RIBOSWITCH INDUCIBLE GENE EXPRESSION

Abstract

An intronic, self-splicing riboswitch is configured for enzyme-product specificity by introducing an appropriate aptamer. This then provides a sensing-expression construct, whereby the presence of an enzyme product in the cell triggers self-splicing of the intron sequence to restore the reading frame of the reporter gene and as such to drive expression of the gene product. The sensing construct expresses a protein which marks the cell or permits its growth or survival in or on an otherwise selective media. In this way, introduction or the presence of such product sensing-reporter constructs in cells can be harnessed to provide a multi-parallel rapid screening of cells or libraries for desirable enzyme variants.

Claims

1. A method of inducing expression as claimed in claim 3, wherein the cell is provided by a step of transforming an host cell with the polynucleotide expression construct.

2. (canceled)

3. A method of inducing expression of a desired RNA molecule, or a desired protein or a polypeptide in a cell comprising: (i) providing a cell which contains a polynucleotide expression construct, the construct comprising a polynucleotide sequence to be transcribed, and wherein the polynucleotide sequence is interrupted by at least two introns which are self-splicing introns and whose splicing activity is under the control of an aptamer, and wherein the aptamer has binding affinity for an inducer; (ii) exposing the transformed cell to the inducer to activate self-splicing activity of the introns and thereby produce the RNA molecule, or expression of the protein or polypeptide.

4. (canceled)

5. A method of inducing expression as claimed in claim 3, further comprising transforming the host cell with a second expression construct, wherein the second expression construct comprises a polynucleotide sequence whose expression is under the control of a promoter which is under the control of the RNA, protein or polypeptide product of the first expression construct, and the RNA product of the second expression construct is the desired RNA product, or the expressed protein or polypeptide expression product of the second expression construct is the desired protein or polypeptide.

6. (canceled)

7. A method as claimed in claim 5, wherein the cell is transformed with the first and second constructs separately, simultaneously, or sequentially.

8. (canceled)

9. (canceled)

10. A method of inducing expression as claimed in claim 5, wherein the cell contains a second expression construct comprising a polynucleotide sequence whose expression is under the control of a promoter which is selected from the group consisting of a) inducible by the RNA product of the first expression construct, and the transcribed RNA of the second expression construct is the desired RNA; and b) under the control of the expressed RNA, protein or polypeptide of the first expression product and expressed protein or polypeptide of the second expression construct is the desired protein or polypeptide.

11. (canceled)

12. A method of inducing expression as claimed in claim 3, wherein the RNA is a microRNA (miRNA), a small interfering RNA (siRNA), an antisense RNA, a tRNA or a ribozyme.

13. (canceled)

14. A method of inducing expression as claimed in claim 3, wherein the at least two self-splicing introns each comprise an aptamer and wherein the aptamer is the same.

15. A method of inducing expression as claimed in claim 3, wherein the polynucleotide sequence of the first expression construct encodes Phage T7 DNA dependent RNA polymerase, optionally wherein the second expression construct comprises the promoter PT7.

16. A method of inducing expression as claimed in claim 3, wherein the aptamers bind a ligand; optionally wherein the ligand is selected from one of theophylline, tetracycline, neomycin or malachite green.

17. A method of inducing expression as claimed in claim 3, wherein the self-splicing intron is the T4 td gene self-splicing intron.

18. A cell for inducer molecule-controlled expression of a gene product, the cell comprising a polynucleotide expression construct, wherein the construct comprises a polynucleotide sequence which is interrupted by at least two introns which are self-splicing introns and whose splicing activity is under the control of an aptamer, and wherein the aptamer has binding affinity for the inducer.

19. A cell as claimed in claim 18, wherein the expressed gene product is a) an RNA; optionally one of a microRNA (miRNA), a small interfering RNA (siRNA), an antisense RNA, a tRNA or a ribozyme; or b) a protein or polypeptide.

20. (canceled)

21. A cell as claimed in claim 18, further comprising a second expression construct comprising a polynucleotide sequence whose expression is under the control of a promoter whose activity is regulated by the expression product of the first expression construct.

22. A cell as claimed in claim 21, wherein the expressed product of the second expression construct is an RNA molecule; optionally one of a microRNA (miRNA), a small interfering RNA (siRNA), an antisense RNA, a tRNA or a ribozyme.

23. A cell as claimed in claim 21, wherein the expressed product of the second expression construct is a protein or polypeptide.

24. A cell as claimed in claim 18, wherein the two or more self-splicing introns contain the same aptamer.

25. A cell as claimed in claim 18, wherein the or each self-splicing intron is the T4 td gene self-splicing intron.

26. (canceled)

27. A kit for preparing an host cell for inducible host cell expression of a gene product, comprising: (i) a first polynucleotide expression construct, the first construct comprising a polynucleotide sequence to be transcribed, and wherein the polynucleotide sequence is interrupted by at least two introns which are self-splicing introns and whose splicing activity is under the control of an aptamer, and wherein the aptamer has binding affinity for an inducer; (ii) a second polynucleotide expression construct, the second construct for receiving a polynucleotide sequence which encodes an RNA or protein or polypeptide to be expressed, and wherein the polynucleotide to be expressed is under the control of a promoter which is inducible by the transcribed polynucleotide; optionally (iii) a set of instructions for inserting a polynucleotide sequence to be expressed into the second polynucleotide.

28. A kit as claimed in claim 27, further comprising: (i) a host cell comprising a polynucleotide expression construct, wherein the construct comprises a polynucleotide sequence which is interrupted by at least two introns which are self-splicing introns and whose splicing activity is under the control of an aptamer, and wherein the aptamer has binding affinity for an inducer; (ii) a second polynucleotide expression construct, the second construct for receiving a polynucleotide sequence which encodes an RNA or protein or polypeptide to be expressed, and wherein the polynucleotide to be expressed is under the control of a promoter which is inducible by the transcribed polynucleotide; optionally (iii) a set of instructions for (a) inserting a polynucleotide sequence to be expressed into the second polynucleotide and/or (b) transforming the host cell with the second polynucleotide expression construct; optionally further comprising a container containing an inducer; wherein for example the inducer is theophylline.

