NIH Individual Postdoctoral Application
I-tsuen Chen
(1) Specific Aims2) BACKGROUND Since the unexpected discoevry of 'split' genes in adenovirus was discovered1much progress has been made in understanding of the mechanisms of eukaryotic pre-mRNA splicing (3,19. However, the exact requirements for splicing and the rules governing the splite site selection remain unclear (for review, see 1). Much effort to define the splice sites in eukaryotic transcripts are based on the consensus sequence and by site-directed mutagenesis studies. The assays most often used to analyze the effects of splice site mutations have been splicing in cell-free nuclear extracts (in vitro) and splicing during transient expression following transfection (in vivo). Some confusing results have been obtained depending on which of these two assays or which gene systems (normal vs. alternate splicing) have been used (2,8,10). One example of a discrepancy between the in vivo and in vitro systems is the splicing of kappa immunoglobin light chain pre-mRNA. This gene is functionally transcribed only after rearrangement of one of the variable gene segments to one of four junctional (J) regions. In contrast to the in vivo situation, in which the accurate 5' splice site of the recombined J segment was used, there is no discriminating among the J segment splice sites in vitro. Splicing of transcripts of integrated genes adds another condition that can yield different results (1,9). Up to now, relatively little information has been provided by the analysis of splicing mutants of integrated genes. Another approach of studying splicing is to isolate spontaneous mutants of an endogenous gene within the context of its indigenous chromosomal location. Very informative splicing mutants causing exon-skipping were isolated by this method (14). However, this approach is more difficult than the site-directed mutagenesis of cloned genes, and the analysis of large numbers of mutants is needed in order to obtain additional splicing mutants.1. Insertion of a promoterless gene that can be selected for within the intron of another gene that can be selected against.
2. Saturation mutagenesis for all cis~acting mutations affecting splicing of the transcript of a chromosomal gene.
3. Characterization of the isolated splicing mutants.
4. Selection for suppressor mutants.
It has been shown in yeast systems that a genetic approach to the splicing problem is very informative. That is, to construct a gene fusion such that production of an easily scored gene product is dependent on proper splicing. Such a gene fusion includes the yeast actin gene which contains one intron followed by the yeast HIS 4 gene downstream. The expression of the HIS 4 gene product requires the proper splicing of the actin. By in vitro mutagenesis, a unique point mutation that alters the splicing of this gene fusion was isolated (20).
If a We propose to select for all mutations in an in situ gene containing a single intron that can disrupt tbe spiicing process. The strategy is to place a selectable gene into an intron of another gene that can be selected against. The resulting construct will then be introduced into appropriate mammalian cells. If the intron is spliced, the selectable gene gets destroyed and is not expressed (Fig.lA). However, if a mutation prevents efficient splicing, then if the unspliced larger mRNA were sufficiently stable, the internal gene could be translated from its resident AUG (Fig lB). Therefore, the splicing mutants could be directly isolated by selecting for the internal gene and against the host gene. The nature of mutations will be determined by polymerase chain reaction (PCR)-amplification and direct sequencing (7,21,22) and the mutant phenotype analyzed by RNase protection methods (17,32). This saturation mutagenesis would define all bases necessary for correct splicing of an integrated gene.(3) EXPERIMENTAL DESIGN AND METHOD
1. Cell Line The parental cell line DG44 will be used for this study. DG44 is a derivative of Chinese hamster ovary (CHO) cells that is a double deletion for the dihydrofolate reductase (DHFR) gene (28). DHFR-deficient cell lines require glycine, a purine, (e.g., hypoxanthine), and thymidine for growth because these cells are unable to reduce folate supplied in the medium to the active form of cofactor, tetrahydrofolate. Such mutants can be readily selected from hemizygous cells on the basis of their relative resistance to tritiated deoxyuridine suicide (27 ). In this selection, DHFR-positive cells are killed by the incorporation of tritiated deoxyuridine into their DNA. We will use this method to select against the DHFR gene.
Plasmid Constructs and Transfection. We will place the prokaryotic gene for hygromycin resistance into the first and only intron of a DHFR minigene, pDCHIP (see Fig 1). pDCHIP contains the six exons of DHFR gene, intron 1, O.8 kb of the 5' flank and the first major polyadenylated site. This minigene transfects DHFR-deficient CHO cells with high efficiency, about the same efficiency as the full genomic sequences (29 ). The 2 kb Sma I fragment containing the HPH gene (an E. coli gene encoding hygromycin B phosphotransferase and conferring hygromycin resistance) can be isolated from the plasmid pHEBO (available from B. Sugden). This Sma I fragment contains the coding region of the HPH gene followed by polyadenylation signals of the HSV tk gene (thymidine kinase gene of herpes simplex virus type I); no promoter and intron are present in this fragment (26). The Sma I fragment will be inserted into a unique Pst I site at the middle of the 0.3 kb intron of pDCHIP. Briefly, the Pst I site will be treated with T4 polymerase to create blunt ends or exonuclease Bal 31 to trim l0-20 bp around Pst I site to remove a nearby potential AUG site. The resulting plasmid construct will be linearized and transfected into DG44 cells by the calcium-phosphate or Polybrene method (4,31). The DHFR-positiye transfectants will be selected in medium lacking hypoxanthine and one having a single copy of the construct would be chosen. The linearized construct should increase our chances to get a single copy insert in the 'transfected cell (unpublished result).
