Denis Lafontaine and David Tollervey European Molecular Biology Laboratory (EMBL), Postfach 10 22 09D-69012 Heidelberg, Germany
Tel: +49 6221 387 264 Fax: +49 6221 387 518
Keywords: Saccharomyces cerevisiae, epitope tags, immunodetection, PCR
With the availability of the complete yeast genomic sequence, techniques which allow the rapid functional analysis
of genes of interest are of increasing importance. Here we report a technique which allows the initial characterisation
of genes of interest, through the construction of conditionally expressed mutations for functional analyses and the
generation of epitope-tagged fusion proteins for immuno-localisation and immuno-purification, entirely by PCR.
A PCR-based technique for the creation of chromosomal gene disruptions has been described (1). We extend this technique to allow the rapid creation of conditionally expressed alleles (GAL mutants) and the synthesis of proteins fused to epitope tags. The technique relies on the PCR amplification of HIS3-pGAL or HIS3-pGAL-TAG cassettes using two primers containing flanking sequences specific to the target gene followed by the transformation of the PCR product into a his3-strain. Four vectors have been designed and tested (Fig. 1). In these the HIS3 marker is flanked either only by the GAL10 promoter (vector pTL26) or by the GAL10 promoter fused to different epitope tags. The epitope tag sequences are 2x Protein A, 3x c-myc, and His8 (vectors pTL27, pTL28 and pTL32 respectively, see legend of Fig. 1 for full description). As an example of the use of these vectors, the construction of a GAL-regulated, ProtA::Ssb1p fusion is outlined in figure 2. The SSB1 flanking sequences present on the 5' and 3' primers target the chromosomal integration of the PCR construct upstream of, and in frame with, the initiator AUG of SSB1 (Fig. 2).
Figure 1. Structure of the pTL vectors used as PCR templates
The pTL vectors were constructed as follow. In plasmid pTL26, the GAL1-10 promoter region of plasmid pDL503 (4), isolated by EcoRI-BamHI digestion, was subcloned into plasmid pRS313 (6) in order to be fused to a HIS3 marker. The 2x ProtA cassette, which contains the two IgG binding domains of the S. aureus Protein A, was amplified by PCR from plasmid p28NZZtrc (7) using two primers which create flanking NcoI sites (primer 1: 5'-CCCATGGCAGGCCTTGCGCAACAC-3' and primer 2: 5'-CTTTCCCATGGCATTCGCGTCTACTTTCGGCGC-3'). The PCR product was digested by NcoI and subcloned into a NcoI containing vector. The 2xProtA cassette was isolated from this plasmid by NcoI digestion, filled with Klenow DNA polymerase and inserted at the filled EcoRI site of plasmid pTL26 to yield vector pTL27. The 3x c-myc cassette which contains three human c-myc epitopes bridged by glycine residues was amplified by PCR from plasmid pUC119-myc-tag3 using two primers which create flanking NcoI sites (primer 1: 5'-AATTATACCATGGGTACCCGGGGATCCTCTAGA-3' and primer 2: 5'-CTTTCCCATGGCTCTAGAGGATCCGTTCAAGTC-3'). The PCR product was digested by NcoI and subcloned into a NcoI containing vector. The 3x c-myc cassette was isolated from this plasmid by NcoI digestion, filled with the Klenow enzyme and inserted at the filled EcoRI site of plasmid pTL26 to yield vector pTL28. The His8 tag was created by annealing the following oligonucleotides: (i) EcoRI-8His-F 5'-AATTCATGAGAGGTTCTCACCATCACCATCACCATCACCATC-3'; (ii) XhoI-8His-R 5'-TCGAGATGGTGATGGTGATGGTGATGGTGAGAACCTCTCATG-3'. The linker was digested by EcoRI-XhoI and subcloned in pTL26. The resulting plasmid was called pTL32. pTL32 can be used for the in-frame fusion of 6 or 8 histidines residues. All constructions were checked by sequencing. PCR reactions were performed using 30 ng of the appropriate pTL vector and 100 pmol of each primer in buffer containing 10 mM Tris HCl (pH 8.3 at 20¡C), 50 mM KCl, 1.5 mM MgCl2, 200 μM of each dNTP, 10u Taq DNA polymerase (Boehringer) in a final volume of 100 μl. Following an initial denaturation step, 5 min, 94¡C, Taq DNA polymerase was added and amplification was performed 30 x (1 min, 45¡C; 4 min, 72¡C; 30 sec, 94¡C), followed by incubation for 10 min, 72¡C. The sequences of PCR primers used for these amplifications are given in Table 1. For each target gene the 5' primer is common for all templates. The PCR products were digested with XmnI and gel purified using a QIA quick kit (QIAGEN). The optional XmnI digestion step cuts the pTL vector template and avoids problems with contamination of the PCR product by intact vector. The lengths of the PCR products are 2584 bp (HIS3-GAL), 2973 bp (HIS3-GAL-2x ProtA), 2721 bp (HIS3-GAL-3x c-myc), 2613 bp (HIS3-GAL-His6) and 2619 bp (HIS3-GAL-His8).