29. A kit as claimed in claim 27, wherein the self-splicing intron is the T4 td gene self-splicing intron.

30. (canceled)

Description

DETAILED DESCRIPTION

[0090] The invention will now be described in detail with reference to the examples and to the drawings in which:

[0091] FIG. 1A shows a common known mechanism of engineered riboswitches in bacteria. For regulation of translation initiation. When there is no ligand a stem-loop is formed by the aptamer domain and by a sequence 5 to and complementary to the Shine Delgarno region, the latter being accessible to the 30S RNA for binding and therefore translation initiation occurs. When ligand binds aptamer then a folding of the aptamer domain occurs as well as an alternative stem loop structure involving the SD region. This then blocks 30S binding and stops translation initiation.

[0092] FIG. 1B shows a regulation of transcription termination. The aptamer is fused to a short spacer region followed by a sequence complementary to the 3 part of the aptamer and a U stretch. When no ligand is present then the complementary 3 part is base paired with the aptamer forming a terminator structure so that RNA polymerase (RNAP) dissociates and transcription is blocked. When ligand binds aptamer then this terminator structure formation is inhibited and transcription can proceed.

[0093] FIG. 1C shows an illustration of one of the mechanisms of a riboswitch located in the 5-UTR of bacterial mRNA.

[0094] FIG. 2 shows the structures of two introns as shown in Thompson et al., 2002 BMC Biotechnol 2:21. This drawing is from 2002 Thompson et al; licensee BioMed Central ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL: http://www.biomedcentral.com/14726750/2/21. Panel A shows the parental self-splicing intron of phage T4 (which does not need an inducer to splice). Panel B shows an engineered variant, in which the P6a-loop (see panel A, box) has been replaced by a theophylline-binding aptamer; binding of theophylline induces a conformational change that triggers splicing activity (see panel A, arrows indicate cleavage in loops P1 and P10).

[0095] FIG. 3 shows part of the pyrimidine synthesis pathway, dTMP is a key link in the synthesis of DNA. The pathway from thymine to dTMP is not supported by E. coli, while the pathway from DNA to dTMP does not support growth. Bacteria can therefore be grown under non-selective conditions by adding thymidine (and indeed not thymine) to the medium, a compound which has a very low abundancy in common medium components such as tryptone, peptone and yeast extract.

[0096] FIG. 4 shows the relationship between growth rate of single intron thyA constructs (i.e. those with a single self-splicing intron) and promoter strength. E. coli DH10B-thyA carrying the reporter constructs were grown with different amounts of theophylline and monitored for 20 h.

[0097] FIG. 5 shows a comparison of the phage T4 td intron with the mutant td intron. The mutant td intron differs only slightly from the phage T4 td intron in the 5 and 3 flanking sequence and retains its activity. The flanking regions which are part of the riboswitch, but not of the intron (they are part of the exons) are WT and mutated in intron 1 and 2 respectively.

[0098] FIG. 6 shows the growth rate of both single and double intron thyA constructs under control of the P.sub.tacI promoter under varying theophylline concentrations. Intron 1 () is the intron at (F171-P175), intron 2 () is the intron at (H51-I55). A construct featuring no intron in the thyA gene () is also shown.

[0099] FIG. 7 shows vector maps for dTMP auxotrophy complementation. Panel A shows the intron insertion like in the phage T4 td intron situation. Panel B shows the introduction of the functional intron upstream of the wild type position, such that the introduction is silent in the sense that no amino acids were changed in the protein sequence into which the intron is inserted. Panel C shows the tandem introns for improved control of expression.

[0100] FIG. 8 shows the structure of a self-splicing intron (Cech et al., 1994). In loop 6a, sequences have been inserted that affect the splicing activity. Insertion of ligand-binding RNA fragments (aptamers) controls the splicing by a conformational change, triggered by the presence or absence of a specific ligand.

[0101] FIG. 9 shows the structure of a small RNA fragment with an aptamer. The aptamer sequence can be varied using Systematic Evolution of Ligands by Exponential Enrichment (SELEX) and screened for novel specificities.

[0102] FIG. 10A shows a single self-splicing intron construct.

[0103] FIG. 10B shows a double self-splicing intron construct.

[0104] FIG. 10C compares the structures of pSC024, pSC026, pSC018 and pSC022.

[0105] FIG. 10D shows a schematic of the reporter cascade described for GFPuv production, although any protein may be produced. The T7 polymerase gene is carried on the bacterial genome. The enzyme's functionality depends on maturation (RNA splicing that restores open reading frame), which is controlled by theophylline in case of the heterologous expression system, or a compound derived from an enzymatic reaction when the system is utilised as a screening method.

[0106] FIG. 11 shows the structure of expression vectors for GFPuv expression dependent on rhamnose and theophylline.

[0107] FIG. 12 shows the vector map for the reporter plasmid pSC028-GFPuv-term. pSC028-GFPuv-term was constructed using pACYC184 as a base.

[0108] FIG. 13 shows theophylline dependency of the cascade at 0.8 mg/L rhamnose. The PtacI serves as benchmark for a strong promoter dependent on E. coli RNA polymerase and is the pSC034f construct.

[0109] FIG. 14 shows expression of GFPuv over a wide range of rhamnose and theophylline concentrations.

[0110] FIG. 15 shows the sensitivity of the cascade, showing the theophylline dependency of the cascade in response to differing concentrations of rhamnose. The cascade is controlled by one theophylline dependent intron at the first position depicted in FIG. 10. In particular expression of GFPuv using the single intron driven cascade over a wide range of rhamnose and theophylline concentrations is shown.