We expect that failure to remove the intron would produce a DHFR-deficient phenotype. However we don't know that failure to remove the intron will lead to HPH gene expression through the DHFR promoter. For instance, this scheme depends on the ability of the HPH coding sequences to be translated starting from a downstream AUG. Although such secondary starts have been documented previously (12,23), they may be inefficient. As an important control before the experiment is initiated, we will not insert HPH gene into the wild-type minigene but rather into a mutant minigene that can't splice. Such a mutant minigene was produced by site-directed mutagenesis in which the onserved GT (15) at the 5' end of the intron 1 has been mutated to GC (obtained from A. Weiner). We found that when this mutated minigene was introduced into DG44 cells they accumulated unspliced transcripts and remained DHFR-negative, even if the construct was present in multiple copies (unpublished results). Therefore, we will first place the HPH gene into the Pst I site of the mutated minigene to see whether the HPH gene can be expressed and confer resistance to hygromycin in transfected cells. If the DHFR promoter is unable to activate the HPH gene, we will replace it with stronger promoter such as the SV4O early region promoter. If the upstream AUG near the Pst I site causes a problem (the HPH gene may use it as start site, which will not be in-frame), we will remove it. If HPH will not be translated efficiently using its own AUG, we could try putting HPH in-frame with the DHFR start. If we succeed with the mutated minigene, we will proceed to place HPH gene into the wild type minigene and select mutants. If the HPH gene itself interferes with splicing, i.e. it contains cryptic splice sites (30), then no DHFR-positive starting cell line could be produced; we would then try other selectable gene is such as neomycin (NEO) or GPT (Ecogpt for xanthine-guanine phosphoribosyltransferase) (16,25). A more general approach would be to select mutants carrying chimeric genes that have become DHFR-positive because the cryptic splice site has been mutated away.
3. Mutagenesis and Isolation of Mutants A culture of cells containing a sngle copy of an HPH insert will be mutagenized (1% survival) with ethyl methane sulfonate (EMS) or other mutagens to insure vulnerability at every position. EMS is useful because it induces primarily point mutations rather than major DNA rearrangements in mammalian cells (18 ). Mutagenized cells will be allow 6-7 days of growth for phenotypic expression. They will then be subject to selection in a medium containing hygromycin for expression of the HPH gene and tritiated-deoxyuridine to select against the DHFR gene. Positive colonies will be picked and recloned.
4. Characterization of mutants. For rapid screening of mutations in a target sequence, both RNase protection (17,32) and the polymerase chain reaction (PCR) method (21,22 ) will be used. To use the PCR technique, we will synthesize oligomers that flank the region of interest. The nature of mutation in the minigene would be examined by PCR, amplifyiing two 200-300 bp DNA fragments spanning the 5' and 3' splice sites, starting directly with mutant cells or DNA isolated from mutant cells. The amplified DNAs would be used for direct sequencing (7 ). At the same time, a riboprobe spanning an exon-intron junction would be synthesized with the use of an SP6 transcription system.(13), and used to analyze RNA from the mutant cells. This probe should reveal the ratio of spliced (partial protection) to unspliced (full protection) RNA in a given preparation. The presence of the HSV-tk poly A site may provide some competition for the splicing which might allow leaky DHFR-deficent mutant to be isolated. If this poly A interferes with splicing in the wild type gene, it will be removed. If it is necessary we will clone the mutant gene and do cell-free splicing studies to determine the nature of splicing event that has been affected.If our project works, we would be generating a series of splicing-deficient mutations. These mutant should provide a complete picture about integrated DHFR minigene splicing rather than a limited collection of point mutations.
5. Selection for suppression mutants. The splicing mutants could also serve as a substrate for studying suppression of splicing mutants. Since selection for the DHFR-positive phenotype after mutagenesis of the splicing mutants is very strong, rare revertants of dhfr splicing mutants can be isolated. We would hope to get second site mutations that compensate for (suppress) the first mutation. Even more interesting would be suppression by mutated trans-acting factors (suppressor). This suppression approach has been evidenced in a yeast system (5).If, after limited trials, no internal gene can be expressed in the mutated minigene (control experimentt) then we would have nothing to select for. This failure could be due to unstable unspliced transcripts, transcripts that can't get out of nucleus, and or that can't translate efficiently or other unknown mechanisms. We will then try an alternative approach.
