IpO: Plasmids and methods for simplified, PCR-based DNA transplant in yeast

Abstract

Many yeast experiments require strains modified by recombinant DNA methods. Some experiments require precise insertion of a DNA segment into the genome without a selectable marker remaining. For these applications, we developed a new PCR-based method for marker-free DNA transplant. The current PCR-based method requires the labor-intensive construction of a PCR template plasmid with repeats of the DNA segment flanking URA3. The design of a new vector, IpO, reduces the work to cloning a single copy of the DNA segment between overlapping URA3 fragments present in the vector. Two PCRs are performed that capture the DNA segment and one or the other URA3 fragment. When the PCR products are co-transformed into yeast, recombination between the overlapping URA3 fragments restores URA3 and transposes the cloned DNA segment inside out, creating a repeat-URA3-repeat cassette. Sequences designed into the PCR primers target integration of the cassette into the genome. Subsequent selection with 5-fluoro-orotic acid yields strains that have popped out URA3 via recombination between the DNA repeats, with the result being the precise insertion of the DNA segment minus the selectable marker. An additional advantage of the IpO method is that it eliminates PCR artifacts that can plague the current method’s repeat-containing templates.

INTRODUCTION

Many yeast experiments require strains modified by insertion of a DNA segment at a specific site in the genome. In some experiments, such as deleting a gene by one-step gene replacement, the DNA segment serves only as a selectable marker for transformation. In other experiments, the DNA segment plays a wider role in the experimental analysis. Examples of this class of DNA segments include transcriptional regulatory elements, transcriptional reporters, and recombinant tags for protein detection. In this report, we focus on the latter class of DNA segments, and in particular the methods used to insert them at specific sites in the genome. The DNA segment can be derived from yeast, another organism, or be purely synthetic. Because the methods entail the transfer of the DNA segment from its source to the genome of the host yeast, we refer to process as DNA transplant.

Two common approaches for transplanting DNA segments are one-step methods that retain the selectable marker used for transformation and two-step methods that can result in marker-free DNA insertion. The choice of which to use is dictated by the particular experiment. Although marker-free DNA transplant requires significantly more work, some experiments demand it. Examples include inserting a transcriptional regulatory element to place a gene under heterologous control while avoiding effects that might be associated with the selectable marker, inserting DNA encoding a fluorescent tag within an open reading frame (ORF), and constructing a strain with several modifications that would make the accumulation of selectable markers impractical.

The original method for marker-free DNA insertion requires two steps and is referred to as “pop-in/pop-out” gene replacement [Scherer and Davis, 1979]. A length of genomic target DNA is cloned into an integrating plasmid containing the selectable/counterselectable URA3 marker. Next, the DNA segment that will be transplanted into the genome is inserted into the cloned DNA by site-specific mutagenesis. The first step is transformation of yeast with the new plasmid that has been cut at a restriction site within the cloned DNA to target “pop-in” integration. Transformation with a circular plasmid produces on the chromosome a direct repeat of the cloned genomic DNA separated by the URA3 plasmid, with the inserted DNA segment on one side and wildtype on the other. The second step is initiated by culturing cells in the absence of uracil selection. At a low frequency, cells arise in the population that have undergone recombination between the direct repeats flanking the URA3 plasmid. These rare “pop-out” recombinants are selected on 5-fluoro-orotic acid (FOA) medium, which is toxic to URA3 cells [Boeke et al. 1984]. The resulting strains retain either the inserted DNA segment or wildtype DNA, depending on where the pop-out crossover occurred. The original method, although elegant, is cumbersome because it is a multi-week process that requires construction of two new plasmids for each desired chromosomal change.