EXAMPLE 1

Evaluation of T4 Intron-Controlled ThyA Auxotrophy

[0111] An in vivo biosensor is usually composed of a control element and a reporter gene. The reporter gene can confer antibiotic resistance, fluorescence, auxotrophy complementation or luminescence. The control element can act on several stages in the protein production process. Protein based control elements like LacI typically intervene with the transcription of the gene, while inteins and post-translational modification deal with activation of the protein itself. In between, there is the control on translational level predominantly performed by riboswitches. The riboswitches are mostly located in the 5-UTR sequestering and releasing the Shine-Dalgarno sequence to block or allow translation by the ribosome. For example, FIG. 1 shows the operation of a generalised riboswitch, illustrating how structural changes in the RNA fragment induced by the binding of a signal metabolite may result in the reduced accessibility of the Shine-Dalgarno sequence and the blockage of translation. Several cases have been described where the 5 untranslated region (5 UTR) of mRNA transcripts form secondary structures (e.g. stem-loop structures) via intra-molecular base paring. The structure of an RNA fragment may change upon specific binding of a ligand, and may affect the accessibility of the ribosome binding site (Shine-Dalgarno sequence) and as such the translation efficiency (riboswitch on/off). Variation of the sequence of the signal metabolite (ligand)-binding sub-fragment (aptamer) may adjust the specificity of the riboswitch.

[0112] It may be difficult to alter the ligand of these riboswitches, since the anti-Ribosome Binding Site (anti-RBS) or anti-terminator may be part of the aptamer domain. Altering the aptamer domain will change the anti-RBS or anti-terminator rendering the riboswitch inactive. Randomising the 5-UTR and testing for riboswitch activity is one solution to that problem. Another possibility is changing a ribozyme, either a synthetic or a natural one, into a riboswitch by attaching an aptamer domain. This creates an allosteric ribozyme also referred to as aptazyme. An aptazyme that was based on the hammerhead ribozyme was designed by Ogawa and Maeda (2007) Bioorg Med Chem Lett 17:3156-3160.) (This aptazyme is based on SD sequestering and the block is released by the endonuclease activity of the ribozyme upon induction. A different approach using the same mechanism can be applied in eukaryotes cutting of the poly-A tail upon induction. A type of synthetic riboswitch that has not been studied extensively is the group I aptazyme. This aptazyme is a modified version of a group I self-splicing intron. The intron that was modified is derived from the phage T4 td gene encoding thymidylate synthase (see FIG. 2). This gene has a homologue in E. coli named thyA. In the engineered variant, derived from the parental self-splicing intron of phage T4 (both shown in FIG. 2), the P6a-loop (see panel A, box) has been replaced by a theophylline-binding aptamer. Binding of theophylline induces a conformational change that triggers splicing activity (see panel A, arrows indicate cleavage in loops P1 and P10).

[0113] This system has many properties that make it suitable as an in vivo biosensor. Contrary to riboswitches that block the Ribosome Binding Site (RBS), there is only one way leakage can occur; when blocking the RBS the block may be released by complete unfolding of part of the mRNA. When the intron is unfolded it does not splice out of the mRNA, still disallowing functional translation. The leakage that will occur in both instances is when the aptamer is not completely destabilised when no ligand is present and the switch is flipped in the absence of a trigger. The gene the intron is naturally present in is an essential gene that can be complemented easily by adding thymidine to the medium or supplying the gene on a plasmid. By default, no large amount of thymidine is present in most widely used media like LB, so all experiments can be performed on rich media. The selection allows for a great number of variants to be tested without expensive equipment or labour intensive experiments. A property the td intron based riboswitch shares with other riboswitches is the transferability to other organisms, as riboswitches are unaffected by post-translational modification. Group I introns have the extra advantage that no species-specific elements like SD sequence or poly-A tail are involved. This experiment focuses on the exact conditions the theophylline responsive T4 td intron needs to function as selection tool in E. coli.

[0114] Auxotrophy Complementation with thyA Under Control of a Theophylline Dependent intron

[0115] thyA is a gene encoding thymidylate synthase, a crucial part of the pyrimidine synthesis pathway. It catalyses the reaction from dUMP to dTMP using THF as a cofactor (see FIG. 3). As a precursor of dTTP, which is required for DNA synthesis dTMP is a key link in the synthesis of DNA. The pathway from thymine to dTMP is not supported by E. coli, while the pathway from DNA to dTMP does not support growth. dTMP can also be derived from thymidine when this is added to the medium, but not thymine as E. coli DH10B lacks the enzyme to convert thymine into thymidine. The ThyA deficient bacteria cannot grow on rich medium without addition of thymidine.

[0116] The relationship between the growth rate of single intron constructs (i.e. those with a single self-splicing intron) and promoter strength was determined.

[0117] The theophylline dependent phage T4 td intron was designed according to Thompson et al (2002) BMC Biotechnol 2: 21. The intron flanking regions are identical on protein level between the thyA gene of E. coli and the td gene of phage T4 allowing for introduction of the intron into thyA with silent mutations only.

[0118] Reporter constructs carry a p15A origin of replication derived from pACYC184, a kanamycin resistance gene from pET24d and the 5-UTR and CDS of E. coli thyA. A terminator and promoters of different strength were placed upstream of the 5-UTR. All prompters are listed in Table 1. E. coli DH10B-thyA carrying the reporter constructs were grown with different amounts of theophylline and monitored for 20 h.

[0119] The log phase growth rate was determined in biological triplicate for each construct and theophylline concentration (FIG. 4).

[0120] The constructs with promoters P.sub.tacI () and P.sub.tet (O) show near maximum growth while not being induced and maximum growth with slight induction. The P.sub.lacUVS () construct shows the highest theophylline dependency having no growth without induction and a dynamic range of 0 mM-0.4 mM theophylline resulting in growth rate 0 h.sup.1-0.69 h.sup.1. The promoters P.sub.ara, P.sub.bla, P.sub.cat and P.sub.lac do not support log phase growth nor does the negative control (frame shifted thyA). E. coli DH10B containing the frame-shifted thyA serves as positive control (closed square) (FIG. 4).

