTCACCGCGATGTCCGTGTACTGGTACGTCGACACCAA GGACCAGTTGCTCGAGCTCGCCCTGGACGCCGTCTTCGGCGAGCTGCGCCACCCGGACCCGG ACGCCGGGCTCGACTGGCGCGAGGAACTGCGGGCCCTGGCCCGGGAGAACCGGGCGCTGCTG GTGCGCCACCCCTGGTCGTCCCGGCTGGTCGGCACCTACCTCAACATCGGCCCGCACTCGCT GGCCTTCTCCCGCGCGGTGCAGAACGTCGTGCGCCGCAGCGGGCTGCCCGCGCACCGCCTGA CCGGCGCCATCTCGGCCGTCTTCCAGTTCGTCTACGGCTACGGCACCATCGAGGGCCGCTTC CTCGCCCGGGTGGCGGACACCGGGCTGAGTCCGGAGGAGTACTTCCAGGACTCGATGACCGC GGTGACCGAGGTGCCGGACACCGCGGGCGTCATCGAGGACGCGCAGGACATCATGGCGGCCC GGGGCGGCGACACCGTGGCGGAGATGCTGGACCGGGACTTCGAGTTCGCCCTCGACCTGCTC GTCGCGGGCATCGACGCGATGGTCGAACAGGCCTCCGCGTACAGCCGCGCGCCAGATCCAAA AAAGAAGAGAAAGGTAGATCCAAAAAAGAAGAGAAAGGTAGATCCAAAAAAGAAGAGAAAGG TAATGGATGCTAAGTCACTAACTGCCTGGTCCCGGACACTGGTGACCTTCAAGGATGTATTT GTGGACTTCACCAGGGAGGAGTGGAAGCTGCTGGACACTGCTCAGCAGATCGTGTACAGAAA TGTGATGCTGGAGAACTATAAGAACCTGGTTTCCTTGGGTTATCAGCTTACTAAGCCAGATG TGATCCTCCGGTTGGAGAAGGGAGAAGAGCCCTGGCTGGTGGAGAGAGAAATTCACCAAGAG ACCCATCCTGATTCAGAGACTGCATTTGAAATCAAATCATCAGTTTCCAGCAGGAGCATTTT TAAAGATAAGCAATCCTGTGACATTAAAATGGAAGGAATGGCAAGGAATGATCTCTGGTAA (SEQ ID NO:27). For construction of PIT4, the Pip protein of Streptomyces coelicolor was amplified from pGemT: :Epip4 with oligos

OMF63:GTACGAATTCCCACCATGAGTCGAGGAGAGGTG (SEQ ID NOT) and OMF103: GATCAAGCTTTCAGGCCTGTTCGACCATC (SEQ ID NO:28) followed by cloning into vector pcDNA3.1/N5/His-TOPO (Invitrogen) under control of P_CMV. The resulting plasmid pMF225 contains the following PIT4 sequence which corresponds to the sequence of the pip gene of Streptomyces coelicolor.

ATGAGTCGAGGAGAGGTGCGCATGGCGAAGGCAGGGCGGGAGGGGCCGCGGGACAGCGTGTG GCTGTCGGGGGAGGGGCGGCGCGGCGGTCGCCGTGGGGGGCAGCCGTCCGGGCTCGACCGGG ACCGGATCACCGGGGTCACCGTCCGGCTGCTGGACACGGAGGGCCTGACGGGGTTCTCGATG CGCCGCCTGGCCGCCGAGCTGAACGTCACCGCGATGTCCGTGTACTGGTACGTCGACACCAA GGACCAGTTGCTCGAGCTCGCCCTGGACGCCGTCTTCGGCGAGCTGCGCCACCCGGACCCGG ACGCCGGGCTCGACTGGCGCGAGGAACTGCGGGCCCTGGCCCGGGAGAACCGGGCGCTGCTG GTGCGCCACCCCTGGTCGTCCCGGCTGGTCGGCACCTACCTCAACATCGGCCCGCACTCGCT GGCCTTCTCCCGCGCGGTGCAGAACGTCGTGCGCCGCAGCGGGCTGCCCGCGCACCGCCTGA CCGGCGCCATCTCGGCCGTCTTCCAGTTCGTCTACGGCTACGGCACCATCGAGGGCCGCTTC CTCGCCCGGGTGGCGGACACCGGGCTGAGTCCGGAGGAGTACTTCCAGGACTCGATGACCGC GGTGACCGAGGTGCCGGACACCGCGGGCGTCATCGAGGACGCGCAGGACATCATGGCGGCCC GGGGCGGCGACACCGTGGCGGAGATGCTGGACCGGGACTTCGAGTTCGCCCTCGACCTGCTC GTCGCGGGCATCGACGCGATGGTCGAACAGGCCTGA (SEQ ID NO: 29). pMF150 is almost identical to pMF225 but the TGA stop codon of Pip encoded on pMF150 has been mutated in order to allow fusion to the various transactivating and transrepressing domains. Because of the mutation of this stop codon the pip gene contained in pMF150 is terminated a few bp downstream at the next stop codon encountered in the vector sequence. However, both Pip proteins perform equally in the PipON regulation concept.

RESULTS

Regulation characteristics of the PipON system CHO-Kl cells were cotransfected with pMF208 (P_PIRON-SEAP) and either fransrepressor-containing plasmid pMF207 (PIT3; Pip-KRAB) or pMF225 (PIT4; Pip). Cells were grown in the absence and presence of PI. In the absence of PI, expression of P_PIRON was fully repressed and induction could be observed in the presence of PI. Table 9 shows the regulation characteristics of both inducible regulation concepts.

Table 9: Comparison of the regulation characteristics of the PipON systems. CHO-Kl cells 5 were simultaneously cotransfected with the plasmids indicated and the SEAP readout (slope of the p-nitrophenolphosphate-based light absorbance time courses) was determined in the presence and absence of pristinamycin 48 hr post transfection. The SEAP expression levels of the PipON systems were compared to the pSEAP2-control construct which is isogenic to pMF208 but lacks the PIR3 binding element in front of the P_sv40 promoter.

While both PipON systems are induced in the presence of PI, their maximal as well as their basal expression levels differ significantly. Whereas the PIT4-P_PIRON configuration (pMF225-pMF208) reaches expression levels comparable to the strong constitutive viral SN40 promoter and shows an induction factor of 25 which is more than double the induction 25 factor of the classical PIT/P_PIR system, the PIT3-P_PIROΝ configuration resulted in the tightest repression and an induction factor of about 150, more than 15 times higher compared to the PIT/P_PIR system. However, the maximal expression level of the PIT-P_PIRON configuration is about 3-fold lower compared to the maximal expression level of the PIT/P_PIR configuration.

The difference in their regulation characteristics enables the use of the two PIPON system for

20 a broad range of different applications. In configurations which require high expression levels the use of the PIT4/P_PIRON configuration is advantageous whereas situations which require tightest repression of basal expression of the transgene, the PIT3/P_PIRON concept is the preferred system.

25 The modular setup of P_PIRON consisting of an independent operator sequence (PIR3) and a fully functional promoter element allows straightforward exchange of PSV40 of P_PIRON by any type of promoter to enable, for example, tissue-specific regulated expression or adaptation of this inducible regulation concept to other organisms such as yeast and plants.