There is a PCR-based method available that simplifies marker-free DNA transplant in yeast [Schneider et al., 1995; Khmelinskii et al., 2011]. It is also a two-step method and uses the basic principles of the original: transformation with linear DNA that has yeast sequences on each end to target pop-in integration, and FOA selection for pop-out recombination between direct DNA repeats flanking URA3. A clever twist of the PCR-based method is that the DNA segment itself becomes the repeated DNA that flanks URA3. Because of this reconfiguration, the only plasmid that needs to be made is a PCR template that contains, in this order, the DNA segment, URA3, and a second copy of the DNA segment (repeat-URA3-repeat, see for example Figure 1A). To create a repeat-URA3-repeat cassette for yeast transformation, a PCR is performed using the plasmid template and primers that contain both yeast sequences to target integration and plasmid sequences to prime PCR.

Figure 1.

PCR amplification from a repeat-URA3-repeat template. (A) pJH104, a TEF1p-URA3-TEF1p plasmid. U2 and D2 priming sequences flank the cassette. Primers URA3.34.7 and URA3.34.8 are used for split-URA3 PCRs. (B) PCRs with different primer combinations. The BAR1 sequences (grey bars) have 60 bp genomic sequence followed by U2 or D2. PCR products are illustrated below the three PCRs. The 556 bp artifact product in PCR 1 corresponds to BAR1-U2-TEF1p-D2-BAR1, as determined by cloning and sequencing of similar products. (C) Gel analysis of the three PCRs shown in (B).

The PCR-based method simplifies DNA transplant by eliminating the need to construct two plasmids for each genomic target. Another advantage is that having the DNA segment as the direct repeat means that all pop-out recombinants retain the DNA segment. Nonetheless, the method has room for improvement. Construction of the repeat-URA3-repeat PCR template plasmid is labor-intensive, requiring either a four-part ligation or construction of three sequential plasmids made by two-part ligation. The work is multiplied when one wishes to make template plasmids for a variety of DNA segments. Moreover, PCR amplification of repeat-containing templates is problematic, producing artifacts or failing altogether.

In a study to interrogate long non-coding RNA (ncRNA) function in yeast, we wished to place a set of ncRNAs under heterologous transcriptional control by inserting a variety of promoters and terminators near ncRNA transcription start sites (TSS). Because ncRNAs are often close to or overlap other annotated genomic features (e.g. see [Lardenois et al., 2011]), it is essential to use marker-free DNA transplant to insert the transcriptional control elements. Our first approach was to use the existing PCR-based method. However, while constructing and using PCR template plasmids, we encountered the problems noted above: labor-intensive plasmid construction and problematic PCRs using repeat-URA3-repeat templates. In this report, we describe how we solved both problems by creating a new PCR template vector and by reconfiguring the PCRs used to make DNA for yeast transformation.

MATERIALS AND METHODS

Standard techniques were used for DNA manipulations [Ausubel et al., 1995]. DNA oligonucleotides (Table 1) were synthesized by Eurofins MWG Operon (Huntsville, AL). We used the following PCR polymerases: Phusion Hot Start II High Fidelity (Thermo Scientific) to amplify DNA that would be cloned or directly sequenced, Takara ExTaq to amplify DNA for yeast transformation, and Phire Hot Start II (Thermo Scientific) for genomic confirmation PCR. Cloned DNA segments were sequenced in their entirety, except where noted for pJH104. Yeast media, culture conditions, and molecular techniques were as described by [Amberg et al., 2005]. Assay of α-factor secretion was as described by [Sprague, 1991].

Table 1.