TABLE-US-00001 TABLE1 Promotersequencesusedintheexpressionof reportercontructsinE.coliSequence Sequence Code Name -35box-10box Consensusseq TTGACANNNNNNNNNNNNNNNNN-TATAAT SC019a P_ara CTGACGCTTTTTATCGCAACTC--TCTACT SC019b P_bla TTCAAATATGTATCCGCTCATGA-GACAAT 5C019c P_cat ATGAAATAAGATCACTACCGGGCGTATTTT SC019d P_lac TTTACACTTTATGCTTCCGGCTCGTATGTT SC019e P_lacUV5 TTTACACTTTATGCTTCCGGCTCGTATAAT SC019f P_tacI TTGACAATTAATCATCGGCTCG--TATAAT SC019g P_tet TTGACAGCTTATCATCGATAAGC-TTTAAT

[0121] In all cases a higher thyA expression results in more growth. More growth does not necessarily mean that the ThyA expression is high enough to sustain the growth. No growth above OD.sub.600 of 0.040 was observed for bacteria carrying thyA under control of the P.sub.ara, P.sub.bla, P.sub.cat and P.sub.lac promoters. These promoters do not support log phase growth, but can extend the period the bacteria can grow on the carry-over thymidine depending on the promoter strength and induction by theophylline.

[0122] Theophylline dependent log phase growth was observed for P.sub.lacUVS, P.sub.tacI and P.sub.tet. These promoters more closely resemble the consensus sequence of the 35 and 10 regions. To observe better growth when the auxotrophy complementation is under control of a stronger promoter is to be expected. The promoters P.sub.tacI and P.sub.tet are strong enough to support log phase growth without induction by theophylline, while the P.sub.lacUVS promoter does not. The positive control, E. coli DH10B with the frame-shifted thyA gene, appears to be slightly theophylline dependent. This effect is marginal and is most likely caused by the position the samples have on the microtiter plate rather than the theophylline concentration (FIG. 4).

[0123] Auxotrophy complementation indirectly depends on the concentration of mature mRNA. The concentration of mature mRNA depends on the concentration of immature mRNA and the maturation rate. The concentration of immature mRNA is mostly dependent on promoter strength, while the maturation rate is dependent on theophylline induction. The maturation rate does not equal zero when the no inducer is present. This leakage is shown by the constructs having a strong promoter in front of the coding sequence. Where there no leakage, promoter strength should have no effect when no inducer is present. The weak promoter constructs do not generate enough mature mRNA even when the maturation rate is high. The amount of ThyA is not enough to reach the minimal concentration of dTTP required in the cell. A concentration below the minimal requirement will result in thymineless death. It appears there is a fine line between never enough ThyA and always enough ThyA. The balance is matched rather well with the P.sub.lacUV promoter. No growth is observed when no inducer is present and maximum growth is observed at full induction (FIG. 4).

[0124] Although the growth of E. coli DH10B-thyA carrying the P.sub.lacUVS construct was not observed in microtiter plates, it sometimes was observed in 5 mL cultures in a 50 mL Greiner tube. Evaporation is a serious issue in the microtiter plate only causing problems after several hours of growth. By that time all exponential growth was finished already and carry-over thymidine was consumed staggering the growth. The bacteria in the Greiner tube did not suffer from evaporation, so a very small subpopulation having slightly increased expression may become dominant overnight indicating that the background expression of ThyA is only just below the minimal requirement, sometimes exceeding it. While this background growth may not be interfering with competition experiments between induced and uninduced bacteria, it may lead to false negatives.

EXAMPLE 2

Introduction of a Second Intron Into the thyA Coding Sequence

[0125] The strong promoters P.sub.tacI and P.sub.tet showed leakage exceeding the minimal requirement of ThyA (FIG. 4). A second intron was introduced into the coding region of thyA. No other part of the thyA gene matches the native flanking regions of the phage T4 td intron, so another strategy was applied. The absence of an easily identified insertion site presents a challenge as the intron ribozyme is composed not only of the intron region itself, but of the 3 flank of exon 1 and the 5 flank of exon 2 as well. The intron flanks are therefore part of both the ribozyme and the coding region (FIG. 5).

[0126] Previously, Pichler et al. (Pichler and Schroeder, 2002, J Biol Chem 277:17987-17993) showed that the flanking sequence does not need to match the native flanking sequence perfectly for a functional phage T4 td intron. Next to some tolerance in the intron flanking regions, the coding sequence can be composed of different codons. An algorithm was written to analyse the thyA coding sequence for possible locations for the intron to be inserted. The position had to match several requirements: 1) Mutations in the coding sequence were to be translationally silent for both the intron flanking regions and the restriction sites to clone the second intron into the thyA gene; 2) The flanking regions of the second intron had to match the flanking regions of the native intron as closely as possible; 3) No mutations in the flanking regions were allowed other than described by Pichler et al. (Pichler and Schroeder, 2002, J Biol Chem 277:17987-17993); and 4) Possibility of silent introduction a restriction site next to either flank was preferential.

[0127] The top candidate position was identified as HLRSI (amino acids 51-55) with the intron in frame 2 and only one mutation in the intron flanking region changing a wobble base pair into a U-A base pair. Two unique restriction sites could be mutated close to the insertion site: Psp1406I upstream and PstI downstream. A construct with a tandem intron at (H51-I55) and (F171-P175) and a construct with the (H51-I55) intron only were made. In both cases, the thyA gene was under control of the P.sub.tacI promoter. The constructs were tested in E. coli DH10B-thyA according to the same protocol as the single intron constructs (FIG. 6).

[0128] The growth rate of both single and double intron constructs under control of the P.sub.tacI promoter was measured (FIG. 8). Intron 1 () is the intron at (P171-P175), intron 2 () is the intron at (H51-I55). Introns 1 and 2 in tandem (O) shows a theophylline dependency with a dynamic range of 0 mM-0.4 mM theophylline resulting in a growth rate of 0 h.sup.1-0.67 h.sup.1. Less leakage is observed with intron 2, but less maximum growth rate as well, indicating a lower splice rate. No intron in the thyA gene () results in maximum growth regardless of the theophylline concentration.