Induction potential of different streptogramin antibiotics on the Pip/P_PIRON regulation system

^J 30 In order to assess the regulation potential of different streptogramin antibiotics, pMF225 or pMF150 encoding PIT4 were cotransfected with pMF208 (P_PIRON-SEAP) into

CHO-Kl cells and the SEAP induction was measured 48 following transfection (as described above in Example 1) and addition of the following streptogramin antibiotics: (i) PI, the group

_ . B streptogramin component of pristinamycin; (ii) Pyostacin, the human oral antibiotic; (iii)

Synercid, the first injectable semi-synthetic pristinamycin derivative effective against most multidrug resistant human pathogenic bacteria and (iv) virginiamycin used as growth promotant in livestock feed. 2μg/ml group B component was used for the regulation studies (for mixed streptogramins the concentration of the group B component was determined based on the fixed 70%:30% ratio of group A and group B components).

Table 10 shows the induction potential of the different streptogramins. Pyostacin is as efficient in inducing the Pip/P_PIRON as is pure PI component, virginiamycin exhibits slightly lower regulation performance and Synercid has almost no inducing capacity on the Pip/P_PIRON system. Therefore, the use of Synercid in antibiotic chemotherapy is likely to be compatible with the use of the Pip/P_PIRON system in gene therapy and tissue engineering applications.

Table 10: 1-nduction performance of different streptogramin antibiotics. CHO-Kl cells were simultaneously cotransfected with the plasmids pMF225 (PIT4) and pMF208 (P_PIROΝ-SEAP) and the SEAP readout (slope of the p-nifrophenolphosphate-based light absorbance time courses) was determined in the presence and absence of the streptogramins indicated 48 hr post transfection. Example 5: Construction of multi-purpose expression vectors In order to use the pristinamycin-responsive mammalian gene regulation system in a wide variety of applications, a set of 6 mammalian gene expression vectors which are compatible with the use of all pristinamycin-dependent transactivators (PIT, PIT2) and transrepressors (PIT3, PIT4) was constructed. The first set of vectors, pMF189 and pMF229, consists of monocistronic expression vectors containing the P_PIR (pMF189) and P_PIRON (pMF229) promoters followed by a multiple cloning site of up to 22 unique restriction sites 6 of which are rare-cutting sites for enzymes recognizing 8 bp (Figure 8).

The second set of vectors is shown in Figure 9 and contains the same promoters P_PIR and P_PIRON integrated in the pTRIDENT family of tricistronic expression vectors (Fussenegger et al. 1998, Biotechnol. Bioeng. 57,1-10). pTRIDENT vectors contain a single tricistronic expression unit which is driven either by P_PIR (pTRIDENT9 and pTRIDENT 10) or P_PIRON (pTRIDENTl 1 and pTRIDENT12). While the first cistron is translated in the classical cap-dependent manner, the following two cistrons rely on cap-independent translation initiation mediated by internal ribosomal binding sites of the encephalomyocarditis virus or of polio viral origin (IRES). While pTRIDENT9 (P_P ) and pTRIDENTl 1 (P_PIRON) contain two IRES elements, pTRLDENTlO (P_PIR) and pTRIDENT 12

(P_PIRON) contain an IRES as well as a CITE element. Both IRES elements are among the strongest currently available, showing high translation initiation in a wide variety of mammalian cells and tissues (Borman et al., 1997; Fussenegger et al, 1998, Biotechnol. Bioeng. 57,1-10, Fussenegger et al, 1998, Nat. Biotechnol. 16,468-472). Both IRES elements are flanked by large polylinkers which allow convenient movement of genes into pTRIDENT derivatives. pTRIDENT vectors have proven to be useful tools for a wide variety of applications (Fussenegger et al, 1998, Biotechnol. Bioeng. 57,1-10; Fussenegger et al, 1998, Nat. Biotechnol. 16, 468-472)

Construction of multi-purpose expression vectors pMF189 the P_PIR expression vector was constructed by a three step cloning procedure: (i) the yellowish fluorescent protein YFP was amplified from pEYFP-Cl (Clontech) with oligos OMF90:

GTACGAATTCGATATCATGCATGGCGCCGTTTAAACGCGTATTTAAATGATCCGCTAGCGCT ACCG (SEQ ID NO:30) and OMF91:GATCAAGCTTGCGGCCGCGGATCCGCCCGGGCCACACAAAAAACCAACACACAGAT GTAATGAAAATAAAGATATTTTATTATCGATACTAGTGCGATCGCTTAATTAATTTAAATTT ACTTGTACAGCTCGTCC (SEQ ID NO:31) and ligated into pcDNA3.1/V5/His-TOPO to result in pSAM222. (ii) YFP was excised from pSAM222 by EcoKUNotl and ligated to ρMF164 (cut withEcoRI/Notl) thereby replacing GFP of pMF164 (P_PIR-GFP) and resulting in pSAM226 (P_PIR-YFP). (iii) YFP was eliminated from pSAM226 by Sw l and the remaining vector was religated to result in plasmid pMF189. pMF229 was constructed by a three-step cloning procedure: (i) The PIR3 element was excised from pMF208

and ligated to pTBCl (XhoUEcdRI) thereby replacing the tetracycline-responsive promoter P_hCMV*-ι of pTBCl and resulting in pMF210. (ii) pMF210 was restricted with SspllStul and the viral SN40 promoter excised from pSBCl with SspllStul was cloned in front of the PΕR.3 element to reconstitute P_PIROΝ and result in plasmid pMF222. (iii) The P_PIRON promoter was subsequently excised from pMF222 by SspUEcoRl and ligated to pMF189 restricted with SspVEcόRI thereby replacing P_PIR of pMF189 with P_PIRON and resulting in pMF229.

For construction of the tricistronic pTRIDENT expression vectors the P_PIR promoter was excised from pMF164 by SspVEcόRI and the P_PIRON promoter was excised from pMF222 by SspVEcoRI.

The P_PIR promoter elements were subsequently cloned:

1. into the SspVEcoRI sites of pTRIDENTl to replace P_hCVM and result in pTRIDENT 9 (P_PIR-MCSI-IRESI-MCSII-rRESII-MCSIII-pA).

2. into the SspVEc RI sites of pTRIDENT3 P^_VM^ and result in pTRIDENT 10 (P_PIR-MCSI- IRES-MCSII-CITE-MCSπi-pA).

The P_PIRON promoter elements were subsequently cloned:

1. into the SspVEcόRI sites of pTRIDENTl to replace P_hCVM*-ι and result in pTRIDENT 11 (P_PIRON-MCSI-IRESI-MCSII-IRESII-MCSIII-pA).

2. into the SspVEcόRI sites of pTRIDENT3 to replace P_hCVM*-ι and result in pTRIDENT 12 (P_P]RON-MCSI-IRES-MCSII-CITE-MCSiπ-pA).

Construction of the positive feedback regulation system using the streptogramin-responsive regulation concept

In contrast to the classical PIT/P_PIR system in which PIT and P_m reside on different plasmids, the positive feedback regulation concept places both elements in a single, often mutlticistronic, expression unit. In particular, the transactivator PIT is placed under control of its target promoter P_PIR.