PCR Primers

Name	Purpose	Sequence (5′ to 3′)
TEF1.34.1	Forward XhoI-U2-TEF1p-XbaI, pJH104	TTCCTCGAGCGTACGCTGCAGGTCGACGA ATCCTTACATCACACCCAATC
TEF1.34.2	Reverse XhoI-U2-TEF1p-XbaI, pJH104	GGTTCTAGAATCAATGGGAGGTCATCGAA AG
URA3.34.1	Forward XbaI-URA3-BamHI, pJH104	GGTTCTAGAAGAGTGCACCACGCTTTTCAA T
URA3.34.2	Reverse XbaI-URA3-BamHI, pJH104	CGCGGATCCTTTCACACCGCAGGGTAATA AC
TEF1.34.3	Forward BamHI-TEF1p-D2-EagI, pJH104	GGTTCTAGACGCGGATCCCACTTCAAAAC ACCCAAGCACA
TEF1.34.4	Reverse BamHI-TEF1p-D2-EagI, pJH104	CTTCGGCCGTCATCGATGAATTCGAGCTC GTTTGTAATTAAAACTTAGATTAGATTGC TATGCTTTC
URA3.34.7	Reverse split-URA3 PCR, pJH104	AGTATATTCTCCAGTAGATAGGGAGCC
URA3.34.8	Forward split-URA3 PCR, pJH104	TGCTGCTACTCATCCTAGTCCTG
BAR1.U2	60 nt homology upstream of BAR1 ATG + U2	AGTGGTTCGTATCGCCTAAAATCATACCA AAATAAAAAGAGTGTCTAGAAGGGTCAT ATACGTACGCTGCAGGTCGAC
BAR1.D2	Reverse complement of 60 nt homology from BAR1 ATG to +60 of ORF + D2	GTTAATAATCGCGAAACTCGCCAAAATAA GTTTCAAACAAAGATGATTAATTGCAGAC ATATCGATGAATTCGAGCTCG
URA3.39.3	Forward XhoI-URA3 5′-KpnI, pJH121	TTCCTCGAGTGTGGTTTCAGGGTCCATAAA G
URA3.39.4	Reverse XhoI-URA3 5′-KpnI, pJH121	GTGGGTACCAGTATATTCTCCAGTAGCTA GGGAGCC
URA3.39.1	Forward SacI-URA3 3′-EagI, IpO	AGGGAGCTCACAGTTAAGCCGCTAAAGGC ATTA
URA3.39.2	Reverse SacI-URA3 3′-EagI, IpO	CTTCGGCCGGGTAATAACTGATAT AATTAAATTGAAGC
aTEF1.39.1	Forward XbaI-UP-AgTEF1p-DN- XhoI, pJH124	GGTTCTAGACCGGTGCATGCGCTAGCTGTG ACTGTCGCCCGTACAT
aTEF1.39.2	Reverse XbaI-UP-AgTEF1p-DN- XhoI, pJH124	TTCCTCGAGGACGTCTGTACACCGCGGTTT TGTTTATGTTCGGATGTGATGTG
GAL1.50.1	Forward NheI-GALp-SacII, pJH131	AAAAGCTAGCTTATATTGAATTTTCAAAA ATTCTTACTTTTTTTTTG
GAL1.50.2	Reverse NheI-GALp-SacII, pJH131	AAAACCGCGGTATAGTTTTTTCTCCTTGAC GTTAAAGTATAGAGG
BAR1.61.1	60 nt homology upstream of BAR1 ATG + UP	AGTGGTTCGTATCGCCTAAAATCATACCA AAATAAAAAGAGTGTCTAGAAGGGTCAT ATAACCGGTGCATGCGCTAGC
BAR1.61.2	60 nt homology upstream of BAR1 control sequences (−338 to −279) + UP	ACATTTGTTGCATTTATTATATTAGAGATG CGTTGTCCCTGTTTTTCTACCTCCGACATC ACCGGTGCATGCGCTAGC
BAR1.61.3	Reverse complement of 60 nt homology from BAR1 ATG to +60 of ORF + DN	GTTAATAATCGCGAAACTCGCCAAAATAA GTTTCAAACAAAGATGATTAATTGCAGAC ATGACGTCTGTACACCGCGG
URA3.for	Forward primer for split-URA3 PCR, IpO plasmids	CACAGTTAAGCCGCTAAAGGC
URA3.rev	Reverse primer for split-URA3 PCR, IpO plasmids	AGTATATTCTCCAGTAGCTAGGGAGCC

U2 and D2, single underline. UP and DN, double underline. Restriction sites, italics.