[0129] A reduction of background was discovered with the second intron (FIG. 6). The tandem intron completely erases the background growth even with the P.sub.tacI promoter. Aside from the absence of background growth on microtiter plate, no bacterial growth was observed lacking both thymidine and theophylline. The second intron at (H51-I55) alone shows a slightly lower ThyA expression compared to the first intron at (F171-P175). This result may be caused by the difference in position, the difference in sequence or both. The difference in position means that the surrounding parts of the mRNA will have different secondary and tertiary structures as well as a different translation speed. This may affect the splicing rate of the intron. The difference in sequence will affect the ThyA production in two ways. Firstly, the intron splice rate is directly dependent on the intron flanking region (Pichler and Schroeder, 2002, J Biol Chem 277:17987-17993) which is not the same for intron 1 and 2. Furthermore, by introducing the silent mutations for the restriction sites and the intron flanking region, the amino acid sequence may not be altered, but the codon usage is. The difference in codon usage may influence the ThyA expression either in a positive or negative way.

[0130] The phage T4 td intron is therefore a useful tool for selection of E. coli that have a small molecule inside their plasma membrane. The ability of this system to completely select against bacteria that have no such small molecule present makes it relatively straightforward to select for the bacteria that do. Leakage and fully-induced expression can be carefully adjusted so bacteria without small molecule do not grow at all, whilst the bacteria with small molecule do. It was shown that the P.sub.lacUVS promoter can balance the leakage and the induced expression so that the dynamic range is between 0 mM and 0.4 mM theophylline resulting in a growth rate between 0 h.sup.1 and 0.69 h.sup.1 on microtiter plate. However, this particular promoter is not expected to support Log phase growth and is net so preferred. A direct route to manage balance between leakage and full expression is the introduction of a second intron at an upstream position. Tandem introns are significantly more effective in reducing background splicing, while maintaining the dynamic range in both inducer concentration and growth rate.

[0131] Materials and Methods

[0132] Chemicals and Plasmids

[0133] Thymidine and theophylline were purchased from Sigma-Aldrich (St. Louis, Mo.). A plasmid containing the E. coli thyA gene interrupted by a modified phage T4 td intron between G173 and L174 was commissioned at GeneArt (pMA-ThyA-SI001) as well as an intron version containing a theophylline responsive aptamer (pMA-ThyA-Theo). Plasmid pET24d was purchased from Novagen. Plasmid pRham C-His was purchased from Lucigen.

[0134] Enzymes were purchased from Thermo Scientific and used according to the manufacturer's instructions, unless stated otherwise.

[0135] Bacterial Strains and Media

[0136] E. coli DH10B T1.sup.R was purchased from invitrogen (C6400-03) and used for plasmid propagation and standard molecular techniques, as well as a parent strain for the thyA deficient E. coli DH10B-thyA strain. Transformation was performed with a ECM 63 electroporator (BTX) at 2500 V, 200 and 25 F, 2 mm cuvettes, 20-40 L of electro-competent cells and recovery in LB.

[0137] Bacteria were generally grown at 37 C. on LB medium (Miller) containing the appropriate antibiotics: kanamycin (50 mg/L), ampicillin (100 mg/L), chloramphenicol (35 mg/L) and tetracycline (15 mg/L). In addition, the auxotrophic E. coli DH10B-thyA was complemented with thymidine (100 mg/L) when necessary.

[0138] Construction of Reporter Plasmids

[0139] The reporter plasmids pSC018a-gTheo were constructed using pACYC184 as a base. The steps include exchange of the chloramphenicol acetyltransferase (cat) for the aminoglycoside 3-phosphotransferase (kan) from pET24d (Novagen), exchanging the TetA(C) for the thyA gene encoded on the pMA-ThyA-SI001 plasmid and exchanging the 6b hairpin for the theophylline responsive aptamer from pMA-ThyA-Theo).

[0140] Promoter variants were made by polymerase chain reaction (PCR) and ligating the PCR product into pSC018f-Theo (FIG. 7A) between the KpnI and BcuI sites. pSC022f-Theo (FIG. 7C) was constructed by cloning a second theophylline responsive intron into pSC018f-Theo between R53 and S54 using the Psp1406I and PstI sites. The second intron was generated by PCR using pMA-ThyA-Theo as a template. pSC024f and pSC026f-Theo (FIG. 7B) were constructed by using pSC018f-Theo and pSC022f-Theo respectively as template for PCR. Ligation of the PCR products into pSC018f between the Acc65I and MluI site removed the intron between G173 and L174, leaving no intron in pSC024f and one intron between R53 and S54 in pSC026f-Theo. A frame-shift construct in the thyA gene of pSC024f was constructed by digestion with MluI, Klenow fragment 5 fill-in and re-ligation of the plasmid.

[0141] DNA purification was performed with the DNA Clean & Concentrator-5 kit of Zymo Research (D4004) or the Zymoclean Gel DNA Recovery Kit (D4002). Plasmid was isolated with the Plasmid Miniprep kit of Thermo Scientific (#K0503). Ligation was performed at 22 C. for 1 h, followed by 10 min heat inactivation. All plasmids were verified by PCR and/or restriction analysis and sequencing by GATC Biotech (Konstanz, Germany).

[0142] Construction of the Thymidine Synthase Deficient Strain

[0143] The thyA deficient strain DH10B-thyA was made according to a standard protocol (Datsenko and Wanner (2000) PNAS 97: 6640-6645) with the exception of the PCR template and the competent cells protocol and the PCR template for the insertion cassette.

[0144] Electro-competent cells were made by growing DH10B T1.sup.R (Invitrogen) containing pKD46 at 30 C. on 16 g/L peptone, 10 g/L yeast extract and appropriate antibiotic to an OD.sub.600 of 0.4 and cooled down to 4 C., washed with ultrapure water once and 10% glycerol twice. Finally the bacteria were concentrated 250 in 10% glycerol.

[0145] DH10B T1.sup.R containing pKD46 was transformed with a PCR product generated from pMA-RQ-Lox71-kan-Lox66, kindly provided by Teunke van Rossum, containing a kanamycin resistance gene flanked by Lox71 and Lox66. The Lox sites can be recombined by cre recombinase removing kanamycin resistance, but do not form a functional Lox site. Transformed bacteria were recovered in LB medium containing thymidine (100 mg/L) for 2.5 h at 37 C. and plated on LB agar plates containing kanamycin (50 mg/L) and thymidine (20 mg/L). Colonies were verified for thyA deficiency by plating on LB agar plates containing kanamycin (50 mg/L). Plasmid curation was assessed by growing on LB agar plates containing ampicillin (100 mg/L) and thymidine (20 g/L).