In this configuration initial transcripts originating from the leakiness of the P_PIR promoter lead to few PIT molecules which are inactivated in the presence of stregtogramins. However, in the absence of this class of antibiotics initial PIT molecules can bind to and induce P_PIR. Since a PIT transcript is produced in every round of transcription, a principle called positive feedback, PIT accumulates in the cell and ensures high-level expression of the transgene of interest, yet this system retains full regulatability. Advantages of the positive feedback regulation system over classical binary regulated expression systems are:

33926. Tighter repression of gene expression since PIT is not expressed constitutively but originates from rare leaky transcripts. Therefore, in the repressed situation (in the presence of pristinamycin) little PIT is present in the cell which initiate transcription from P_PIR in contrast to the situation in which PIT is constitutively expressed from a separate vector.

33927. The positive feedback system produces a PIT molecule in every round of transcription leading to higher intracellular PIT levels and therefore also higher expression of the transgene of interest.

33928. The positive feedback regulation concept establishes regulated gene expression in a single step. The classical binary PIT/P_PIR expression systems requires first installation of PIT and then installation of the P_PIR-responsive gene. Two subsequent rounds of transfection and selection is not only tedious and time consuming but also undesired for advanced future therapies such as tissue engineering and gene therapy since the genome is changed significantly more than in a one-step engineering approach.

These positive feedback regulation vectors that were constructed contain both the green fluorescent protein GFP and PIT in a dicistronic, P_PIR-driven configuration. When pMF170 (PPIR-GFP-IRES-PIT-pA) was transfected in CHO-Kl, BHK-21 or HeLa cells bright green fluorescence could be observed by fluorescence microscopy in the absence of PI whereas GFP-expression was completely repressed in the presence of PI. Also constructed was pTPJDENT-PIT (pMF 175), which contains the pristinamycin- dependent transactivator in the first cisfron of pTRJ_DENT9. Cisfrons 2 and 3 of pTRIDENT-

PIT could accommodate two different genes of interest. Therefore, pTRDDENT-PIT derivatives enable one-step installation of streptogramin-responsive expression of up to two independent genes.

Construction of the positive feedback regulation vectors

5 For construction of the autoregulatory system, PIT was excised from pMF156 by

EcoRI/Hindlll and cloned into the polylinker of the eukaryotic expression vector pSBC-2 (EcoRI/Hindlll). As a pSBC-2 derivative, the resulting plasmid pMF169 is compatible to the pSBC-1 derivative pMF164 (P_PIR-GFP) and allows fusion of both expression units via their SspVNotl restriction sites, thereby generating pMF170. pMF170 contains a dicistronic P_PIR- regulated expression unit encoding GFP in the first and PIT in the second cistron. Cap- independent translation initiation of the second cistron is enabled by a picornaviral internal ribosomal entry site (IRES) encoded upstream of PIT.

For construction of pTRTDENT-PIT, PIT was excised from pMF156 by

2₅ EcoRI/Hindlll and cloned into pTRIDENT 1 was restricted with EcoRI/HmcOTI and PIT to result in plasmid pMF168. pMF168 was then restricted with SspVEcoRI and the P_hCMv_*-ι ^was replaced by P_PIR excised from ρMF164 (P_PIR-GFP) with SspVEcόRI to result in pMF175 which is designated pTRIDΕNT-PIT.

20 Example 6: Detection of novel antibiotic activities using the streptogramin-responsive expression technology

Streptogramins are unique antibiotics. Due to their composite nature (group A and group B streptogramins), they are less likely to elicit resistance mechanisms and show higher

25 bacteriocidal acitivity because of a synergistic effect of both compounds. For example, one particularly useful streptogramin antibiotic is Synercid®, a semi-synthetic, injectable pristinamycin derivative, which is active against a wide variety of multidrug resistant human pathogenic bacteria including vancomycin-resistant species. Infections of these vancomycin- resistant specie are causing increasing health concerns in both the US and Europe.

30 A sensitive streptogramin detection technology will have two major impacts on current environmental and medical concerns: (i) Detection of trace amounts of banned streptogramins such as virginiamycin in the human food chain for inspection purposes and

(ii) detection of novel streptogramin antibiotics in the culture supernatants of Actinomycetes

~_. and fungi.

RESULTS

The streptogramin detection technology is based on the pristinamycin-inducible

(PipON) system (although the pristinamycin-repressible version can also be used) which has been established in mammalian cells (CHO-Kl) stably or by cotransfection of pMF150 or 5 pMF225 (Pip expression vector) and pMF208 (P_PIRON-SEAP). The assay is based on the fact that presence of streptogramins in the test sample will induce the P_PIRON promoter and lead to a quantitative change in the SEAP readout.

Using 100 microliters of Streptomyces pristinaespiralis (a major producer of pristinamycin) supernatant as standard, we compared the detection potential of the pristinamycin-responsive system to classical antibiogram tests using a collection of streptogramin-sensitive human pathogenic bacteria including Staphylococcus aureus, Streptococcus pneumoniae, Streptococcus pyogenes, Listeria monocytogenes, Corynebacterium diphtheriae, Haemophilus influenzae and Neisseria gonorrhoeae as well as 25 the non-pathogenic Bacillus subtilis which is routinely used to screen for novel antibiotics effective against Gram positive bacteria. The presence of pristinamycin could be reliably detected in S. pristinaespiralis supernatants using the PipON system which resulted in about 20% of the SEAP readout typically reached using 2 μg/ml of pure pristinamycin I. However, using the same amount of supernatant in antibiogram tests, antibiotic activity could only be

20 detected in the hypersensitive strain of C. diphtheriae (19 mm of inhibitory diameter). All other pathogens tested including the standard antibiogram reference strain B. subtilis failed to detect pristinamycin in the S. pristinaespiralis supernatant despite of the synergy of streptogramins which has been reported to increase the bacteriocidal activity by a factor of

25 100 (Cocito et al, 1997, J. Antimicrob. Chemother. 39: 7-13).

Following efforts to detect new streptogramin derivatives we tested the culture supernatants of 14 novel Streptomyces isolates for the presence of antibiotic activity.

Antibiogram analysis of the supernatants using S. aureus, S. pyogenes, L. monocytogenes, C. diphteriae, andN. gonorrhoeae, and 10 ml of culture supernatant, only two supernatants

30 showed an antibiotic profile typical for streptogramins. One of these supernatants also induced the SEAP readout of the PipOΝ system about 3 -fold using only 100 microliters of supernatant which shows that the Pip-based streptogramin detection technology outperforms classical antibiogram test by at least a factor of 100. The second supernatant which was

~ positive in the antibiogram test decreased the cell viability when applied to the PipOΝ setup, indicating the potential antibiotic is cytotoxic to human cells. The correlation between antibiotic acitivity and induction of SEAP suggests that a novel Streptomyces species has been isolated which produces a streptogramine-like antibiotic activity. The screening technology using the pristinamycin-responsive expression strategy has the following advantages over classical antibiogram tests: (i) Antibiotic screening is not limited by the sensitivity of indicator bacteria to a yet uncharacterized streptogramin (sensitivity of bacteria to antibotics greatly varies between strains and even isolates), (ii) the mammalian cell-based streptogramin detection concept shows at least two order of magnitude higher sensitivity to this class of antibiotics than antibiogram tests based on bacteriocidal activity, (iii) the mammalian screening context ensures exclusive detection of bioavailable and non-cytotoxic antibiotics.

Example 7: Streptogramin-responsive expression systems for plant cells The ability to regulate transgene expression in plant cells or entire plants is an important tool for functional genomic research, repression of cloned genes which may be toxic (especially during the regeneration process) and production of protein therapeutics in plant tissue cultures.