PCR template plasmid pJH104 (Addgene 48262) carries two copies of the TEF1 promoter flanking URA3. U2 and D2 sequences provide priming sites for PCR [Chu and Davis, 2008]. pJH104 was constructed as follows. The XhoI-U2-TEF1p-XbaI segment was PCR amplified from NKY210 [Raymond and Kleckner, 1993] genomic DNA with primers TEF1.34.1 and TEF1.34.2; TEF1p sequences are from −460 to −128 relative to the TEF1 start codon. XbaI-URA3-BamHI was amplified from pRS316 [Sikorski and Hieter, 1989] with primers URA3.34.1 and URA3.34.2. The BamHI-TEF1p-D2-EagI segment was amplified from NKY210 with primers TEF1.34.3 and TEF1.34.4; TEF1p sequences are from −343 to −1. The three DNA segments were digested with restriction enzymes and combined in a four-part ligation with XhoI and EagI digested pBluescript II SK (−) (Agilent Technologies). The cloned TEF1p segments and most of URA3 were confirmed by DNA sequencing. The TEF1p segments share 215 bp overlap. When the TEF1p-URA3-TEF1p cassette is integrated in yeast, recombination between the TEF1p repeats “pops-out” the intervening URA3 marker and restores the full-length, 460 bp TEF1 promoter.

IpO (Addgene 48233) is the base vector for building DNA transplant plasmids. It carries overlapping URA3 fragments flanking a multiple cloning site (Figure 2). IpO was constructed in two steps. First, the URA3 5′ fragment was PCR amplified from JHY222 genomic DNA (S288c background, [Lardenois et al., 2011]) with primers URA3.39.3 and URA3.39.4, which add XhoI and KpnI sites; URA3 sequences are from −242 to +495. The resulting product was digested with XhoI and KpnI and inserted into the same sites of pBluescript II SK (−) to create pJH121. Next, the URA3 3′ fragment was likewise amplified with primers URA3.39.1 and URA3.39.2, which add SacI and EagI sites; URA3 sequences are from +216 to 80 bp downstream of the stop codon. The resulting product was digested with SacI and EagI and inserted into the same sites of pJH121, creating IpO. The URA3 fragments share 280 bp overlap.

Figure 2.

IpO, a vector for building DNA transplant plasmids. The URA3 fragments share 280 bp overlap (ORF coordinates +216 to +495). The multiple cloning site has nine unique restriction sites.

We created two promoter transplant plasmids based on IpO. Each has UP and DN priming sites flanking the DNA insert. Thus, either promoter can be transplanted using a single pair of gene-specific primers with UP and DN sequences on their 3′ ends. pJH124 (Addgene 48259) carries the 284 bp Ashbya gossypii TEF1 promoter. It was created by PCR amplification of the Ag TEF1 promoter from pFA6a-kanMX4 [Wach, et al, 1994] with primers aTEF1.39.1 and aTEF1.39.2, which add XbaI plus UP before Ag TEF1p, and DN plus XhoI after. The resulting product was digested with XbaI and XhoI and cloned into the same sites of IpO. The TEF1 promoter sequences in pJH124 differ from those found in common vectors. pJH124 contains exactly the intergenic region between the tandem-transcribed A. gossypii MRL1 and downstream TEF1 genes. The TEF1 promoter in pFA6-kanMX4 adds 95 bp of the MRL1 ORF and leaves out 2 bp before the TEF1 start codon. Our TEF1 promoter may be advantageous because it is shorter (284 vs. 378 bp) and contains all sequences prior to TEF1. Based on the tandem orientation of MRL1 and TEF1, the TEF1 promoter is likely unidirectional. In pJH124, the TEF1 promoter is oriented towards the DN sequence and the URA3 5′ fragment. Plasmid pJH131 (Addgene 48260) carries the 668 bp S. cerevisiae GAL10/GAL1 bidirectional promoter, corresponding to the entire sequence between the two divergent genes. It was created by PCR amplification of GAL10/GAL1 promoter DNA from JHY222 with primers GAL50.1 and GAL50.2, which add NheI and SacII sites. The resulting product was digested with NheI and SacII and cloned into the same sites of pJH124, which are present in the UP and DN sequences, respectively. The GAL1 promoter is oriented towards the DN sequence and the URA3 5′ fragment.