[0146] Electro-competent cells were made from DH10B T1.sup.R-thyA-kan growing on medium containing kanamycin (50 mg/L) and thymidine (100 mg/L) at 37 C. and transformed with pJW168 containing the cre recombinase. Auxotrophy, recombination of the Lox sites and plasmid curation were assessed by plating on LB agar medium, LB agar containing kanamycin (50 mg/L) and thymidine (20 mg/L) and plating on LB agar medium containing ampicillin (50 mg/L) and thymidine (20 mg/L). Electro-competent cells were made of the knock-out strain and transformed with the auxotrophy reporter constructs.

[0147] E. coli DH10B-thyA Growth Assays

[0148] E. coli DH10B-thyA containing a reporter construct of the pSC series were grown overnight at 37 C. on LB medium containing kanamycin (50 mg/L) and thymidine (100 mg/L). A 10.sup.4 dilution was made and grown in with a variable amount of theophylline in a 96 well microtiter plate (Greiner) in a final volume of 200 L. Culture plates were incubated under continuous shaking for 20 h at 37 C. and the OD.sub.600 was measured every 10 minutes in a Synergy MX plate reader. As carry-over thymidine allows the knock-out strain to grow without ThyA, a lower limit OD.sub.600 was set to 0.040 AU to negate false positive growth. Growth rate () was calculated from at least 1 h of log phase growth exceeding an OD.sub.600 of 0.040 according to

In(C)=In(C.sub.0).Math..Math.t

EXAMPLE 3

Development of a Synthetic Riboswitch-Ribozyme Hybrid

[0149] Group I self-splicing introns are RNA molecules with catalytic activity: i.e. RNA ribozymes. These introns catalyze their own excision from precursors such as mRNA. The well characterized T4 self-splicing intron has been demonstrated to adopt a specific 3D-structure that is required for catalytic activity (FIG. 8). In loop 6a, sequences have been inserted that affect the splicing activity. Insertion of ligand-binding RNA fragments (aptamers) controls the splicing by a conformational change, triggered by the presence or absence of a specific ligand.

[0150] The T4 self-splicing intron has been engineered into a functional catalytic riboswitch by inserting a theophylline-binding aptamer (FIG. 5). The recombinant riboswitch was still able to splice itself, but its cleavage activity was triggered by a conformational change upon binding of the aptamer ligand, i.e. theophylline (Thompson et al, 2002 BMC Biotechnol. 2: 21). The molecular mechanism of the ligand-dependent riboswitch activity is the reversible disruption of the RNA structure (FIG. 5). Replacing the stem-loop by an aptamer destabilizes its formation, but the stem-loop can be restored by a ligand binding to the aptamer. However, the formation of the stem-loop can occur without binding of the ligand. The stem-loop without the bound ligand is less stable than the ligand bound stem-loop, but can cause background provided the screening or selection method is sensitive enough.

[0151] To test for riboswitch functionality, the intron/aptamer fusion was integrated in a reporter gene. Next to the aforementioned antibiotic resistance marker, also auxotrophic markers can be used (essential genes for amino acids or nucleotides are deleted in microbial hosts; growth in the absence of these amino acids or nucleotides is only possible when the corresponding gene is complemented in a plasmid). For the riboswitch test, an E. coli thyA knockout strain; the thyA gene encodes an essential enzyme in biosynthesis of thymidine (one of the four bases of DNA nucleotides). The thymidine auxotrophy is complemented by a plasmid-borne thyA gene. When using a plasmid with ThyA that was interrupted by the hybrid riboswitch, it was demonstrated that the thyA knockout strain of E. coli could survive on minimal medium containing theophylline (without thymidine), but not on minimal medium lacking both theophylline and thymidine (Thompson at al. 2002 BMC Biotechnol. 2: 21). Hence, the presence of the riboswitch ligand theophylline, allows for growth.

EXAMPLE 4

Tandem Riboswitch Provides Improved Stringency

[0152] In the aforementioned experiments (Thompson et at 2002 BMC Biotechnol, 2: 21), there was a background level cell growth was observed in the absence of the theophylline ligand (Figure ). This background was subtracted from the growth observed in the presence of the ligand. Theophylline-induced growth was approximately 40% of the parental intron, so maximally 40% of the wild type growth. In an experiment based on the latter publication, the induction of the thyA gene by theophylline turned out to be far from stringent. Non-induced splicing of the intron was thought to be the cause of growth on minimal medium lacking both theophylline and thymidine. As a solution to the problem of non-induced splicing, a suitable second insertion site was identified, for the insertion of a second self-splicing intron.

[0153] A comparison of a single intron construct with a construct of the present invention featuring a tandem self-splicing intron was made (FIG. 9). Growth (OD600) was plotted against time (minutes) in the presence (1 mM) and absence (0 mM) of theophylline for the construct designed by Thompson et al. (Thompson et al. 2002 BMC Biotechnol. 2: 21) utilising one theophylline-dependent intron in thyA, in a thyA-deficient E. coli DH10B strain (FIG. 9A). On this time scale the theophylline-induced bacteria outcompeted the non-induced bacteria with ease, but the non-induced bacteria continued growth until the OD600 reached a value comparable to the induced bacteria overnight (not shown). Growth was measured in the same way for a construct containing an E. coli thyA gene interrupted by two theophylline-dependent introns, which was transformed into a thyA-deficient E. coli DH10B strain. The overnight growth was plotted against the theophylline concentration in the medium (FIG. 9B). The growth shows a non-linear relationship with theophylline concentration. The line drawn through the first five points represents a Michaelis-Menten-like kinetics profile with an apparent kd value of 0.1-0.2 mM theophylline (FIG. 9B). Surprisingly, the absence of growth in the absence of theophylline indicates a selection method with no detectable background (FIG. 9B) in contrast with the prior art construct (FIG. 9A) where a low level of cell growth continues in the absence of theophylline). Above 1 mM, theophylline became slightly toxic.

