Tetrahymena bifunctional Dihvdrofolate Reductase - Thymidylate Synthase

Deficiency and its Use

Field of the invention

The present invention addresses fields of recombinant molecular biology, in particular the use of a bifunctional marker enzyme in Tetrahymena enabling selection of transformands and facilitating drug discovery in parasitic Protozoa research.

Background of the invention

Tetrahymena is a ciliated eukaryotic unicellular organism belonging to the regnum of Protozoa and bearing two nuclei, a transcriptionally silent, diploid germline micronucleus (MIC) and a transcriptionally active, polyploid somatic macronucleus (MAC). In 1923, when Nobel Laureate Andre Lwoff succeeded in growing Tetrahymena in pure culture, the basis for exploiting this alveolate as a model organism was laid. Milestone discoveries made in Tetrahymena are the discovery of dynein motors, telomeres, RNA-mediated catalysis, telomerase and the function of histone acetyltransferases in transcription regulation. Within the last decades molecular biological techniques have been developed to alter Tetrahymena's genome and proteome: DNA transfection methods comprise inter alia microinjection into the MAC by electroporation and biolistic bombardement of MIC and MAC. Episomal plasmids based on an rDNA-replicon are available, as well as knock-out/-in techniques based on homologous recombination. On protein level, heterologous expression of related species has been performed and also endogenous proteins were silenced by a novel antisense-ribosome- technique. The advantages of using Tetrahymena in biotechnological applications include fast growth, high biomass, fermentation in ordinary bacterial/yeast

equipment, up-scalability as well as existence of cheap and chemically defined media.

So far, only a few markers that can be used in Tetrahymena have been described: ribosomal point mutation mediated resistances, a plasmid based neomycin resistance and a complicated beta-tubulin selection marker making use of an inducible promotor in combination with mutated tubulins being resistant / sensitive to the mitotic drug taxol \ Yet no true auxotrophic marker is available that permits selection without the use of antibiotics or drugs. This is where the present invention applies.

Description of the invention

Critical enzymes in pyrimidine biosynthesis are the enzymes dihydrofolate reductase (DHFR) and thymidylate synthase (TS). DHFR catalyses the production of tetrahydrofolate from dihydrofolate; TS is in charge of transferring a methyl-group from N ⁵ , N ¹⁰ -methylene-tetrahydrofolate to dUMP thereby generating dTMP and tetrahydrofolate. These enzymes being crucial for pyrimidine synthesis have been used as auxotrophic markers in various systems by targeted gene disruption, but also a number of inhibitors (antifolates) have been developed as anti parasite drugs. In animals, fungi and eubacteria the DHFR and TS gene are separately translated, whereas plants, Alveolata and Euglenozoa have a bifunctional fusion gene with both enzyme activities combined in one protein ("DHFR-TS").

The occurrence of the bifunctional enzyme in Tetrahymena pyriformis has been postulated in 1984 ² and 1985 ³ but no functional or even molecular biological analysis had been performed. A partial amino acid sequence of DHFR-TS of a non determined ¹ T. pyriformis- 1 ike strain" has been published in 2001 ⁴ , but this work is lacking any proof of linkage of the described partial cDNA to enzyme function.

The present invention provides a detailed characterization of the T. thermophila DHFR-TS gene including gene structure and functional data on the enzyme including data on in vivo function. In addition to a simple auxotrophic marker that is described for the first time in ciliates, the combination of these results with certain uncommon properties of Tetrahymena yields a surprisingly powerful tool for development and discovery of new antifolates against parasitic apicomplexans like for example Plasmodium sp. (malaria), Toxoplasma gondii (toxoplasmosis) and Cryptosporidium sp. (cryptosporidiosis).