6. Alternative Approach We will not place a foreign gene into DHFR minigene, but create tandemly arranged exons which would disrupt the normal DHFR sequence. We would put a DHFR genomic fragment containing exon 2 and its flanking sequence into intron 1 of the minigene so that the new construct contains the complete intron 1, two exon 2 sequences and part of intron 2 and intron 1 sequences between the two exon 2 sequences (see Fig 2). This construct will be co-transfected with NEO into DG44 cells and transfectants selected by their resistance to the drug G418. We will do rapid screening of G418-resistant clones looking for the integrated dhfr gene using PCR. If the hypothetical "first come1,first served" principle in splicing is correct (1), we would expect that the results should be DHFR-deficient because the exon 2 repeat will disrupt the nomal DHFR RNA sequence (Fig 2A). In this scheme, we expect that the initial cell line is DHFR-deficient. We would then select for the DHFR-positive phenotype, produced by the failure to splice in the interfering redundant exon 2.
If the whole redundant exon 2 is treated as an intron and spliced out, a DHFR-positive phenotype would occur from the start (Fig 2B). To test this possibility we transfected the construct into DG44 cells using the pDCH1P wild type minigene as a control and grew in the selective media. If the transfection frequencies are much lower in the construct-transfected cells, we will proceed with the experiments with confidence. We found that there was a hundred-fold transfection frequency decrease in the experimental transfected cells compared to the control pDCH1P transfected cells (unpublished result). Another experiment we could do is introducing a splicing mutation into the genomic fragment of the minigene. Its transfection frequencies as compared to the wild type construct in the selective medium should be much higher.
Specifically, a 0.8 kb Pst I to Bst ElI fragment from genomic clone PMH8 (29 ) was ligated into the Pst I site of minigene pDCH1P. After one end ligation, four deoxynucleotides were added along with Klenow fragment polymerase to create blunt ends which were ligated with each other. The construct (15-20 ng) was co-transfected with NEO DNA in a 1:2 ratio. This strategy increases the chances of finding both genes integrated into chromosomes of the same cell so that fewer G4l8-resistant colonies need to be screened. We found that when 0.5-1 microgram of a DHFR cosmid clone is transfected into a million DG44 cells, there is good chance to get single copy insertion (unpublished results). Using PCR with unique intronic primers, cells containing the DHF'R construct can be screened rapidly. Cells with different levels of DHFR activity can be distinguished by growth in medium lacking hypoxanthine, glycine and thymidine (-GHT) or lacking hypoxanthine only, which is less stringent. We and others have previously shown that about 1% of wild type levels of DHFR will suffice for growth in a purine-free medium (6, 29). We also can screen for single copy insertion by Southern blot analysis of genomic DNA (24). The mutagenesis selection of DHTR positives will proceed as soon as the correct cell line is characterized (single-copy, intact construct, and DHFR-deficient).
If we succeed with this method, we would construct another minigene which contains intron 1 and 2. We would place the exon 2 fragment into intron 2 to do similar experiments. Since the complexity of several spliced sites are involved, we would get more information about the nature of the splicing process.
(4) LITERATURE
1.Aebi, N. and Weissmann, C. (1987). Precision and orderliness in splicing. Trends Genet. 3:102-107.
2.Aebi, N., Horning, H. and Weissmann, C. (1987). 5' cleavage site in eukaryotic pre-mRNA splicing is determined by the overall 5' splice region, not by the conserved 5' GU. Cell 50:237-246.
3.Berget, S.N., Moore, C. and Sharp, P. (1977). Spliced segments at the 5' terminus of adenovirus 2 late ~A. Proc. Natl. Acad. Sci. USA 74:3171-3175.
4.Chaney, W.G., Howard, D.R., Pollard, J.W., Sallustio, S. and Stanley, P. (1986). High frequency transfection of CHO cells using polybrene. Somatic Cell Mol. Genet. 12:237-244.
5.Couto, J.R., Tamm, J., Parker, R. and Gutherie, C. (1987). A trans-acting suppressor restores splicing of a yeast intron with a branch point mutation. Genes and Develop. 1:445-455~
6.Crouse, G.F., McEwan,,R.N. and Pearson, N.L. (1983). Expression and replication of engineered mouse dihydrofolate reductase minigenes. Mol. Cell Biol. 3:257-266.
7. Gyllenstein, U. and Ehrlich, H. (1988). Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the ~LA-DQA locus. Proc. Natl. Acad. Sci. USA 85:7652-7656.
8. Kedes, D'. and Steit:, J. (1987). Accurate 5' splice site selection in mouse kappa light chain pre-mRNAs is not cell-type specific. Proc. Natl. Acad. Sci.