[0154] The introduction of a tandem-riboswitch resulted in a 70% recovery of growth compared to wild type with theophylline induction, while no growth at all was observed without induction by theophylline, i.e. black and white selection.

[0155] Each mRNA only possesses a single 5 UTR and this limits the potential positions for riboswitches to be inserted. It has been shown that the introduction of more than one riboswitch can confer improved stringency on the activity of the riboswitch, by removing background levels of non-induced splicing (and therefore expression). As there is only one 5 UTR and multiple positions for an intron, instead of inserting the riboswitch in the 5 UTR, use of the coding sequence allows multiple riboswitches to be inserted into the coding sequence and therefore improved flexibility and stringency of the engineered switch. The presence of a specific ligand (e.g. theophylline) controls self-splicing of the riboswitch, thereby restoring the reading frame of the gene encoding the desired gene product.

EXAMPLE 5

Auxotrophy Complementation

[0156] Auxotrophy complementation is based on interruption of an important step in the pyrimidine synthesis pathway (FIG. 3). The thyA gene is knocked out in the host strain E. coli K12 substrain DH10B and complemented by a either a plasmid encoded thyA copy or thymidine supplement in the growth medium.

[0157] The plasmid encoded thyA gene is interrupted by a theophylline responsive self-splicing intron; single or in tandem. The vector maps are depleted in FIG. 7. The pSC018-Theo (FIG. 7A) contains one intron in the coding sequence and does slow down the growth significantly when not induced, however during prolonged incubation (overnight) the non-induced bacteria grow to the a similar density as the induced bacteria. While induction does give a growth advantage, the selection is not black and white. A single intron insertion on another position (FIG. 7B) yields a similar picture as an intron inserted on the wild type position. Only two introns in tandem (FIG. 7C) provide enough control to have no growth at all while the bacteria are not induced and a dose dependent growth when they are.

EXAMPLE 6

Intron Controlled Expression Cascade

[0158] Being an enzyme, a single molecule of ThyA can convert a vast amount of dUMP to dTMP, thereby enhancing the signal. For other reporters, like GFPuv, this is not the case as the bacteria are only as fluorescent as there are GFPuv molecules. Signal below the detection limit is a problem in case of GFPuv as is illustrated in Table 2. The fluorescence of the bacteria themselves (pSC012) is in the same order of magnitude as the induction independent self-splicing intron (pSC034f-SI001). The induction by theophylline is not significantly observed in pSC034f-Theo. Possibly there is a difference in expression between the induction independent self-splicing intron and the non-induced and induced theophylline dependent intron, but it cannot be concluded from these data. When no intron is present in the GFPuv (pSC034f) the fluorescence is multiple orders of magnitude higher than the background fluorescence.

TABLE-US-00002 TABLE 2 Direct interruption of GFPuv Induction Fluorescence ID GFPuv Intron dependent Induced (AU/OD) pSC012 23 pSC034f + 7204 pSC034f-Si001 + + 33 pSC034f-Theo + + + 16 PSC034f-Theo + + + + 25

[0159] To enhance the signal from GFPuv, a cascade was made using DNA dependent RNA polymerase from phage T7. The GFPuv expression is controlled by T7 polymerase and the T7 polymerase is in turn controlled by the theophylline responsive intron (FIG. 10). The T7 polymerase gene is carried on the bacterial genome. The enzyme's functionality depends on maturation, which is controlled by theophylline in case of the heterologous expression system, or a compound derived from an enzymatic reaction when the system is utilised as a screening method.

[0160] A few copies of the T7 polymerase will result in a myriad of GFPuv molecules, so a small change in T7 polymerase concentration will result in a large change in GFPuv concentration, which can be measured. Since the T7 polymerase is a very processive enzyme, it needs a tight control of expression. FIG. 11 shows the polymerase is controlled on the transcription level (L-rhamnose dependent promoter) and translation level (theophylline dependent intron). Read-through GFPuv expression is reduced by an upstream terminator.

[0161] The reporter plasmid pSC028-GFPuv-term was constructed using pACYC184 as a base (vector map shown in FIG. 12). The cat gene conferring chloramphenicol resistance was exchanged with the kanamycin resistance gene from pET24d. The tetA(C) gene was replaced causing an insertion site flanked by Acc65I and BcuI. The pGFPuv NdeI.sup. XhoI.sup. was used as template for a polymerase chain reaction (PCR) adding an NdeI site to the 5-end of the GFPuv gene and a BcuI site to the 3-end. The PCR fragment was ligated to pRham-CHis (Lucigen) digested with NdeI. This ligation added the 5-UTR to the GFPuv gene. A PCR was performed on the ligation reaction adding an Acc65I site, a terminator and PT7 promoter in front of the 5-UTR and a BcuI site to the 3-end of the GFPuv gene. The secondary PCR fragment was digested with Acc65I and BcuI and ligated into the Acc65I/BcuI insertion site. Finally, the T7 terminator from pET24d was amplified by PCR adding a BcuI site to the 5-end and an XbaI site on the 3-end. The terminator was ligated into the BcuI site 3 of the GFPuv gene, causing the BcuI site to persist on the 5-end of the terminator and to be removed on the 3-end of the terminator.

[0162] The performance of the cascade was measured using GFPuv expression (FIG. 12). Specifically, the theophylline dependency of the cascade at 0.8 mg/L rhamnose was measured. The PtacI serves as benchmark for a strong promoter dependent on E. coli RNA polymerase and is the pSC034f construct. Additionally, expression of GFPuv over a wide range of rhamnose and theophylline concentrations was determined. The cascade is controlled by one theophylline dependent intron at the first position depicted in FIG. 12. The system is virtually off when either rhamnose or theophylline is absent, although the absence of L-rhamnose is more important than the absence of theophylline. The background observed when fully induced with rhamnose is about 7.5% of the maximum expression. Also there is a correlation between the background expression in the absence of one inducer and the expression level upon induction. There is a good dose-dependent relationship and the fully induced system yields a signal 5-6 times the signal observed from the strong TacI promoter. A small amount of theophylline causes a measurable signal already, which can be distinguished by techniques like FACS to separate the bacteria producing the enzyme of interest from the rest.