Great efforts are being made to fight these most severe worldwide problems by targeting the DHFR-TS of the above mentioned parasites by antifolates. Initial success in the fight against malaria has been overcome by the rise of drug resistant Plasmodium strains. Especially in the case of malaria the diversity of DHFR-TS is manifold due to the occurrence of different strains (P. falciparum, P. vivax, P. malariae, P. ovale) and high frequency of DNA-recombination / - mutations producing new resistances over and over again. By revealing the crystal structure of the mutated and thereby resistant DHFR-TS enzymes, new small molecule drugs are being developed by computer aided molecular modelling. The major problem is the testing of the drug candidates for efficacy: As endoparasites are difficult to culture and have an unusual complex life cycle direct testing of the new compounds is nearly impossible. A detour has been described, that makes use of dhrl gene deficient yeast strains ⁵ expressing the functional, parasitic DHFR-TS enzyme of interest thereby reconstituting the dhrl deficiency. However this approach has a lot of disadvantages: The codon usage of parasitic apicomplexans, is very different from all commonly used expression- systems like E. coli, yeast and mammalian cell-lines and some tRNAs needed for translation are not abundant in these model organisms. Apicomplexa / Sporozoa, belonging to the phylogenetic group of Alveolata, possess a number of special cell biological features, the most notable being the presence of cortical alveoli, flattened vesicles packed into a continuous layer supporting the membrane. No model organism described so far has this assortment of special properties that will influence drug uptake and virtue.

- A -

AII the problems mentioned above can be overcome by the use of Tetrahymena:

It belongs to the Alveolata and is the most closely related species to apicomplexans. Tetrahymena is capable of growing in chemical defined media without the presence of thymidine arguing for a functional salvage pathway for dTMP synthesis by a bifunctional DHFR-TS enzyme. Deficiency in this enzyme activity should yield cell lines that only grow in media supplemented with thymidine. To destruct DHFR-TS activity various common techniques can be used like e.g. targeted gene knockout, site directed mutagenesis of essential amino acids and random mutagenesis. Screening for DHFR-TS deficient clones can be performed by replica plated culturing in media with (T+) or without thymidine (T-). Potential clones should only grow in T+ but not in T-. Successful reconstitution of enzyme activity can be achieved by transforming aforementioned cells with DNA fragments encoding for a functional DHFR-TS enzyme yielding cells that will grow on T-. This method allows selection of transformands without using drugs or antibiotics. In case of the exogenous, recombinant DHFR-TS being derived from parasitic apicomplexans an ideal model system for antifolate drug screening is achieved that can readily be used in a high throughput system.

A method for producing a ciliate cell with reduced or essentially no dihydrofolate reductase (DHFS) activity or reduced or essentially no thymidylate synthase (TS) activity or both reduced or essentially no dihydrofolate reductase and thymidylate synthase (DHFR-TS) activity is claimed, comprising the steps of

a) transforming ciliate cells by inserting a construct containing an allele altering the gene encoding the endogenous DHFR-TS into at least one of the endogenous DHFR-TS genes of the ciliate macronucleus (MAC),

b) inducing an allelic assortment process in the transformed ciliate cells to generate cells having the construct inserted in most or all functional DHFR-TS genes of the MAC, and

c) identifying the cells generated in step b) by cultivation with or without thymidine.

The method according to the invention can also utilize the fact that the micronucleus (MIC) stores the genetic information for sexual progeny. Consequently the method according to the invention also encompasses that in step a) the construct containing an allele altering the gene encoding the endogenous DHFR-TS is inserted into at least one of the endogenous DHFR-TS genes of the ciliate micronucleus (MIC), and in step b) the cells having the construct inserted in most or all functional DHFR-TS genes of the MAC are generated by breeding the cells of step a) with other ciliate cells to produce progeny that contains a new MAC derived from the altered MIC.

According to the invention it can be preferred that the ciliate is Tetrahymena, preferably Tetrahymena thermophila.

The DHFR-TS gene according to the invention can have a nucleotide sequence according to Seq ID No. 1.

It may be preferred that the region 1.5 kb up- and downstream of the cell's endogenous gene coding for the DHFR-TS bifunctional enzyme is additionally altered.

The method according to the invention can comprise an additional step of reconstituting the DHFR-TS activity by transfection of the cells with a DNA or RNA molecule coding for a functional non-endogenous DHFR-TS enzyme.

The term "non-endogenous" means that the DNA or RNA molecule is derived from a different organism, preferably from a different alveolate species.

Furthermore the method according to the invention can comprise an additional step of reconstituting the DHFR-TS activity by transfection of the cells with a DNA or RNA molecule coding for a functional endogenous DHFR-TS enzyme as well as another protein.

The DNA or RNA molecule coding for a functional endogenous or non- endogenous DHFR-TS enzyme can be derived from an alveolate, preferably a ciliate, even more preferably from Tetrahymena and most preferably from Tetrahymena thermophila. It may also be derived from an apicomplexan.