MATERIAL AND METHODS

Cloning of the PIPpOFF and PIPpON systems

The PIPpOFF and PIPpON systems are binary systems which require a transactivator/transrepressor and a responsive target promoter. Here we describe the construction of the PIPpOFF system comprising the pristinamycin-dependent transactivator (PIT) and the cognate plant-specific pristinamycin-repressible promoters P_pPIR1 and P_pPIR2 and the PIPpON system which is based on plant expression constructs encoding the pristinamycin-inducible protein Pip/Pip-NLS and the pristinamycin-inducible plant promoters

^ PpPIRONl ^M αn^lr^ul ^ ppPIRONZ-

PIPpOFF

(a) PIT: Construction of a plant-specific expression configuration of PIT (Pip-NP16; SEQ ID ΝO:4; Example 1) was accomplished with a two-step cloning procedure: (i) PIT was amplified from the mammalian PIT expression vector pMF156 (P_hCMV-PIT-pA) using oligonucleotides OMF147: gatcggtaccggatccCCACCATGAGTCGAGGAGAGGTG (annealing sequence in capital letters) (SEQ ID NO: 32) and OMF149: gatcctgcagtctagaCTACCCACCGTACTCGTC (annealing sequence in capital letters) (SEQ ID NO: 33) and cloned into the eukaryotic expression vector pcDNA3.1/N5/His-TOPO

(Invitrogen) under control of P_hCMV to result in pMF270 (P_hCMV-PIT-pA). (ii) PIT was excised from pMF270 by BamΗVPstl (contained in OMF147: BamBΪ and OMF149: Pstl, underlined) and cloned into the corresponding sites (BamHVPstl) of pUCA7, a pUC18-derived plant expression vector, to result in pMF276. The PIT-encoding expression unit encoded on pMF276 is driven by the constitutive cauliflower mosaic virus 35S promoter (P_CaMV35sJ Odell et al., 1985, Nature 313: 810-812) and contains a 3' poly-adenylation signal derived from the octopine synthase gene (pA_0CS; Gatz et al, 1991, Mol. Gen. Genet 227: 229-237) (pMF276: P_CaMV35s-PIT-pA_ocs; Figure 10). (b) P_pPIR1 and P_pPIR2: The PTR3 element containing 9 Pip binding sites (SEQ ID NO:21,

Example 3) was excised from plasmid pPIR3 using EcoRN ISmal and cloned in sense (pMF279; P_pPIϊα; Figure 10B) and antisense (pMF281; P_p^; Figure 10B) orientation into the polished EcoRI site of pTT-GUS (EcoRI-TATA-box-GUS-pA_35S; Bδhner et al., 1999, Plant J.

19: 87-95) thus resulting in a PIR3-TATA-box-GUS-ρA_35S cassette in which the E. coli β- glucuronidase (GUS; Vancanneyt et al., 1990, Mol. Gen. Genet. 220: 245-250) is driven by

P_pPi_R! and P_pPIR2, respectively (PIR3-TATA-box; Figure 10B). The P_pPIR-driven GUS expression unit harbors a polyadenylation signal (pA_35S) derived from the cauliflower 35S gene (Vancanneyt et al., 1990, Mol. Gen. Genet. 220: 245-250). The TATA-box-containing minimal P_CaMV35 promoter (P_CaMV35s_mi_n) ^used here, has been previously optimized for low background activity in tetracycline-responsive plant gene expression systems (positions - 48/+1; Zuo and Chua, 2000, Curr. Opin. Biotechnol. 11: 146-151). In addition, P_CaMV35Smi_n harbors a C to A transition at position -45 to eliminate the potential CG methylation site thought to be responsible for the silencing phenomenon associated with this promoter (Bδhner et al., 1999, Plant J. 19: 87-95; Weinmann et al, 1994, Plant J. 5: 559-569).

PIPpON

(a) Pip/Pip-NLS: Cloning of the transrepressor Pip (SΕQ ID NO: 29; Example 4) into a plant cell-specific expression configuration included a two-step procedure: (i) Pip was amplified from pMF150 (P_hCMV-Pip-pA) using oligonucleotides OMF147: gatcggtaccggatccCCACCATGAGTCGAGGAGAGGTG (annealing sequence in capital letters) (SEQ ID NO: 34) and OMF148: gatcctgcagtctagaTCAGGCCTGTTCGACCATC (annealing sequence in capital letters) (SEQ ID NO: 35) and cloned in the sense orientation into pcDNA3. l/V5/His-TOPO to yield pMF269 (P_hCMV-Pip-pA). Pip was subsequently excised from pMF269 by BamHVPstl (contained in the oligos: OMF147: BamH , OMF148: Pstl underlined) and ligated into the corresponding sites (Bam VPstT) of the plant expression vector pUCA7 to generate pMF275 (P_CaMV35S-Pip-pA_0CS; Figure 11). For construction of Pip- NLS, Pip was amplified from pMF150 (P_hCMV-Pip-pA; Fussenegger et al., 2000, Nature Biotechnol., in press) using oligonucleotides OMF137: gatcgaattcggtaccgatatcggatocccaccATGAGTCGAGGAGAGGTG (annealing sequence in capital letters) ( SEQ ID NO: 36) and OMF138: gatccctgcagggcatgcttatcaagtcgactcgagcttctttctttgccttgacagttgagcactttccctattcctaaccaatctag ctctcttcttttcaggcctagtcgacgcagctggcgcgcggctgtacgcggaGGCCTGTTCGCCATC (annealing sequence in capital letters) and cloned in sense orientation) (SEQ ID NO: 37) into pcDNA3.1/V5/His-TOPO to result in pMF267 (P_hCMV-Pip-NLS-pA). OMF138 contains in its 5' extension the TGAlb NLS (bold; Bδhner et al., 1999, Plant J. 19: 87-95). Pip-NLS was excised from pMF267 by BamHVPstl (contained in the oligos OMF137: BamHI, OMF138:

Pstl underlined) and cloned into the corresponding sites (BamHVPstl) of pUCA7 to result in pMF273 (P_CaMV35S-Pip-NLS-pA) (Figure 11).

(b) P_pPIR0Ni and P_pPIRON2 ^: The pristinamycin-inducible plant gene promoter (P_pPIRON) was constructed following a two-step cloning procedure: (i) The P_CaMV35S-MCS-GUS-pA_nos cassette which contains a polyadenylation site derived from the nopaline synthase gene (Bevan et al, 1983, Nucleic Acids Res. 11 : 369-385) was excised from pBI121 (Jefferson et al., 1987, EMBO J. 6: 3901-3907) by HindllVEcoRI and cloned into pUC18 which resulted in pMF252. (ii) pMF252 was linearized by Smal and ligated to the PJ.R3 Pip-binding module excised from pPJJ .3 to result in pMF265 (PIR3 inserted in the antisense orientation; P_pPIRON.; P_CaMV35S-PIR3(AS)-GUS-pA_nos) and pMF266 (PIR3 inserted in sense orientation; P_pHRON2; Pc_aMV35s-PIR3(S)-GUS-pA_110S) (Figure 11).

Construction of the Nicotiana tabacum cell line SRI and cultivation

Callus formation was induced by placing leaf discs of in v/tro-propagated Nicotiana tabacum SRI plantlets on solidified LS (Linsmaier-Skoog) medium supplemented with sucrose (30 g/1), 2,4-dichlorophenoxyacetic acid (0.2 mg/1) and naphtalene acetic acid (0.19 mg/1). Leaf discs were incubated in the dark at 25°C. Developed callus tissue was removed from the discs and transferred to liquid media and adapted to single-cell suspension by shaking at 110 rpm.