We used the IpO method to create several MATα strains with either the TEF1 or GAL1 promoter transplanted upstream of the MATa-specific BAR1 gene. In one set of strains the promoters were used to replace the BAR1 upstream control region from −278 to −1 [Kronstad et al., 1987]. Transplant of the Ag TEF1p promoter is illustrated in Figure 3. Two split-URA3 PCRs (Figure 4A) were performed for each pJH124 and pJH131 using the following primer pairs: BAR1.61.2 and URA3.rev for the promoter-URA3 5′ fragment, and URA3.for and BAR1.61.3 for the URA3 3′-promoter segment. In another set of strains the promoters were inserted directly upstream of the BAR1 start codon without replacing any DNA. The split-URA3 PCRs were the same as for the promoter replacement PCRs above, except that the promoter-URA3 5′ fragment was amplified with BAR1.61.1 and URA3.rev primers. Split-URA3 PCR pairs were mixed 1:1 (22.5 μL each, ~7 μg total DNA) and used without concentration or purification to transform JHY337 (MATα ura3Δ0 leu2Δ0 lys2Δ0; [Horecka and Davis, 2014]) using the method of [Gietz and Schiestl, 2007]. One fifth of the transformations were spread on SC-Ura plates, which were replica plated to fresh plates after two days to eliminate the background lawn that is often present with PCR transformations. Over 100 colonies were obtained for each transformation. Correct transformants were identified by genomic PCR using primers specific to one side of the integrated URA3 cassette. For the heterologous Ag TEF1 promoter, 16 of 16 scored as correct. For the Sc GAL1 promoter, 10 of 80 scored as correct. Several independent transformants were streak purified on YPD plates and then a single colony of each was used to inoculate 3 mL YPEG broth (2.5% ethanol, 2% glycerol; used to prevent growth of petites). Cultures were grown to saturation (~2 × 10⁸ cells/mL) for 2 days at 30C. Twenty microlitres culture fluid (~4×10⁶ cells) was spread onto a plate of minimal complete medium containing 0.8 mg/mL FOA (US Biological). An average of 114 colonies arose on each plate (range 5 to ~200 colonies, six plates scored).

Figure 3.

Overview of the IpO method for DNA transplant. Illustrated is the replacement of BAR1 regulatory sequences with the Ag TEF1 promoter. (A) A portion of IpO-based plasmid pJH124, which contains the 284 bp Ag TEF1 promoter flanked by UP and DN priming sites. The two split-URA3 PCRs use primer pairs BAR1.61.2 + URA3.rev and URA3.for + BAR1.61.3. (B) Recombination and integration (dotted lines, “pop-in”) of the split-URA3 PCRs from (A) at the genomic BAR1 locus (coordinates relative to start codon). The grey bars attached to the Ag TEF1p (T) sequences are the 60 bp BAR1 sequences designed into the primers. (C) Structure of the genomic BAR1 locus, with the AgTEF1p-URA3-AgTEF1p cassette replacing precisely BAR1 sequences from −278 to −1. The dotted line indicates the directly repeated Ag TEF1p segments that can undergo homologous recombination (“pop-out”) to yield the product shown in (D), which is the Ag TEF1 promoter precisely transplanted upstream of the BAR1 ORF.

Figure 4.

Use of the IpO method for heterologous regulation of BAR1 in MATα cells. Numbers correspond to: 1, Ag TEF1p inserted at BAR1 start (@ATG); 2, Ag TEF1p replacing BAR1 UAS (@UAS); 3, GAL1p inserted at BAR1 start; 4, GAL1p replacing BAR1 UAS. (A) Split-URA3 PCR amplification of DNA for transformation. Primer pairs are BAR1.61.1 (or BAR1.61.2) + URA3.rev (left) and URA3.for + BAR1.61.3 (right). Marker (M) sizes for (A) and (B) are given in kilobasepairs (kb); the bright band is 3.0 kb. (B) Genomic PCR of wildtype (wt) BAR1 and promoter transplant strains following FOA selection. Expected sizes (bp): wt (649); 1, Ag TEF1p@ATG (969); 2, Ag TEF1p@UAS (691); 3, GAL1p@ATG (1353); 4, GAL1p@UAS (1075). (C) Heterologous expression in MATα cells of barrier, the α-factor protease. Halo assays were performed by suspending MATa bar1 cells in water and spreading on YPGalactose (Gal) and YPDextrose (Dex) plates. MATα cells were taken from Gal or Dex stock plates and patched onto the lawns. Plates were incubated at 30C and photographed after one (Dex) or two (Gal) days. Secreted α-factor inhibits growth of the MATa bar1 lawn, creating a halo (e.g. wt). Simultaneous expression of barrier protease in the same cell inactivates α-factor.