In order to test the strategy, fusions constructs were made for the genes SSB1, RRP3 (C.L. O'Day, F. Chavanikamannil and J. Ableson, submitted for publication) and RRP41 (see Tables 1 and 2).
To avoid ectopic integration at the HIS3 and/or GAL1-10 locus, recipient strains carrying both the his3-Æ200 and a GAL1-10 deletion were used (strains YDL401 and YDL402, Table 3). 250-500 ng of the purified PCR fragment were used for transformation with the LiAc technique (2). 5-10 transformants were typically obtained per transformation (Table 2). Integration at the correct chromosomal locus was verified by PCR amplification on DNA from yeast colonies, using primers flanking the sites of integration (data not shown). Different constructs gave frequencies of correct integration ranging from 40-100% (Table 2).
Figure 2. Construction and integration of fusions. In the example illustrated, PCR amplifications were performed on the pTL27 vector (pGAL-ProtA) using 5' and 3' primers. Each primer contains a sequence required for amplification on the template DNA and a sequence from SSB1 required to specifically target the integration. The 5' primer includes 45 nucleotides of the SSB1 promoter region. The 3' primer includes the sequence complementary to the first 45 nucleotides of the SSB1 ORF. PCR products were transformed into strain YDL401 and transformants selected for histidine prototrophy. Homologous recombination leads to the integration of the PCR cassette generating a chromosomal GAL-ProtA::SSB1 allele.
Expression of the tagged alleles was checked by Western blotting (shown for the strain expressing the ProtA::Ssb1p fusion in Fig. 3). SSB1 is a non-essential gene and transformants were directly plated on 2% glucose minimal medium lacking histidine (SD-his). For the essential genes RRP3 (C.L. O'Day, F. Chavanikamannil and J. Ableson, submitted for publication) and RRP41 (P. Mitchell, D. Lafontaine and D. Tollervey, unpublished), transformants were plated under permissive conditions for GAL transcription; minimal medium lacking histidine and containing 2% sucrose, 2% raffinose and 2% galactose. Strains were then streaked on SD-his to check for the effects of depletion. The expression level of the ProtA::Ssb1p fusion was tested on medium lacking histidine and containing 4% sucrose + 2% galactose (Fig. 3, lane 4), 2% sucrose + 2% raffinose + 2% galactose (Fig. 3, lane 5) and 4% raffinose + 2% galactose (Fig. 3, Lane 6). Galactose functions as a non metabolizable inducer in strains YDL401 and YDL402 since they carry the galÆ108 mutation (3) (see Table 3).
Figure 3. Western-blot analysis of the GAL-ProtA::SSB1 strain.