Furthermore the DNA or RNA molecule coding for a functional DHFR-TS enzyme can be represented by Seq. ID No. 1 and the amino acid sequence of the enzyme can be represented by Seq. ID No. 3.

The present invention also comprises a ciliate cell with an auxotrophy for thymidine. Preferably the ciliate cell according to the invention has reduced or essentially no DHFR activity or reduced or essentially no TS activity or both reduced or essentially no DHFR-TS activity. Said ciliate cell may be obtainable by using the methods as described herein.

A ciliate cell with reconstituted DHFR-TS activity is also disclosed, whereby the DHFR-TS enzyme is not the endogenous form. Said ciliate cell may be obtainable by using the methods as described herein.

Furthermore the present invention encompasses a ciliate cell with reconstituted endogenous DHFR-TS activity expressing another protein. Said ciliate cell may be obtainable by using the methods as described herein.

The use of the ciliate cells according to the present invention having a reconstituted non-endogenous DHFR-TS enzyme activity in an assay to detect chemical compounds affecting DHFR-TS enzyme activity is also claimed, comprising the steps of

a) bringing the cells into contact with the compound to be tested,

b) measuring the DHFR-TS enzyme activity of the cells, and

c) comparing the DHFR-TS enzyme activity to the DHFR- TS enzyme activity of control cells.

The ciliate cells according to the invention having a reconstituted endogenous DHFR-TS enzyme activity can be used for the production of another protein.

An isolated nucleic acid coding for the DHFR-TS protein having the nucleotide sequence according to Seq ID No.l and an isolated DHFR-TS protein having the amino acid sequence according to Seq ID No. 3 is also within the scope of the present invention.

Figure 1 : Genomic structure of T. thermophila DHFR-TS bifunctional enzyme

The DHFR-TS gene structure of T. thermophila consists of three exons (grey) and 2 introns (black). Primer pairs to amplify DNA for homologous integration and to amplify the CDS or cDNA are shown in the bottom part.

Figure 2: pKOI: DHFR-TS knockout construct

Figure 2 shows the knockout construct used for a targeted knockout of the DHFR-TS gene in T. thermophila. It consists of 3' and 5' flanking regions of the T. thermophila DHFR-TS gene and parts of its coding sequence (CDS), disrupted by a functional neomycin cassette conferring resistance to paromomycin.

Figure 3: Generation of DHFR-TS deficient strains by allelic assortment

Wildtype strains are transfected with pKOI. In one copy of the 45 ARPs the endogenous DHFR-TS gene is substituted by the knock out construct (step 2). By amitotic division of the MAC and high selection pressure clones will arise that have sorted out all endogenous DHFR-TS genes and retain only recombinant and defect DHFR-TS genes (step 3).

Figure 4: Selection of DHFR-TS knockout cells by growth on thymidine

Tetrahymena cells with disrupted DHFR-TS gene (clone 1-3) do not grow without the presence of thymidine (CDM-T), whereas wildtype cells do. Addition of thymidine to the medium (CDM+T) recovers growth.

Figure 5: Growth of DHFR-TS deficient Tetrahymena compared to wildtype

Growth kinetics of DHFR-TS knock out cells compared to wildtype cells in media with or without thymidine show that the knock out strain (pKOI) is growing as fast as wildtype cells on thymidine supplemented media (CDM+T). Knock out cells die without thymidine present (CDM-T). The curves are calculated by mean values of at least three independent experiments.

Figure 6: Proper integration of DHFR-TS knockout construct

This figure points up the PCR approach to determine that the knock out / -in construct has integrated into the DHFR-TS gene locus.

Three different cells were tested: Kl is a wildtype control, K2 are cells transfected with a plasmid carrying only the disrupting cassettes but no DHFR- TS gene sequences and pKOI D _V ι_ are cells transformed with the pKOI D _V ι_ plasmid. PCRl is a control reaction amplyfing 369 bp of the beta- hexosaminidase gene. PCR2 is to detect endogenous DHFR-TS (note that in the pKOI D _VL cells there still is a wildtype gene in the MIC!). PCR3 only yields PCR- product for correctly integrated pKOI D _VL DNA. PCR4 shows that the full-length expression cassette has integrated.