Nicotiana tabacum suspension cultures were grown at 25°C in the dark in Linsmeyer-Skoog (LS) medium supplemented with sucrose (30 g/1) and: Thiamine (9.9 mg/ml), Mio-Inositol (100 mg/ml), 2,4-Dichlorophenoxyacetic acid (1 mg/ml), 1-Naphtalene acetic acid (2 mg/ml) under constant rocketing at 110 rpm. All ingredients were obtained from Sigma or Duchefa (Haarlem, NL). The suspension culture was maintained by splitting every 10 days 1:10 and removing cell aggregates by filtration using a metal sieve with a pore size of 0.125 mm.

Transient transformation of tobacco cells The transformation protocol of SRI cells was adapted from Wu and Feng (1999, Plant

Cell Reports 18: 381-386). In brief, 200 mg of exponentially growing SRI cells were harvested by centrifugation. The cells were incubated for 15 min. in 2% DMSO and washed twice in fresh medium. Cells were subsequently resuspended in electroporation buffer (5mM

CaCl₂; 10 mM NaCl₂; 0.4 M sucrose; 8.7% glycerol; 4 mM ascorbate; 10 mM HEPES; pH

6.8) and mixed with desired DNA at a concentration of 0.2 μg/μl. The cell-DNA mixture was incubated for 10 min. on ice electroporated at 700 N/cm and 980 μF using a BioRad Gene

Pulser and appropriate cuvettes. Following elecfroporation, cells were kept on ice for 10 min. and deplasmolyzed by 4 subsequent additions of 100 μl LS medium at 5 min. intervals. Transformation rates 20% ± 5% are typically reached with SRI cells using this modified protocol.

Quantification of β-glucuronidase (GUS) expression levels

Transformed SRI cells were incubated at 25°C for 48 h while rocking at 110 rpm.

During this time, GUS expression was induced (PIPpON) or repressed (PIPpOFF) by addition of 50 μg/ml of the human oral antibiotic Pyostacin^® which contains besides pristinamycin II about 15 μg/ml of the regulating compound pristinamycin I (PI). Cells were lysed in extraction buffer (50 mM ΝaHPO₄; 10 mM EDTA; 0.1% Triton X-100; 0.1% sodium lauryl sarcosine; 10 mM ?-mercaptoethanol; pH 7.0) by freezing them in liquid nitrogen and grinding them using a micropestle. Soluble protein extract was collected by centrifugation for 20 min. at 15,000 rpm and 4°C. Protein concentration of plant extracts was determined using a Bradford assay (Bradford, 1976, Anal. Biochem. 72: 248-254). The GUS expression assay was performed following a modified protocol by Jefferson et al. (1987; EMBO J. 6: 3901-3907). In brief, 100 μl of cell extract were incubated in extraction buffer (see above) also containing 1 mM of the 2-gmcuronidase substrate 4-methylumbelliferyl glucuronide (MUG; Duchefa, NL). The reaction mixture was incubated at 37°C and 100 μl aliquots were removed at appropriate time intervals and the reaction stopped by addition of

1.9 ml of 0.2 M K₂CO₃. The fluorescence time course was determined using typical 365 nm/455 nm excitation/emission profiles and a Shimadzu RF-5001 PC spectrofluorophotometer.

RESULTS

Construction of a pristinamycin-repressible gene regulation system for plant cells

(PIPpOFF)

The pristinamycin-repressible plant gene expression technology (PIPpOFF) is based on a pristinamycin-responsive plant transactivator (PIT; SEQ JX> NO: 4; Example NO: 1) which binds and activates chimeric plant promoters (R_pn ) in an antibiotic-dependent manner.

PIT comprises the Streptomyces coelicolor Pip repressor protein (SEQ ID NO: 29) fused to the VP16 transactivation domain of Herpes simplex virus and has been successfully used for streptogramin-responsive gene expression in mammalian cells. For use in plant cells, PIT has been cloned into a plant-specific expression configuration (pMF276: P_CaMV35S-PIT-pA_0CS,

Figure 10).

The PIT-responsive plant promoters P_pPj_R, and P_pPIR2 were constructed by fusing an artificial PIR3 element (SEQ ID NO: 21, Example 3) containing 9 Pip binding sites at regular intervals of two helical turns (21 bp) (placing them similar to the native configuration of the

Streptomyces pristinaespiralis pristinamycin resistance gene (ptr) on the same face of the

DNA double helix (Blanc et al., 1995, Mol. Microbiol. 17: 989-999; Salah-Bey et al., 1995,

Mol. Microbiol. 17: 1001-1012; Salah-Bey and Thompson, 1995, Mol. Microbiol., 17: 1109-

1119) in the sense (P_pPIR1) and the antisense (P_pHR2) orientation to a TATA-box (TATATAA) element derived from P_CaMV35S (-48/+1; Bδhner et al., 1999, Plant J. 19: 87-95; Odell et al.,

1985, Nature 313: 810-812). Expression vectors pMF279 and pMF281 harbor P_pPIRr (P_pPIR1-

GUS-pA_35S) and P_pPIR2- (P_pPIR1-GUS-pA_35S) driven GUS expression units, respectively (Figure

10). 5 Regulation performance of the PIPpOFF system in tobacco suspension cultures pMF276 (P_CaMV35S-PIT-pA_ocs, Figure 10) was cotransformed either with pMF279 (P_ppi_R.-GUS-pA_35S; Figure 10) or pMF281 (P_pPIR2-GUS-pA_35S; Figure 10) by electroporation into the Nicotiana tabacum cell line SRI. GUS activity was assessed after 48 hours using a fluorescence-based detection technology. Figure 3 shows the regulation performance of the PJPpOFF system in tobacco suspension cultures. Whereas in the presence of pristinamycin (50 μg/ml Pyostacin^®) the pristinamycin-repressible plant promoters P_pPIR1 and P_pPIR2 showed typical basal GUS expression levels (40.89 ± 10.28 and 31.89 ± 6.83 pmoles 4-MU min^"1 mg^"1 protein^"1, respectively) comparable with untransfected control cells (25.33 ± 17.67

25 pmoles 4-MU min^"1 mg^"1 protein^"1), reporter gene expression was induced about 6-fold in the absence of pristinamycin (Figure 12). The maximum expression levels of P_pPIR1 and P_pPIR2 was about 60% of the expression levels typically provided by the strong constititive P_{Ca V35}s promoter (Figure 13). The maximum expression levels of P_pPIR2 were significantly lower compared to P_pPIR1 -driven GUS expression suggesting that the orientation of the PIR3 cassette

20 (SEQ ID NO: 21, Example 3) and/or its distance from the TATA-box may influence the strength of such pristinamycin-repressible promoter configurations in plant cells.