RESULTS

In a study to be published elsewhere, we wished to alter the expression of many long ncRNAs at their genomic loci. A class of ncRNAs of particular interest are those that are close to or overlap with genes of known function (e.g. see [Lardenois et al., 2011]), which makes them a challenge for heterologous regulation. One method we considered was PCR-based, one-step promoter insertion to place a kanMX6-PGAL1 promoter cassette at the ncRNA TSS [Longtine et al., 1998]. Although this method is amenable to the study of many ncRNAs, we were concerned about off-target effects on nearby genes that might be caused by the kanMX6 marker. Ideally, a minimal regulatory DNA segment could be inserted at an ncRNA TSS without a marker or other DNA remaining. One method that can accomplish this is classical two-step gene replacement [Scherer and Davis, 1979]. This method requires that new plasmids be built for each target. Because we wanted to study many ncRNAs, the plasmid-based method was not a practical option. Instead, we chose as a first approach to adapt a PCR-based, two-step method that has been used to introduce marker-free epitope and fluorescent tags into genomic coding sequences [Schneider et al., 1995; Khmelinskii et al., 2011].

The present PCR-based DNA insertion method requires construction of a PCR template plasmid containing two repeats of the DNA to be inserted flanking URA3. The repeat-URA3-repeat construct also has priming sites that allow PCR amplification of the entire cassette using primers that include genomic targeting sequences [Schneider et al., 1995; Khmelinskii et al., 2011]. We used this base architecture to construct pJH104, which has repeats of the S. cerevisiae TEF1 promoter instead of sequences encoding a recombinant protein tag (Figure 1A). As a proof of principle for promoter transplant using pJH104, we chose to target BAR1, which encodes the secreted α-factor protease for which there is a convenient plate assay. When attempting to PCR amplify the entire cassette for transformation using BAR1.U2 and BAR1.D2 primers, we analyzed the reaction on a gel and found not only the expected full-length product, but also a shorter PCR product (Figure 1C). The shorter product appeared with all of the PCR polymerases we tried (ExTaq, Phusion Hot Start II, Phire Hot Start II, Invitrogen Platinum Taq, Agilent Herculase II), and even with the optimized PCR conditions recommended for similar repeat-containing templates [Khmelinskii et al., 2011]; some of our repeat-URA3-repeat PCRs failed altogether (data not shown). To determine the structure of the smaller product, we cloned and sequenced it. Two independent clones had the structure of BAR1-U2-repeat-D2-BAR1 (illustrated in Figure 1B, data not shown), which suggests melting and reannealing of nascent PCR products during the extension step. Our finding of PCR artifacts when amplifying from a repeat-URA3-repeat template is not unique [de Hoogt et al., 2000; Khmelinskii et al., 2011; Schneider et al., 1995].