Figure 7: Enzyme function (DNase) of knock in target

Supernatants of three clones transformed with pKOI D _V ι_ were assayed for DNase activity. Only induced cells (+) show high levels of DNase activity. Transformed, but uninduced cells show a slightly elevated enzyme activity compared to wildtype cells due to low basal promotor activity.

Figure 8: Expression of knock in target

Western blots show expression of recombinant human DNase I: Only transformed and induced cells (+) show strong signals due to anti-DNase I antibodies. The intracellular PLAi-DNase-fusionprotein is visualized on the left blot by samples of cell lysates. Bands are running at higher molecular weight than the mature and processed positive control. On the right, supernatants were subjected to western blot; the size of the secreted protein of Tetrahymena argues for a correct processing when compared to the positive control.

Figure 9: Overview reconstitution of DHFR-TS activity by knock in

Step I shows the constructs needed for homologous recombination of different exogenous DHFR-TS genes into the disrupted Tetrahymena DHFR-TS gene locus. The strain described in example 1 is transfected with these DNAs enabling selection by medium deficient of thymidine (step III). After allelic assortment stable cell-lines with different recombinant DHFR-TS properties are obtained (step IV) that can be used in antifolate drug screening.

Figure 10: Enzyme (engogenous lipase) function of knock in target

Comparable to Figure 6 the gene of an endogenous Ciliate-Lipase were combined with an inducible promoter and flanking regions of the Tetrahymena

DHFR-TS gene. Subsequently the construct was integrated in a transformation vector (construct called pKOIX_M_Lipase) and transformed into Tetrahymena.

After homologous integration of DHFR-TS construct into Tetrahymena DHFR-TS locus and following selection of the clones (comparable to the described procedure in Fig. 3 - Fig. 6), supernatants of one clone transformed with

pKOIX_M_l_ipase were assayed for Lipase activity. Only cells, that have integrated pKOIX_M_Lipase construct, show high levels of Lipase activity. Wildtype cells show a low level of Lipase enzyme activity due to low promoter activity of the endogenous Lipase promoter. The lower graph shows the full- length integrative construct. Figure 10 shows the Lipase activity in the supernatant of wildtype cells and induced transformants measured with a Lipase enzyme assay according to the Reflectoquant® lipase test instructions of the manufacturer.

Figure 11 :

Figure 11 shows the construction of the full-length integration/expression cassette for lipase (pKOIX_M_Lipase), which was transformed into Tetrahymena cells in order to obtain over-expression of endogenous Ciliate-Lipase.

EXAMPLES

The following examples are provided to illustrate the embodiments of the present invention, but are not intended to limit its scope.

Cells and cell culture

Tetrahymena thermophila strains B 1868.4, B 1868.7 and B 2068.1 were kindly provided by Peter J. Bruns and cultivated in skimmed milk medium (2% skimmed milk, 0.5% yeast extract, 0.1% ferrous sulphate chelate solution and 1% glucose) in SPP or in CDM medium.

Amplification of the DHFR-TS gene of T. thermophila

The DHFR-TS cDNA gene and its 5 ^' and 3 ^' flanking sites can be amplified using the following primer pairs. Nucleotides in small letters encode sites for restriction endonucleases.

Amplification of the DHFR-TS 5 ^' flanking region :

DHFR 5 ¹ I F Notl: 5 ^' -cccgcggccgcACAGAGTTAATGGAAATGGAGC-3 ^' ,

DHFR 5'2 R BamHI: 5 ^' -gggggatccATATTTAAGCGATCTTTCAATGG-3;

Amplification of the DHFR-TS cDNA and gene with introns:

DHFR CDS F: 5 ^' -cgcGAATTCATGAAAACAAGACATTTTGATATAGTTTTAGC-3 ^' ,

DHFR CDS R: 5 ^' -gcgCTCGAGTCAGACAGCCATTTTCATTTATATTTTAGGG-3 ^' ,

Amplification of the DHFR-TS 3 ^' flanking region :

DHFR 3 ¹ I F XhoI: 5 ^' -gggctcgagATGCTCATGTTTACTCTAATCACG-3 ^' ,

DHFR 3 ¹ I R ACC65I: 5 ^' -gggggtaccAGTAAAAATAGAGTAGAAGGAG-3 ^' .