Construction of pristinamycin-inducible plant gene regulation systems (PIPpON) 2 For regulated expression in transgenic plants and plant tissue culture an inducible rather than a repressible gene regulation system would be desirable. This enables induction of transgene repression by addition of the antibiotic rather than by its withdrawal. We therefore constructed a pristinamycin-inducible plant gene regulation system (PIPpON) which consisted of the Streptomyces coelicolor pristinamycin resistance gene repressor Pip (SEQ ID

^J 30^J NO: 29, Example 4) (Fussenegger et al., 2000, Nature Biotechnol., in press) which binds, in the absence of streptogramin antibiotics, to the PIR3 binding module (SEQ ID NO: 21,

Example 3) cloned downstream of a constitutive plant promoter (P_CaMV35s)- Binding of Pip to the PIR3 module blocks P_CaMV35S-mediated target gene expression. In the presence of

„ pristinamycin, Pip dissociates from PJ-R3 and full P_CaMV35S-driven expression is induced. For construction of PIPpON S. coelicolor Pip was cloned in a plant-specific expression configuration (pMF275, P_CaMV35s-Pip-PA,_os; Figure 11). In order to increase Pip concentrations in the plant nucleus which is expected to reduce basal expression levels of the PIPpON system, Pip was fused to a nuclear localization signal (NLS) derived from the plant 5 transcription factor TGAlb (PEKKRARLNRNRESAQLSRQRKKLEST (SEQ ID NO:32); Katagiri et al., 1989, Nature 340: 727-730; Van der Krol and Chua, 1991, Plant Cell 3: 667- 675) (pMF273, P_CaMV35S-Pip-NLS-pA_0CS; Figure 11).

The pristinamycin-inducible plant promoters P_pPι_R0Nι and P_pPIRON2 were constructed by cloning the PIR3 element (SEQ ID NO: 21, Example 3) in antisense (P_PPIRONI) ^and sense orientation (P_pPIR0N2) (Figure 11). Expression vectors pMF265 and pMF266 contain P_PPIRONΓ (Pc_aMV35s-PIR3(AS)-GUS-pA_ΗOS) and P_pPIR0N2- (P_CaMV35S-PIR3(S)-GUS-pA_nos) driven GUS expression units (Figure 11). Both plasmids show expression levels similar to the isogenic construct pMF252 (P_CaMV35S-MCS-pA-_nos) in the absence of a Pip-encoding construct or in the 25 absence of pristinamycin (see below and Figure 13). However, the maximal expression levels of P_pPIRON2 containing PIR3 in antisense orientation were significantly lower compared to P_pPIRONi suggesting that the orientation as well as the distance of a highly repetitive sequence like PIR3 may influence promoter strength (Figure 13).

20 Regulation performance of the PIPpON system in tobacco suspension cultures

PIPpON system was introduced into SRI tobacco suspension cultures by cotransformation of various combinations of transrepressor-encoding plasmids pMF273

(Pc_aMV35s-Pip-NLS-pA_0CS) or pMF275 (P_CaMV35s-Pip-pA_0cs) and P_pPIR0N-driven reporter

₂₅ constructs pMF265 (P_pPIR0Nι; P_CaMv₃₅s-PIR³(AS)-GUS-pA_nos) or pMF266 (P_pPIR0N2; P_CaMv_35S-

PIR3(S)-GUS-pA_I10S). In all PIPpON configurations tested (pMF273/265; pMF273/pMF266; pMF275/265; pMF275/266) the strong constitutive viral promoter P_CaMV3ss was completely repressed in the absence of pristinamycin (Figure 13). However, upon addition of 50 μg/ml pristinamycin to the cell culture medium the Pip-based transrepressors are released from the

3 ^u0 PIR3 module as shown in in vitro studies and in mammalian cells. Addition of the streptogramin antibiotic pristinamycin to the plant tissue culture resulted in 10-fold induction of GUS activity and maximum expression levels comparable to P_CaMV35S-driven expression

(Figure 13; pMF255, Pc_aMvsss-pAi_os) which compares favourably to induction factors (5.6-

„ fold) achieved by the ecdysone system in transient tobacco cultures (Martinez et al., 1999,

Plant J. 19: 97-106) (Figure 13). Fusion of Pip to a NLS (pMF273) neither increased the overall regulation performance of the PIPpON systems nor reduced their basal expression levels (Figure 13). DISCUSSION

We have established an alternative antibiotic-inducible gene regulation technology by adapting determinants of the Streptomyces pristinamycin resistance operon for use in plant cells. Ideal plant and mammalian cell-specific gene regulation systems share several characteristics. For example, they should show high induction ratios (low basal expression levels and high expression levels upon induction), rapid kinetics (fast induction and repression), and no pleiotropic effects or cytotoxicity.

Transgenic plants grown in contained greenhouses and plant tissue culture (mainly tobacco suspension culture) become increasingly important for the production of protein pharmaceuticals such as human interleukin (IL)-2 and IL-4 (Magnuson et al., 1998, Protein Expr. Purif. 13: 45-52), various therapeutic antibodies (Ma et al., 1998; Nature Med. 4: 601- 606) or edible human vaccines (Tacket et al., 1998; Nature Med. 4: 607-609). This is due to the simple and inexpensive cultivation technology and the absence of contaminating animal viruses, bloodborne pathogens, oncogenes and bacterial toxins. In this context, plant gene regulation systems may be essential for safe production of difficult-to-express proteins and agriculture-based large scale expression of desired proteins in a particular developmental stage. We have demonstrated here the successful use of novel plant gene regulation systems which are based on a Streptomyces pristinamycin resistance operon. The pristinamycin- responsive plant expression technology offers an attractive alternative to existing gene regulation technologies for challenging applications in basic plant research, agricultural applications and biopharmaceutical manufacturing.

Example 8: Retroviral expression vectors containing strepto ramin- and tetracycline- dependent transactivators

Successful gene therapy requires reliable delivery of therapeutic transgenes into a variety of human cell types. Replication-incompetent refroviruses are ideal vectors since they mediate DNA transfer, single-copy chromosomal integrations and expression of therapeutic fransgenes in target cell lines (Ausubel et al. 1995, Current Protocols in Molecular Biology (John Wiley &Sons, NY). In combination with pantropic packaging systems (Yee et al., 1994, Methods Cell Biol. 43: 99-112; Burns et al., 1993; Proc. Natl. Acad. Sci. USA 90: 8033-8037) refroviruses can be produced which infect a wide variety of cell types.

We have constructed retroviral vectors which can deliver the streptogramin- and tetracycline-dependent transactivators to variety of mammalian cell lines. pMF311 (5'LTR- ψ⁺-PIT-IRES-tTA-CITE*-zeo-3 'LTR) and pMF312 ((5'LTR-ψ⁺-PIT-IRES-rtTA-CITE*-zeo-

3 'LTR) were constructed following a multi-step cloning procedure: pMF311: (i) The streptogramin-dependent transactivator PIT (SEQ ID NO: 4, Example 1) was excised from pMF156 with EcoRI/Hindlll and cloned into the corresponding sites (EcoRVHindllT) ofpSBCl (Dirks et al., 1993, Gene 128: 247-249) to result in pMF167. (ii) The PIT-IRES fragment was excised from pMF167 with EcoRI/Notl and cloned into the corresponding sites (Ec RI/Notϊ) of pTRLDENT19 (Moser et al., 2000, Biotechnol. Prog., in press) resulting in pMF296 (P_hCMV*-l-PIT-IRES-MCS-CITE*-zeo-pA). (iii) the tetracycline- dependent transactivator (tTA, Gossen and Bujard, 1992, Proc. Natl. Acad. Sci. USA 89:

5546-5551) was amplified from pSAM201 (Fussenegger et al., 1997, Biotechnol. Prog. 13:

733-740) with oligonucleotides OMF159:

GATCGCGGCCGCCCACCatgtctagattagataaaagtaaa gtg (SEQ ID NO: 38) and OMF161: GATCTGATCAGATATCctacccaccgtactc gtc (SEQ ID NO: 39) and ligated in antisense orientation into pcDNA3.1/V5/His-TOPO (Invitrogen) to result in pMF307. (iv) tTA was excised from pMF307 with NotVEcoRN and ligated to pMF296 cut with NotVSrfl to give pMF309 (P_hCMV*__rPIT-IRES-tTA-CITE*-zeo- pA). (vi) pMF309 was restricted with EcoRI/RgtTI and the PIT-IRΕS-tTA-CITΕ*-zeo cassette was inserted into the retroviral vector pMSCVneo (Hawley et al, 1994, Gene Ther. 1: 136-

138) cut with EcoRI/R HI which resulted in pMF311 (5'LTR-ψ⁺-PIT-IRΕS-tTA-CITΕ*- zeo-3'LTR) (Figure 14). pMF312: (i) rtTA was amplified with OMF159 and OMF161 from pTET-ON

(Clontech) and ligated in antisense oritentation into pcDNA3.1/V5/His-TOPO (Invitrogen) to result in pMF308. (ii) rtTA was excised from pMF308 by NotVEcoRN and ligated to pMF296 cut with NotVSrfl to give pMF310 (P_hCMV*.,-PIT-IRES-rtTA-CITE*-zeo-pA). (iii) pMF310 was restricted with EcoRUBglR and the PIT-IRES-rtTA-CITE*-zeo cassette was inserted into the retroviral vector pMSCVneo cut with EcoRUBamHl which resulted in pMF312 (5'LTR-ψ⁺-PIT-mES-rtTA-CITE*-zeo-3'LTR) (Figure 14). Replication-incompetent pMF311 - and pMF312-derived refroviruses were produced using the panfropic retroviral packaging system (GP-293; Clontech). pMF311- and pMF312- based refroviruses were used to infect a variety of mammalian cell lines including TF-1 (ATCC# CRL-2003), K-562 (ATCC# CCL-243), PC- 12 (ATCC# CRL-1721), P19 (ATCC# CRL-1825), CHO-Kl (ATCC# CCL-61), C2C12 (ATCC# CRL-1772) and NIH3T3 (ATCC#

CRL-1772). After two days these cell lines were transfected with pMF201 (P_hCMV -CFP-pA_r

/-P_HR-YFP-pA_π) (Figure 5; Example 2) und examined by fluorescence microscopy. In all cell lines both fluorescent transgenes could be independently regulated by tetracycline (tet) and streptogramins (PI) (pMF311 -based retrovirus +pMF201: +tet/+PI: CFP_OFF/YFP_OFF; -tet/-PI:

CFP_ON/YFP_ON; -tet/+PI: CFP_ON/YFP_OFF; +tet/-PI: CFP_0FF/YFP_0N) (pMF312-based retrovirus +pMF201: +tet/+PI: CFP_0N/YFP_0FF; -tet/-PI: CFP_OFF/YFP_ON; -tet/+PI: CFP_OFF/YFP_OFF; +tet/-

PI: CFP_ON/YFP_ON).

Example 9: Streptogramin-responsive gene expression in transgenic mice

In order to demonstrate the potential of streptogramin-responsive expression technology in vivo, we produced transgenic mice. The expression units P_EF1._α-PIT-pA and

P_pi_R-GFP-pA ere excised from pMF255 (a pMF156 derivative: The human

EFlα (elongation factor lα) promoter (P_EFι._α) was excised from pEF4/Myc-His A

(Invitrogen) with SspVEcoRI and cloned SspVEcoRI into pMF156 thereby replacing P_hCMv) and pMF164, respectively and independently injected into mouse oocytes. These oocytes were reimplanted into separate pseudo-pregnant foster mice. Offspring were screened by

PCR for P_EF1._α-PIT-pA- and P_PIR-GFP-pA-containing offsprings (Hogan et al., 1994,

Manipulating the mouse embryo, 2nd edition, Cold Spring Harbor Laboratory Press, Cold

Spring Harbor, New York). Positive offsprings were crossed to generate double transgenic mice which contain the P_EF1._α-PIT-pA as well as the P_PIR-GFP-pA expression units. Vibratome sections of different organs of positive double transgenic offsprings were analyzed by GFP- mediated fluorescence. Double transgenic mice containing P_EFι._α-PIT-pA and P_PIR-GFP-pA expression units showed no GFP-mediated fluorescence in different organs (including muscle and brain) when streptogramins were supplemented in the drinking water. However, when these mice were watered with streptogramin-free drinking water, high level GFP expression could be observed. The foregoing written specification is sufficient to enable one skilled in the art to practice the invention. Indeed, various modifications of the above-described means for carrying out the invention which are obvious to those skilled in the field of molecular biology, medicine or related fields are intended to be within the scope of the following claims. All references cited herein are incorporated by reference.

Claims

What is claimed is:

1. A plant cell comprising a nucleic acid encoding a polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic.

2. The plant cell of claim 1 , wherein the nucleic acid hybridizes under moderate stringency conditions to a sequence that encodes the polypeptide encoded by SEQ ID NO:29.

3. The plant cell of claim 1, wherein the nucleic acid hybridizes under high stringency conditions to a sequence that encodes the polypeptide encoded by SEQ ID NO:29. 5

4. The plant cell of claim 1, wherein the nucleic acid hybridizes under moderate stringency conditions to the sequence of SEQ ID NO:29.

5. The plant cell of claim 1, wherein the polypeptide which binds to the P_ptr sequence is 0 derived from a bacterium selected from the group consisting of Actinomycetes.

6. The plant cell of claim 5, wherein the Actinomycete is a Streptomycete.

7. The plant cell of claim 1 , wherein the polypeptide which binds to the P_ptr sequence is operatively linked to a second polypeptide which activates or represses transcription in eukaryotic cells.

8. The plant cell of claim 7, wherein the polypeptide that activates transcription is ^υ selected from the group consisting of a VP16 activating domain, a GAL4 activating domain, a CTF/NFl activating domain, an AP2 activating domain, an ITF1 activating domain, an ITF2 activating domain, an Octl activating domain, a Spl activating domain, and the p65 domain of NF-κB. 5

9. The plant cell of claim 7, wherein the polypeptide that represses transcription is selected from the group consisting of a v-erbA oncogene product repressor domain, a thyroid hormone receptor repressor domain, a retinoic acid receptor repressor domain, a Drosophila Krueppel (Kr) protein repressor domain, a KRAB domain of the koxl gene family, an

5 even-skipped repressor domain, an S. cerevisiae Ssn6/Tupl protein complex, a yeast SIRI protein, NePl, a Drosophila dorsal protein, TSF3, SFl, a Drosophila hunchback protein, a

Drosophila knirps protein, a WT1 protein, an Oct-2.1 repressor domain, a Drosophila engrailed protein, E4BP4, ZF5, a 65 amino acid repressor domain of E4BP4 and an

N-terminal zinc finger domain of ZF5. 10

10. The plant cell cell of claim 8, further comprising a nucleotide sequence to be transcribed operatively linked to a P_ptr sequence.

2 11. The plant cell cell of claim 9, wherein the nucleotide sequence to be transcribed is endogenous to the host cell.

12. The plant cell of claim 9, wherein the nucleotide sequence to be transcribed is exogenous to the host cell. 20

13. The plant cell of claim 9, wherein the nucleotide sequence to be transcribed is a reporter gene.

25 14. The plant cell of claim 10, wherein the reporter gene is SEAP or GFP.