We wanted to eliminate the PCR artifact because it reduces desired product yield and may reduce transformation efficiency. We took advantage of the fact that selectable markers split into overlapping fragments can be co-transformed into yeast and recombine with each other to restore a full-length, functional marker [Fairhead et al., 1996]. With this in mind, we designed two URA3 primers that could be used individually with primers BAR1.U2 and BAR1.D2 and template pJH104 to create PCR products that share 215 bp URA3 overlap (Figure 1A). The product of one PCR was predicted to be BAR1 sequences, TEF1p, and the URA3 5′ fragment (PCR 2, Figure 1B). The other was predicted to be the URA3 3′ fragment, TEF1p, and BAR1 sequences (PCR 3, Figure 1B). Gel analysis showed robust amplification of DNA molecules of the expected sizes and, notably, none of the artifact product seen with the single PCR (Figure 1C). Thus, separating the repeats into two PCRs solved the artifact problem. The pair of split-URA3 products were mixed and used to transform yeast. After identifying correct integrants, cells were grown in the absence of uracil selection to allow for recombination between the TEF1p repeats and pop out of URA3. Following selection on FOA plates, strains were isolated and confirmed by genomic PCR, DNA sequencing, and plate assay to have the TEF1 promoter inserted upstream of BAR1 (data not shown).

We learned two lessons from constructing and using pJH104 for TEF1p promoter transplant. First, creating a repeat-URA3-repeat PCR template plasmid is not trivial. We built pJH104 by combining three DNA segments with the vector in a four-part ligation. An alternative approach requires three sequential, two-part plasmid constructions. Laborious plasmid construction can be a bottleneck if one wants to build many PCR template plasmids. Second, PCR amplification of repeat-URA3-repeat templates is problematic, producing an artifact product that lowers desired product yield or failing altogether. By reconfiguring PCR template plasmid design and the protocol for cassette amplification, the method can be greatly simplified and made more robust.

The ability of split-URA3 PCRs to remedy the problematic amplification of repeat-URA3-repeat cassettes provided the foundation for reconfiguring PCR template plasmid design. If two separate PCRs are required, then there is no need to build a complicated repeat-URA3-repeat plasmid. The template requirements then become: 1) the DNA segment cloned upstream of a URA3 5′ fragment and 2) the DNA segment cloned downstream of a URA3 3′ fragment. Instead of building two independent vectors [Fairhead et al., 1996; Reid et al., 2002], we incorporated both requirements into a single vector. Plasmid IpO contains both URA3 5′ and URA3 3′ fragments flanking a multiple cloning site (Figure 2). The DNA segment that one wishes to transplant into the genome is cloned between the overlapping URA3 fragments in a simple, two-part ligation. The resulting plasmid is then used for split-URA3 PCRs, with one reaction amplifying the cloned DNA coupled to the URA3 5′ fragment, and the other amplifying the URA3 3′ fragment coupled to the cloned DNA (Figure 3). When the products are co-transformed into yeast and recombine, the cloned DNA segment that was located inside the URA3 fragments in the plasmid is now repeated and transposed outside the restored URA3 marker (inside-out, “IpO”). Adding genomic sequences to the 5′ ends of two of the split-URA3 PCR primers allows targeted integration of the repeat-URA3-repeat cassette.

As a proof of principle for the IpO vector and method, we again chose to target the BAR1 gene. We built two promoter transplant plasmids based on IpO. Plasmids pJH124 and pJH131 carry the A. gossypii TEF1 and S. cerevisiae GAL10/GAL1 promoters, respectively. We added UP and DN primer binding sites flanking the promoters to allow transplant of either promoter with the same set of BAR1 primers. One approach was to replace the BAR1 upstream control sequences (−278 to −1) with the promoters. The method for transplanting TEF1p in place of the BAR1 upstream sequences is shown in Figure 3. Transplanting the GAL10/GAL1 promoter is the same, except the template is pJH131 instead of pJH124. A second approach was to insert the promoters upstream of the BAR1 start codon without replacing any sequences. Split-URA3 cassettes were amplified by PCR (Figure 4A) and used to transform MATα cells, which do not express BAR1. Following transformation, screening for correct integrants, and selecting for URA3 pop-out strains on FOA plates, we identified correct promoter transplant strains of each type, which were verified by genomic PCR (Figure 4B) and DNA sequencing (data not shown). Halo assays to detect barrier activity confirmed the correct pattern of heterologous BAR1 regulation (Figure 4C). In this assay, secreted α-factor inhibits growth of the MATa bar1 lawn, creating a halo (e.g. wt in Figure 4C). Simultaneous expression of barrier protease in the same MATα cell inactivates α-factor, and therefore the lawn is not affected. The constitutive TEF1 promoter directed BAR1 expression on both YPDextrose and YPGalactose plates, as expected (Figure 4C). Likewise, the regulatable GAL1 promoter showed the correct pattern of BAR1 expression, with barrier being expressed on YPGalactose plates, but not on YPDextrose plates (Figure 4C). We found no observable difference between replacing the BAR1 upstream sequences and simply inserting the promoters before the start codon. This result was not necessarily expected, because the α2-Mcm1 operator located between −264 and −235 upstream of BAR1 has been shown to have a dominant, repressive effect when placed between heterologous CYC1 UAS and TATA promoter sequences [Zhong and Vershon, 1997]. It is interesting that even the shorter 284 bp Ag TEF1 promoter inserted before the BAR1 start codon was not measurably repressed by the presence of the α2-Mcm1 operator.