Construction of plasmids

The pKOI (knock out/in) plasmid was constructed as follows: As backbone for selection and propagation in E. coli the pBS II SK plasmid was used. The 1.5 kb 5 ^' -DHFR-TS integration site was amplified using the primer pair DHFR 5 ¹ I F Notl and DHFR 5 ¹ 2 R BamHI cloned into pBS II SK by using Notl and BamHI sites. Next the 1.4 kb paromomycin selection cassette from the pH4T2 (neo2) was cloned into the intermediate pBS IISK by BamHI and Smal sites. Finally, the 3 ^' -DHFR-TS integration site was amplified by primers DHFR 3' 1 F Xhol and DHFR 3' 1 R Acc65I and cloned by using the Xhol and Acc65I sites to finish the DHFR-TS knock out cassette.

The Sad site of the pBS II SK backbone had been destroyed by site directed mutagenesis to facilitate the use of the endogenous Sad site in the 3 ^' -DHFR- TS integrating sequence and the Xhol site as unique cloning site in pKOI. These sites were used to insert cassettes for the expression of recombinant enzymes (knock in). In this study we used a human recombinant DNase I expression/secretion cassette and a Ciliate-endogenous lipase expression/- secretion cassette. For DNase it consists of the first 115 aa of the endogenous PLAi precursor and the aa 23-281 of the mature human DNase I. To ensure proper translation of this fusion protein a codon optimised synthetic human DNase I gene was used. The expression was controlled by the previously

described ⁶ inducible MTT-I promotor, termination is regulated by the BTU2 terminator, like in the neo2 cassette of pH4T2 ⁷ .

Transformation of pKOI plasmids (biolistic bombardment)

We used conjugating cells, as well as vegetative, growing and non-conjugating stationary T. thermophila strains. The transformation of the T. thermophila cells was performed as previously described by Gaertig et al ⁸ .

Selection, allelic assortment and DHFR-TS knock out assay

T. thermophila cell proliferation assay: For the first ca. 16h after biolistic bombardment transformants were grown in skimmed milk medium. After that transformed cells were grown on SPP medium with in increasing concentrations of paromomycin (from 100 μg/mL to 1000 μg/mL) to support the allelic assortment process. After 3-4 weeks each clone was cultivated on CDM replica plates with or without thymidine (10 mg/mL). Functional DHFR-TS knock out clones are only able to grow in CDM medium supplemented with thymidine. The viability of the DHFR-TS knock out strains was monitored by determining the growth kinetic (Figure 5).

The lack of the endogenous wildtype DHFR-TS gene in the MAC as well as the complete integration of the DHFR-TS knock out and rhDNase I knock in cassette was confirmed by PCR using the following primers:

DHFROlF: 5 ^' -CTTTTTAACAGCCTGCTGCTCG-3 ^' ,

DHFR02R: 5 ^' -GATTTTGATGCTTCAATAAGGTTG-B ^' ,

DHFR03F: 5 ^' -TTATTTGTTTTATCATAGTGGAAAAGG-3 ^' ,

DHFR04R: 5 ^' -CAGACACCTCAATCATATCAAAG-3 ^' ,

DHFR05F: 5 ^' -GGTCCTCCATCAG ATTGTGG- 3 ^'

DHFR06R: 5 ^' -CGCGTCGAGTCAGACAGCCATπTCATTTA-3 ^'

HexOlF: 5 ^' -ATGCAAAAGATACTTTTAATTACTTTC-S ^'

HexO2R: 5 ^' - TATATTTTAGGAATGTTGTAATC-3 ^'

A pH4T2 plasmid carrying the same neo2 and DNase I expression/secretion cassettes was used a PCR control. The PCR strategy is illustrated in Figure 6.

SDS-PAGE and Western blot

Aliquots of transformed cells and of SPP supernatants were resuspended in sample buffer and separated on 15 % SDS-PAGE. The gels were blotted onto nitrocellulose membranes and blocked in PBS containing 0.05 % Tween 20 and 5 % skim milk (PBS-TM). The expression of recombinant human DNase I in transformed Ciliates was detected by two specific anti sera from rabbit against human DNase I (antigen : recombinant human DNase I, Pulmozyme, Roche). Both sera detected the recombinant DNase I antigen. The serum was used in a 1 : 500 dilution in PBS-TM. After washing with PBS/T an HRP-conjugated anti rabbit serum was applied. The blots were developed by using chemiluminescence.