9. Kedes, D. and Steitz, J. (1988). Correct in vivo splicing of the mouse immunoglobulin kappa light-chain pre-mRNA is dependent on 5' splice-site position even in the absence of transcription. Genes and Dev. 2:1448-145
10. Lang, K. and Spritz, R. (1983). RNA splice site selection: evidence for a '-3' scanning model. Sci. 220:1341-1355.
11. Lowery, D. and Van Ness, B. (1987). Comparison of in vitro and in vivo plice site selection in kappa-immunoglobulin precursor mIWA. Mol. Cell Biol.:1346-1351.
12. Hajors, J.E. and Varmus, H.Z. (1981) Nucleotide sequences at host-proviral junctions for mouse mammary tumour virus. Nature. 289:253-258.
13. Melton, D., Kneg, P., Rebagliati, H., Maniatis, T. and Green, H. (1984). .Efficient in vitro synthesis of biologically active RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 12:7035-7056.
14. Mitchell, P., Urlaub, G. and Chasin, L. (1986). Spontaneous splicing mutations at the dihydrofolate reductase locus in CHO cells. Mol. Cell Biol. 1926-1935.
15. Mount, S. H. (1982). A catalogue of splice junction sequences. Nucleic ids Res. 10:459-472.
16. Mulligan, R.C. and Berg, P. (1980). Expression of a bacteria gene in mammalian cells. Sci. 209:1422-1427.
17. Myers, R.M., Larin, Z. and Maniatis, T. (1985). Detection of single 'base substitutions by ribonuclease cleavage at mismatches in RNA:DNA duplex. Sci.230: 1242-1246.
18. Nalbantoglu, J., Goncalves, D. and Meuth, H. (1985). Structure of mutant alleles at the APRT locus of CHO cells. J. Mol. Biol. 167:575-594.
19. Padgett, R.A., Grabovski,.P.j., Konarsku, H.M., Seiler, K.S. and Sharp, P.L. (1986). Splicing of messenger RNA precursors. Annu. Rev. Biochem. :1119-1150.
20. Parker, R. and Guthrie, C. (1985). A point mutation in the conserved lanucleotide at a yeast 5~:splice junction uncouple recognition, cleavage, I ligation. Cell. 41:1077118.
21. Saiki, R.K., Scharf, S., Faloona, F., Mullis, K.B., Horn, G.T., Erlich,. and Arnheim,N.A. (1985). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Sci. 230:1350-1354.
22. Saiki, R.K., Gelfand, D., Stoffel, , Scharf, S4, Higuchi, R., Horn, Mullis, and Ehrlich, LA. .(l988) Enzymatic amplification of globin genomic sequences and restriction site analysis for diagnosis of cell anemia. Sci. 239:487-491.
23. Simonsen, C.C. and Levinson, A.D. (1983). Analysis of processing and nylation signals of the Hepatitis B virus surface antigen gene by using virus 40-Hepatitis B virus chimiric plasmids. Nol. Cell Biol. 2258.
24. Southern, E.M. (1975). Detection of specific sequences among DNA fragments ed by gel electrophoresis. J. 5(01. Biol. 98:503-517.
25. Southern, P.J. and Berg, P. (1982). Transformation of mammalian cells to resistance vith a bacteria gene under control of the SV4O early promoter. J. 5(01. and Applied Genet. 1:327-341.
26. Sugden, B., Marsh, K. and Yates, J. (1985). A vector that replicated as a and can be selected in B-lymphocytes. Mol. Cell Biol. 5:410-413.
27. Urlaub, G. and Chasin, L.A. (1980). Isolation of Chinese hamster cell mutants deficient in dihydrofolate reductase activity. Proc. Natl. Acad. Sci. USA 77:4216-4220.
28. Urllaub, G., Kas, E., Carothers, A. and Chasin, L. (1983). Deletion of the dihydrofolate reductase locus from cultured mammalian cells. Cell 33: 405412.
29. Venolia, L., Urlaub, G. and Chasin, L. (1987). Polyadenylation of Chinese dihydrofolate reductase genomic genes and minigenes after gene transfer. Somat. Cell Mol. Genet. 13:491-501.
30. Wieringa, B., Meyer,F, Reiser, J. and Weissmann, C. (1983). Unusual splice revealed by mutagenic inactivation of an authentic splice site of the beta-globin gene. Nature 301:38-43.
31. Wigler, M., Pellicer, A., Silverstein, S., Urlaub, G. and Chasm, L. A. Transformation of the APRT locus in mammalian cells. Proc. Natl. Acad. A 76:1373-1376.
32. Winter, E., Yamamoto, F., Almoguera, C. and Perucho, M. (1985). A method to detect and characterize point mutations in transcribed genes: amplication and overexpression of the mutant c-Ki-ras allele in human tumor cells. Proc. Nat Acad. Sci. 82:7575-7579.