DISCUSSION

Here we have described the IpO vector and methods for simplified, PCR-based DNA transplant in yeast. Two features that set it apart from the existing PCR-based method are the IpO base vector and the use of split-URA3 PCR to amplify DNA for yeast transformation. The IpO vector streamlines construction of PCR template plasmids by requiring only a two-part ligation. This is a significant improvement over the complicated assembly of repeat-URA3-repeat template plasmids in the present method. The design of IpO places the cloned DNA segment both upstream of a 5′ URA3 fragment and downstream of a 3′ URA3 fragment. Two separate, split-URA3 PCRs from the same template produces DNA molecules that recombine upon yeast transformation to create direct repeats of the cloned DNA segment flanking an intact URA3 selectable marker. It is both the unique design of IpO with its split URA3 fragments and the efficient homologous recombination system in yeast that make the method possible.

Another improvement of the IpO method is the use of split-URA3 PCRs to prepare DNA for transformation. Single PCRs containing repeat-URA3-repeat templates are problematic. Our failed attempts to remedy the problem by trying different PCR polymerases and reaction conditions motivated us to reconfigure the strategy for PCR amplification of DNA from template plasmids. We found that split-URA3 PCR using template pJH104 is robust and free of detectable artifact products. This result, in turn, inspired us to reconfigure template plasmid design, resulting in the creation of the IpO vector. Because split-URA3 PCR with IpO-based plasmids uses two universal URA3 primers, transplanting a DNA segment to a new locus requires only two new genome-targeting primers. Thus, the method is also economical.

In our proof of principle experiment we used the IpO method to transplant the Ag TEF1 and Sc GAL10/GAL1 promoters upstream of BAR1. The method, of course, can be used to transplant virtually any DNA segment. Examples include transcription terminators, reporters such as lacZ or HIS3, epitope tags, fluorescent tags, and purification tags. For economy, we incorporated universal UP and DN primer sites flanking the Ag TEF1 and GAL10/GAL1 promoters so that we could use the same BAR1 primers to amplify both. This approach leaves the UP and DN 18 bp sequences flanking the inserted DNA in the final strain, which is probably acceptable for promoter and terminator transplant. However, there are cases in which seamless DNA transplant is required, such as the insertion of an epitope tag within an ORF. In such cases, genomic-targeting primers can be designed with 3′ ends that hybridize to both sides of the DNA segment cloned into IpO. This is not an option with methods based on exact repeat-URA3-repeat templates, because there would be two forward and two reverse primer-binding sites on the plasmid that would lead to additional problems with PCR.

In summary, we have developed a novel method for PCR-based DNA transplant in yeast. The IpO vector is unique in that it is designed with split URA3 fragments flanking the multiple cloning site and that it takes advantage of the efficient homologous recombination system in yeast to reconstruct an intact URA3 marker. The IpO method is an improvement over current PCR-based methods because it simplifies construction of PCR template plasmids and allows for problem-free, robust amplification of DNA for yeast transformation. The IpO vector (48233) and the two promoter transplant plasmids pJH124 (Ag TEF1p, 48259) and pJH131 (Sc GAL10/GAL1p, 48260) are available from Addgene.