DNase I activity assay

The methyl green based DNase activity assay was performed as already published ⁹ . Samples were incubated at 37 ⁰ C for 24 h on a microtiter plate. Absorbance was measured at 620 nm. Calibration of the assay was achieved by different amounts of defined DNase I Units of Pulmozyme from Roche (CHO derived) in each experiment and linear regression. These results combined with semi-quantitative western blotting were used to calculate the specific activity of expressed DNase I.

Lipase activity assay

The Lipase activity assay (Reflectoquant® test strips together with the Reflectometer RQflex®) was performed according to the instructions of the manufacturer (VWR International GmbH, Hilpertstraβe 20a, 64295 Darmstadt, Germany, article number: 1.05851.0001).

Example 1 : Generating auxotrophic heterokaryons

The protozoan T. thermophila belongs to the ciliates. These eukaryotes consist of two nuclei, a somatic macronucleus (MAC) and the genetic micronucleus (MIC). The MIC is the germline nucleus, i.e. it stores the genetic information for sexual progeny. The MIC is diploid and contains five pairs of chromosomes. In contrast to this the MAC is the somatic nucleus and no MAC DNA is transmitted to sexual progeny. The MAC contains 200-300 autonomously replicating pieces (ARP) that are derived from the MIC. Each of these units is present at about 45 copies except for the rDNA gene that is independently

amplified to ca. 10000 copies per cell. The MAC DNA is the transcribed DNA and therefore responsible for the actual phenotype of T. thermophila cells.

It is obvious that genetic engineering ultimately needs the transformation of the functional MAC that is in charge of protein expression. The first approaches for heterologous expression were done by using plasmids that use the vast amplification of the rDNA gene during MAC development. However the episomal presence of these plasmids depends on the drug concentration and the plasmid may recombine homologously and non-directionally into endogenous rDNA units.

The more promising way is the stable transformation of cells. This means the stable integration of expression cassettes into the MAC and or/the MIC of T. thermophila. MIC transformation can be achieved by stably transforming the chromosomal DNA of the diploid MIC. After conjugation of two different mating types the old MACs of the conjugating cells disappear and new MACs were built in the progenies derived from the recombinant MICs that carry the new information. The whole process follows the statistics of the Mendelian genetics. The advantage of this approach is that one obtains stable clones that maintain the genetic properties and that can be crossed via classical genetics to combine various properties of different T. thermophila strains. This approach is very elaborative and time consuming. Furthermore, it was shown recently that scan RNAs derived from the old MAC play an important role in DNA elimination during the development of the somatic MAC from the germline MIC. The primary sequence of these small RNAs explains how the parental MAC epigenetically controls the genome rearrangement in the new MAC. In the case of stable MIC transformants this RNAi-like mechanism inhibits the establishment and maintenance of foreign expression cassettes in the developing new MAC.

Instead of episomal transformation by rDNA based plasmids or the stable transformation of the MIC a shortcut by the combination of stable MAC

transformation with an immediate allelic assortment approach was used. This combination has several advantages. Firstly, the MAC transformation is much more efficient because there are at least about 45 potential integration sites per gene locus. Secondly, not only conjugating but also non-conjugating and therefore defined strains can be transformed. Finally, the recently reported genome rearrangements in the developing MAC that are regulated by scan RNA mechanisms can be short-circuited by MAC-transformation combined with allelic assortment.

To perform knock out / -in experiments the T. thermophila DHFR-TS gene and its flanking regions were amplified by using the primer pairs DHFR 5'1 F Not,

DHFR 5'2 R BamHI for the 5 ^' region non coding region, the primer pair DHFR

3 ¹ I F Xhol: 5, DHFR 3 ¹ I R Acc65I for the 3 ^' region and the primers DHFR CDS

F, DHFR CDS F for amplification of the coding region. The comparison of the coding region of the DHFR-TS structure gene to the DHFR-TS cDNA (amplified by the same DHFR CDS F /R primer pair) revealed the exon-intron architecture of the structure (Figure 1).