[0163] It is unlikely that an enzyme of interest will provide a concentration as high as 1 mM for every small molecule that is screened. As the enzyme of interest functions inside the cell, the cell membranes acting as a harder will help the production of GFPuv rather than diminishing it. Usually aptamers have a dissociation constant in the low M range, so for maximum signal the intracellular theophylline concentration does not have to be 1 mM, but much less.

EXAMPLE 8

Amplification of the Signal Using a T7 GFPuv Cascade

[0164] Construction of Reporter Plasmids

[0165] The reporter plasmid pSC028-GFPuv-term (FIG. 12) was constructed using pACYC184 as a base. The cat gene conferring chloramphenicol resistance was exchanged with the kanamycin resistance gene from pET24d. The tetA(C) gene was replaced causing an insertion site flanked by Acc65I and BcuI. The pGFPuv NdeI.sup. XhoI.sup. was used as template for a polymerase chain reaction (PCR) adding an NdeI site to the 5-end of the GFPuv gene and a BcuI site to the 3-end. The PCR fragment was ligated to pRham-CHis (Lucigen) digested with NdeI. This ligation adds the 5-UTR to the GFPuv gene. A PCR was performed on the ligation reaction adding an Acc65I site, a terminator and P.sub.T7 promoter in front of the 5-UTR and a BcuI site to the 3-end of the GFPuv gene. The secondary PCR fragment was digested with Acc65I and BcuI and ligated into the Acc65I/BcuI insertion site. Finally, the T7 terminator from pET24d was amplified by PCR adding a BcuI site to the 5-end and an XbaI site on the 3-end. The terminator was ligated into the BcuI site 3 of the GFPuv gene, causing the BcuI site to persist on the 5-end of the terminator and to be removed on the 3-end of the terminator.

[0166] The plasmid pRham-CHis (Lucigen) was used as base for constructing the T7 polymerase variants. The CDS is flanked by an NdeI site and 6His tag on the 5-end and a BgIII site on the 3-end. The intron positions are between G201 and L202 flanked by PscI and HindIII, between G449 and L450 flanked by Bsu15I and XagI and between G671 and L672 flanked by Eco88I and PstI. All have CAAGGGT as 5 intron flank instead of wild type CTTGGGT. The 3 intron flanks are CTAC, CTAC and CTAA respectively.

[0167] GFPuv Fluorescence

[0168] E. coli DH10B-T7His-Theo4 was grown overnight at 37 C. in LB medium containing kanamycin (50 mg/L). A 96 well 2 ml culture plate (Greiner) was filled with a concentrate of theophylline and L-rhamnose. LB medium containing kanamycin and overnight grown bacteria were added so that the final concentration of kanamycin was 50 mg/L, the bacteria had a final dilution of 10.sup.3 and the theophylline and L-rhamnose were diluted to 1 in 500 L total volume. Culture plates were incubated at 37 C. overnight under continuous shaking. The bacteria were centrifuged for 10 minutes at 4700 rpm in a Sorval Legend centrifuge. The supernatant was cleared and the cell pellet was resuspended in 500 L 50 mM Tris-HCl pH 7.5. After resuspension, the plates were incubated at 37 C. for 1 hour to allow maturation of the GFPuv. 100 L of suspension was pipetted into a 98 well black plate with clear bottom (Perkin Elmer) and measured with a Synergy MX plate reader. The cell density was measured by scattering at 600 nm and the fluorescence was measured at an excitation wavelength of 385 nm with a width of 20 nm and an emission wavelength of 508 nm with 20 nm width with a gain of 50. The background fluorescence and background scattering were subtracted and the fluorescence was divided by the scattering at 600 nm. The background fluorescence of bacteria without either GFPuv or T7His polymerase was negligible, but the fluorescence caused by other components than GFPuv in the bacteria was still subtracted.

[0169] Results

[0170] The theophylline dependency of the cascade in response to differing concentrations of rhamnose was measured. E. coli DH10B-T7His-Theo4 diluted from an overnight culture were grown overnight in a 2 mL culture plate containing a variable amount of L-rhamnose and theophylline. The medium was cleared and the bacteria were resuspended in 50 mM Tris-HCl pH 7.5. The fluorescence was measured at an excitation wavelength of 385 nm and an emission wavelength of 508 nm. The cell density was measured by scattering at 600 nm.

[0171] GFPuv fluorescence showed a strong dependency on both L-rhamnose and theophylline (FIG. 14). The fluorescence observed without any induction at all does not significantly differ from the fluorescence observed for the GFPuv reporter plasmid alone. This implies that the double control on the T7 polymerasetranscription control by L-rhamnose and translation control by theophyllinesucceeds to a large extent in keeping the T7 polymerase inactive. The transcription of T7 polymerase itself dictates the dynamic range in fluorescence caused by theophylline. This dynamic range can be adjusted according to the requirements of the application. At any L-rhamnose concentration, the fold-change caused by theophylline is around 15 times. The cascade being a multi-component system with two types of control makes it very hard to model the dependency on both L-rhamnose and theophylline and to have reproducible results. For a final application as a biosensor, the transcription of T7His polymerase may be fixed with a constitutive promoter, while the functional translation will be ligand dependent.

[0172] Since the output dynamic range can be adjusted relatively easily, the intron controlled T7His polymerase can be employed as a generic tool. The intron lowers the maximum translation quite severely, so not all reporter genes will show enough signal when put under control of a ligand dependent intron directly. Enzymes like ThyA or LacZ can handle the lower translations efficiency, but all genes that need an at least decent expression to function, like GFPuv, can now be put under control of one enzyme. An additional advantage is the exchangeability of the reporter plasmids. Expression from these reporter plasmids can be easily adjusted by mutating the T7 promoter.