15. The plant cell of claim 1 which is derived from barley, wheat, rice, soybean, potatoe, arabidopsis, tobacco ox Nicotiana tabacum SRI.

⁰ 16. A plant-derived hairy root culture comprising the plant cell of claim 1.

17. The plant-derived hairy root culture of claim 16 which is derived from Artemisia, Atropa, Beta or Brugmansia.

35

18. A plant cell comprising a nucleic acid, said nucleic acid comprising a P_ptr sequence operatively linked to a first eukaryotic promoter.

19. The plant cell of claim 18, wherein the first eukaryotic promoter is operatively linked 5 to a first coding sequence.

20. The plant cell of claim 18, wherein the nucleic acid further comprises a second eukaryotic promoter.

21. The plant cell of claim 20, wherein the nucleic acid comprises at least one tet operator sequence is operatively linked to the second eukaryotic promoter.

22. The plant cell of claim 20, wherein the second eukaryotic promoter is operatively 25 linked to a second coding sequence.

23. The plant cell of claim 19 or 22, wherein at least the first or the second coding sequence contains an internal ribosomal entry site (IRES).

20 24. The plant cell of claim 23, wherein at least the first or the second coding sequence encodes a P_ptr -binding protein.

25. A method for regulating expression of a P_ptr -linked gene in a plant cell, comprising 25 introducing into the cell a nucleic acid molecule encoding a P_ptr -binding protein, thereby rendering the P_ptr -linked gene capable of regulation by an antibiotic that binds to the P_p(r - binding protein.

26. The method of claim 25, wherein the P_ptr -binding protein further comprises an

30 operably linked second polypeptide that activates or represses transcription in eukaryotic cells.

27. The method of claim 25, further comprising modulating the level of the antibiotic in the cell.

35

28. The method of claim 25, wherein the P_ptr -binding protein is derived from Streptomyces colelicolor, Amycolatopsis mediterranei, Streptomyces peucetius, Streptomyces

11 cyanogenus, Streptomyces hygroscopicus, Streptomyces glaucescens, Streptomyces 5 coelicolor.

29. The method of claim 26, wherein the polypeptide that activates transcription is selected from the group consisting of a NP16 activating domain, a GAL4 activating domain, a CTF/ΝF1 activating domain, an AP2 activating domain, an ITF1 activating domain, an

10

ITF2 activating domain, an Octl activating domain, a Spl activating domain, and the p65 domain of ΝF-κB.

30. The method of claim 26, wherein the polypeptide that represses transcription is

15 selected from the group consisting of a v-erbA oncogene product repressor domain, a thyroid hormone receptor repressor domain, a retinoic acid receptor repressor domain, a Drosophila Krueppel (Kr) protein repressor domain, a KRAB domain of the koxl gene family, an even-skipped repressor domain, an S. cerevisiae Ssn6/Tupl protein complex, a yeast SIRI protein, NePl, a Drosophila dorsal protein, TSF3, SFl, a Drosophila hunchback protein, a

20

Drosophila knirps protein, a WT1 protein, an Oct-2.1 repressor domain, a Drosophila engrailed protein, E4BP4, ZF5, a 65 amino acid repressor domain of E4BP4 and an

N-terminal zinc finger domain of ZF5.

25 31. A process for producing a protein comprising: culturing a plant cell containing: a P_ptr -linked gene that encodes the protein; and a nucleic acid molecule encoding a P_ptr -binding protein, and regulating expression of the P_ptr -linked gene by modulating the level of an antibiotic that binds to the P_ptr -binding

30 protein in the cell.

32. The process of claim 31, further comprising the step of collecting the protein produced by the cell.

35

33. The process of claim 31 or 32, wherein the P_ptr -binding protein further comprises an operably linked polypeptide that activates or represses transcription in plant cells.

34. A method for screening for a candidate antibiotic, the method comprising incubating a plant cell in the presence of a test compound, wherein the host cell comprises a polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic and a nucleotide sequence encoding a reporter gene operatively linked to a P_ptr sequence, and wherein a change in the transcription of the reporter gene indicates that the test compound is a candidate antibiotic.

10

35. The method of claim 34, wherein the polypeptide which binds to the P_ptr sequence is derived from an Actinomycetes bacterium.

25 36. The method of claim 35, wherein the Actinomycete is a Streptomycete.

11. The method of claim 36, wherein the polypeptide which binds to the P_ptr sequence is the pristinamycin induced protein from Streptomyces colelicolor.

20 38. A viral vector comprising a polynucleotide encoding a polypeptide which binds to a

P_ptr sequence in the absence of its cognate antibiotic.

39. The viral vector of claim 38, wherein the polynucleotide hybridizes under moderate 25 stringency conditions to a sequence that encodes the polypeptide encoded by SEQ ID NO:29.

40. The viral vector of claim 38, wherein the polynucleotide hybridizes under high stringency conditions to a sequence that encodes the polypeptide encoded by SEQ ID NO:29.

³ 41. The viral vector of claim 38, wherein the polynucleotide hybridizes under moderate stringency conditions to the sequence of SEQ ID NO:29.

42. The viral vector of claim 38, wherein the polypeptide which binds to the P_ptr sequence „ is derived from a bacterium selected from the group consisting of Actinomycetes.

43. The viral vector of claim 41, wherein the Actinomycete is a Streptomycete.

44. The viral vector of claim 38, wherein the polypeptide which binds to the P_ptr sequence is operatively linked to a second polypeptide which activates or represses transcription in eukaryotic cells.

45. The viral vector of claim 44, wherein the polypeptide that activates transcription is selected from the group consisting of a NP16 activating domain, a GAL4 activating domain, a CTF/ΝF1 activating domain, an AP2 activating domain, an ITF1 activating domain, an ITF2 activating domain, an Octl activating domain, a Spl activating domain, and the p65 domain of ΝF-κB.

5 46. The viral vector of claim 44, wherein the polypeptide that represses transcription is selected from the group consisting of a v-erbA oncogene product repressor domain, a thyroid hormone receptor repressor domain, a retinoic acid receptor repressor domain, a Drosophila Krueppel (Kr) protein repressor domain, a KRAB domain of the koxl gene family, an even-skipped repressor domain, an S. cerevisiae Ssn6/Tupl protein complex, a yeast SIRI

20 protein, NePl, a Drosophila dorsal protein, TSF3, SFl, a Drosophila hunchback protein, a

Drosophila knirps protein, a WT1 protein, an Oct-2.1 repressor domain, a Drosophila engrailed protein, E4BP4, ZF5, a 65 amino acid repressor domain of E4BP4 and an

N-terminal zinc finger domain of ZF5.

25

47. The viral vector of claim 38 which is an adenoviral vector, an adeno-associated viral vector, a retroviral vector, a lentiviral vector, an HIV vector or an HTLN vector.

48. A transgenic mouse having cells comprising a polynucleotide encoding a polypeptide

30 which binds to a P_ptr sequence in the absence of its cognate antibiotic.

49. A transgenic mouse having cells comprising a polynucleotide encoding a polypeptide which binds to a P_ptr sequence in the absence of its cognate antibiotic.

35

50. A transgenic mouse having cells comprising a polynucleotide, said polynucleotide comprising a P_ptr sequence operatively linked to a first eukaryotic promoter.

51. The transgenic mouse of Claim 50 wherein the eukaryotic promoter is operatively linked to a coding sequence.