To start the DHFR-TS knock out experiments the plasmid pKOI was constructed (Figure 2). The neo2 cassette of pH4T2 was used to monitor the successful uptake of the plasmid by selection against paromomycin. The neo2 cassette of pKOI is flanked by the 1.5 kb fragments of the 5 ^' and 3 ^' region of the non-coding regions of the DHFR-TS gene, respectively (Figure 2). Because pKOI lacks an appropriate origin of replication, paromomycin resistant T. thermophila clones argue for a proper homologous recombination event in the DHFR-TS gene locus. Nevertheless, the most convincing evidence for the correct integration into the DHFR-TS locus is a loss of the DHFR-TS activity in the transformed strain. In the case of ciliates this requires the all complete replacement of all chromosomal DHFR-TS wildtype alleles (~45 ARPs) by the ARPs that includes the knock out cassette. We achieved this by allelic assortment. This allelic or phenotypic assortment is based on randomised distribution of the MAC chromosomes units (ARPs) during mitosis. In order to

force the assortment process into the desired direction - namely into the recombinant resistance gene - the transformed cells were cultivated for at least 2-3 weeks using increasing concentrations of the drug paromomycin. Single clones were isolated and tested for DHFR-TS deficiency by using a minimal chemical defined medium (CDM) with (+T) and without thymidine (-T). In Figure 4 and Figure 5 it is demonstrated that the found DHFR-TS knock out clones are real auxotrophic strains: The mutants are able to grow in CDM with thymidine like wildtype strains or strains with an incomplete allelic assortment. In CDM lacking thymidine they are unable to grow (see Figure 4 and Figure 5).

Example 2: Knock out DHFR-TS to knock in the gene of interest

The knock out of the endogenous DHFR-TS gene of T. thermophila also provides the possibility to knock in a further foreign gene that can be expressed heterologously in the DHFR-TS knock out strains. For this purpose the pKOI D _VL plasmid was constructed. It consists of a pKOI backbone with an additional expression cassette that encodes the first 115 amino acids (aa) of the precursor sequence of the PLAi gene and the mature human DNase I (amino acids 23 to 281). Construct for the expression of the endogenous lipase contains the endogenous-Ciliate-lipase-prepro peptide sequence and the mature endogenous Ciliate-lipase.

The PLAi prepro peptide (aa 1 to 110) has significant similarity to members of the cathepsin L family and mediates secretion into the medium. Similar to PLAi prepro peptide the endogenous-Ciliate-lipase prepro peptide mediates the secretion of the lipase into the medium.

In the case of the DNase the five additional amino acids (aa 111 to 115) should ensure an optimal cleavage of the pro PLAi-DNase I fusion protein by

endogenous pro-peptidases. In contrast to the neo2 cassette the expression of the ppPLA115-DNase I fusion protein is regulated by the inducible MTTl promotor. The inducible system was selected because it allows a clear discrimination between the DNase activity of heterologously expressed recombinant human DNase I and the basal activity due to at least two endogenous DNases. According to this the inducible system was also selected in the case of Lipase because it allows a clear discrimination between the basal level of the endogenous Lipase activity and homologous overexpressed endogenous Lipase acitivity. The transformation, selection of positive clones and the directed allelic assortment were done as outlined in example 1. Furthermore, the correct and complete integration of both expression cassettes (neo2 and the DNase I) in the DHFR-TS locus was tested by a PCR approach. Figure 6 shows that this is the case.

In order to demonstrate the pKOI concept, cells of these DHFR-TS knock out strains carrying the ppPLAu ₅ -DNase expression cassette were treated with and without Cadmium. Only induced strains showed an elevated DNase activity in the supernatant (Figure 7). To confirm this enzymatic data a specific antiserum against human DNase I was used to analyse the cell extracts and the supernatant of these human DNase expressing DHFR-TS knock out strains by western blot. The results illustrate that the DHFR-TS knock out strains are capable of expressing and secreting the functional recombinant human protein

(Figure 8).

Example 3 : Recombinant reconstitution of DHFR-TS activity to create strains for high throughput antifolate screening

The reconstitution of the DHFR-TS activity in DHFR-TS deficient Tetrahymena strains can be done by heterologously expressing bifunctional DHFR-TS enzymes of other Alveolata. As already mentioned in the introduction, this

phylogenetic group consists of the Ciliata, Dinoflagellata and Apicomplexa / Sporozoa. Especially the Apicomplexa are of high medical interest because all members are intracellular parasites and some of them cause very severe diseases (malaria, toxoplasmosis, cryptosporidiosis).