Lane 1: strain expressing ProtA::Nop1p, Lane 2: negative control, strain YDL401. Both control strains were grown in YPD. Lanes 3 to 6: GAL-ProtA::SSB1 strain (YDL500). Strain YDL500 was grown in minimal medium lacking histidine supplemented with 2% glucose (Lane 3), 4% sucrose + 2% galactose (Lane 4), 2% sucrose + 2% raffinose + 2% galactose (Lane 5) or 4% raffinose + 2% galactose (Lane 6). Expected sizes of ProtA::Nop1p and ProtA::Ssb1p are 49.177 kDa and 47.530 kDa respectively. Both wild-type proteins migrate abnormally slowly, probably due to the presence of glycine and arginine rich domains (GAR domains); Nop1p (Mr 34,5 kDa) and Ssb1p (Mr 32,853 kDa) migrate with apparent sizes of 38 kDa (8) and 43 kDa (9), respectively. Degradation products were detected for both ProtA::Nop1p and ProtA::Ssb1p. For protein extraction, cells equivalent to 5 OD600 units were harvested and resuspended in 100μl of SDS loading buffer with 25μl glass beads. Cells were vortexed for 1 min and incubated for 1 min at 95¡C three times successively. Lysates were cleared by centrifugation for 10 min at 14.000 rpm and supernatant equivalent to 0.375 OD600 units of cells was loaded per lane. Samples were run on 15% SDS-PAGE gels and blotted according to standard procedures. Western blots were decorated using appropriate antibodies and developed using the ECL detection kit (Amersham). Antibodies used to detect the tagged-proteins are: rabbit peroxidase-anti-peroxydase (PAP; Sigma, Cat. No. P2026) for ProtA fusions, Mouse Mab clone 9E10 (Cambridge Research Biochemicals, Cat. No. OM-11-908) for c-myc fusions and the MRGS.His antibody (QIAGEN, Cat. No. 34610) for poly-His fusions. ÿ These strains also carry a mutation in the galactose permease gene (gal2). The effects of galactose addition to medium containing 2% sucrose and 2% raffinose was tested for the GAL-His8::SSB1 strain. As expected, the presence of galactose in the medium was found to have little effect on the level of expression of the fusion protein (data not shown). A potential problem with the use of GAL-regulated constructs is that many proteins are heavily over- expressed when their genes are transcribed from induced GAL promoters. This, for example, can make the analysis of the sub-cellular localisation of the fusion protein unreliable. Ssb1p is an snoRNP protein and the level of ProtA::Ssb1p was compared to the level of expression of another snoRNP protein, ProtA::Nop1p, expressed under the control of its own promoter (Fig. 3, lane 1). In these strains the level of ProtA::Ssb1p expressed in medium containing 4% sucrose + 2% galactose is similar to that of ProtA::Nop1p, suggesting that its expression is in the same range as expression of endogenous Ssb1p. During growth on medium containing 2% glucose (Fig. 3, lane 2) the level of ProtA::Ssb1p was undetectable. Many GAL-regulated mutants show incomplete growth inhibition on glucose medium due to residual transcription (4). The effects of the transcriptional repression can be enhanced at the translational level through modification of the context of the initiator AUG or by the introduction of an additional, out of frame upstream AUG sequence. In the system reported here, such mutants can simply be made by altering the sequence of the 3' primer (primer 3' in Fig. 2). The URA3 gene of Kluyveromyces lactis is functionally homologous to the S. cerevisiae URA3 gene and fully complements ura3- strains, but has sufficient sequence divergence to prevent genetic recombination (5). To allow epitope-tagging of more than one protein in the same strain we are currently constructing vectors based on the K. lactis URA3 gene. Templates for the construction of carboxy-terminal fusions are also in preparation. The ease with which GAL-regulated and epitope-tagged alleles of genes of interest can be constructed using this strategy allows initial functional analyses of the effects of genetic depletion to be carried out using tagged alleles. This allows the degree of genetic depletion to be followed at the protein level in the absence of specific antibodies (see Fig. 3). The construction of such alleles by conventional techniques typically involves several cloning steps and generally generates only plasmid-borne alleles. In the case of essential genes, these must be transformed into heterozygous diploid stra ins and suitable haploid progeny recovered after sporulation. In contrast, the technique reported here allows mutant alleles of essential genes such as RRP3 and RRP41, to be simply constructed in haploid strains.