Apart from the fact that DHFR-TS deficient T. thermophila strains in combination with a DHFR-TS plasmid/vector provide a new marker for molecular biology, the same system can be used to generate defined test strains for drug development applications. Due to the close phylogenetic relationship it is also possible to recover the DHFR-TS activity in deficient Tetrahymena strains by DHFR-TS enzymes from other members of the Alveolata group. This implies the application of DHFR-TS deficient T. thermophila strains in searching for and testing novel anti DHFR-TS drugs (antifolates). This is of very high importance because antifolates have been one of the most promising drugs in the fight against malaria. The fact that anti DHFR-TS drugs have been administered for a long time in e. g. malaria patients is the reason that during the last decades many of the parasites became resistant. At present many approaches are being explored to find new and more specific and efficient drugs against the parasite DHFR-TS enzyme. As the search for new antifolates relies on structure determination using NMR and crystallization techniques, most of these approaches are very time and cost intensive.

DHFR-TS deficient T. thermophila strains that express an active parasite DHFR-TS will not only be able to grow in medium lacking thymidine, but they are also suitable to search for novel antifolates. Thus the DHFR-TS deficient strains recovered by the homologous bienzyme of the parasites represent a simple and flexible in vivo test system conferrable to other disease causing parasites of the Apicomplexa group. For example antifolates against Plasmodium vivax strains have not been considered because no appropriate and flexible in vivo system for high throughput approaches was available. This can be done easily by recovering the DHFR-TS activity of T. thermophila

strains by using the homologous bienzymes of Plasmodium vivax, Plasmodium malariae, Plasmodium ovale as well as enzymes from Apicomplexa / Sporozoa that cause toxoplasmosis or cryptosporidiosis. Figure 9 illustrates this concept. The DHFR-TS enzyme of the pyrimethamine-sensitive Plasmodium falciparum strain 3D7 was used to recover the DHFR-TS activity in a T. thermophila strain, lacking DHFR-TS activity. In parallel a DHFR-TS from a Plasmodium falciparum strain that is resistant against the pyrimethamine treatment (for example the DHFR-TS S108N mutant) is tested. By measuring the effect of pyrimethamine in dose response curves on the two different transformands drug resistance and sensitivity can be determined easily. Subsequently new potential drugs can be added to the medium to show effectiveness on yet resistant DHFR-TS mutants.

Reference List

1. Gaertig, J., Thatcher, T. H., Gu, L. & Gorovsky, M. A. Electroporation- mediated replacement of a positively and negatively selectable beta- tubulin gene in Tetrahymena thermophila. Proc. Natl. Acad. Sci. U. S. A

91, 4549-4553 (1994).

2. Garrett, C. E. et al. A bifunctional thymidylate synthetase-dihydrofolate reductase in protozoa. MoI. Biochem. Parasitol. 11, 257-265 (1984).

3. Kotsifaki, H., Kapoulas, V. & Deliconstantinos, G. Targeting of liposomes containing methotrexate towards Tetrahymena pyriformis cells. Gen.

Pharmacol. 16, 573-577 (1985).

4. Stechmann, A. & Cavalier-Smith, T. Rooting the eukaryote tree by using a derived gene fusion. Science 297, 89-91 (2002).

5. Hastings, M. D. & Sibley, C. H. Pyrimethamine and WR99210 exert opposing selection on dihydrofolate reductase from Plasmodium vivax.

Proc. Natl. Acad. Sci. U. S. A 99, 13137-13141 (2002).

6. Shang, Y. et al. A robust inducible-repressible promoter greatly facilitates gene knockouts, conditional expression, and overexpression of homologous and heterologous genes in Tetrahymena thermophila. Proc. Natl. Acad. Sci. U. S. A 99, 3734-3739 (2002).

7. Gaertig, J., Gu, L., Hai, B. & Gorovsky, M. A. High frequency vector- mediated transformation and gene replacement in Tetrahymena. Nucleic Acids Res. 22, 5391-5398 (1994).

8. Gaertig, J. & Gorovsky, M. A. DNA-mediated transformation in Tetrahymena. Methods Cell Biol. 47, 559-569 (1995).

9. Sinicropi, D., Baker, D. L., Prince, W. S., Shiffer, K. & Shak, S.

Colorimetric determination of DNase I activity with a DNA-methyl green substrate. Anal. Biochem. Ill, 351-358 (1994).