Transkingdom Genetic Transfer from Escherichia coli to Saccharomyces cerevisiae as a Simple Gene Introduction Tool

All yeast and bacterial strains used in this study are listed in Table 1, including the ones which are provided by the National Collection of Yeast Cultures (NCYC), United Kingdom, and the National Bio-Resource Project (NBRP) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The media used included yeast extract-peptone-dextrose (YPD) broth or agar for S. cerevisiae BY4742 and its derivatives, yeast extract-peptone-adenine-dextrose (YPAD) broth or agar for MaV203, and Luria-Bertani (LB) broth or agar for E. coli, prepared as recommended for the ProQuest two-hybrid system (Invitrogen, Carlsbad, CA). When required for selection of auxotrophic strains, synthetic complete (SC) media or agar lacking leucine and/or tryptophan or uracil were prepared. Antibiotics used were 50 μg ml⁻¹ ampicillin, 30 μg ml⁻¹ chloramphenicol, 50 μg ml⁻¹ kanamycin, 50 μg ml⁻¹ streptomycin, or 7.5 μg ml⁻¹ tetracycline as required. All chemicals were obtained from Wako Pure Chemical Ind., Ltd. (Osaka, Japan). Bacterial cultures were incubated at 37�C, and yeast cultures at 28�C, by following conventional methods.

Plasmids used in this study are listed in Table 1 and shown in Fig. 2. Primers used for the amplification of oriT^P, oriT^Q/mob, and LEU2 regions are listed in Table 2, as are those for the confirmation of the integrity of oriT^P and oriT^Q regions. For the construction of novel TKC vectors, a 287-bp fragment of oriT^P was PCR amplified from plasmid pRH210 (DDBJ accession number AB526839), which is derived from RK2, and inserted at the SnaBI site of plasmid pEXP-AD502 or at the blunt-end-treated Tth111I site of plasmid pDBLeu. Alternatively, a 2,978-bp fragment of oriT^Q/mob was PCR amplified from pAY205 (DDBJ accession number AB526841), which is derived from RSF1010, and inserted into the SnaBI site of pEXP-AD502. For the construction of pRS314::LEU2 2�oriT^P, the 287-bp fragment and another 275-bp SacI-digested fragment of oriT^P were inserted into the SnaBI site and the SacI site of pRS314, respectively, followed by insertion of a 2,091-bp BamHI/PstI-digested LEU2 fragment from pCR-XL-TOPO (Invitrogen, Carlsbad, CA) into the BamHI/PstI site of pRS314.

For plasmid isolation from yeast TKC transformants, each colony from the selection plate was picked up and thinly streaked on another selection plate to purify it. An isolated positive clone from this plate was again picked up and cultured in liquid medium for plasmid preparation. A Zymoprep yeast plasmid miniprep kit (Zymoresearch, Irvine, CA) was used for plasmid preparation from yeast. For plasmid preparation from E. coli, Wizard plus SV miniprep DNA purification systems (Promega, Madison, WI) and a QuickLyse miniprep kit (Qiagen, Duesseldorf, Germany) were used.

Aliquots of 12.5 μl each of E. coli and yeast suspensions in Tris-NaCl buffer (TNB; 10 mM Tris-HCl, pH 7.5, 0.5% NaCl) (7), containing 3.8 � 10⁶ and 1.0 � 10⁶ CFU, respectively, were mixed and incubated for 1 h at 28�C and were further processed on a plate with appropriate auxotrophic selection medium. In case of analysis of TKC carrying no helper plasmid or vector carrying no oriT sequence, the TKC reactions were performed using 3 or 10 times the amount of TNB and cells, respectively. In the case of analysis of vector carrying two oriT^P sequences, pRS314::LEU2 2�oriT^P, the TKC reaction was performed using twice the amount of TNB and cells.

For the analysis of TKC transfer of two plasmids simultaneously, the donor and recipient cells were cultured on LB and YPAD plates, respectively, containing appropriate antibiotics. The TKC reaction was scaled up to include 9.2 � 10⁸ and 8 � 10⁶ CFU of the donor bacterium and recipient yeast, respectively, and incubated in 200 μl of TNB to improve sensitivity and saturate the attachment of recipient with donor.

As the TKC phenomenon is not very well known among researchers (although it was first reported in 1989 [7]), we tried to reconfirm its existence. The TKC reaction was performed by mixing a donor strain from E. coli HB101, carrying the mobilizable vector pAY205 and the helper plasmid pRH210, and a recipient strain, S. cerevisiae BY4742 (Fig. 1). Uracil-prototrophic transformants were detected, while no transformant was observed when a micropore membrane filter was used to partition the donor and recipient cells or when the solution, which was obtained by filtering the donor suspension through the filter membrane, was mixed with the recipient cells (see Fig. S1a in the supplemental material). The TKC phenomenon was also not observed in the donor lacking the helper plasmid, which carries tra genes (see Fig. S1b). These results show that transfer of the vector requires a conjugal transfer system and not incorporation of the vector from the reaction solution released by cell lysis of the donor.

If a shuttle vector transfers as a linear ssDNA molecule without telomeric sequences, it needs to be recircularized to be stably maintained in a yeast cell as an episome. We tested the efficiency of TKC transfer from E. coli HB101 into S. cerevisiae strain BY4742 and compared it to that in five of its mutants known to be defective in DNA repair. We used the shuttle vector pAY205 (Fig. 2a), which possesses the oriT^Q region and mob genes derived from an IncQ-type plasmid, RSF1010, and has been used in TKC studies before (8). In addition to the parental strain, the recipient strains were rad50Δ, yku70Δ, and lig4Δ mutants that are defective in nonhomologous end joining (NHEJ) and also in transfer of T-DNA from Agrobacterium tumefaciens to S. cerevisiae. rad51Δ and rad52Δ are defective in homologous recombination (HR) but are capable of accepting DNA by T-DNA transfer (19). The TKC efficiency of the five mutants, estimated on the basis of the frequency of the URA3 marker, did not show significant differences from that of the wild-type parental strain (Fig. 3a). This result suggested that the DNA repair system in the recipient cell is not critical for transferring and recircularizing the replicative type vector. To investigate the possibility of modifications in the junction sequence around the oriT^Q region, we analyzed the transferred vector DNA sequence. The oriT^Q regions were PCR amplified from transformants (transconjugants) derived from parental and mutant yeast recipients. Ten randomly selected clones from each recipient, or a total of 60 clones, were analyzed. An amplicon of the expected size of 803 bp was detected in 59 transformants, while the remaining transformant that was derived from the wild-type recipient did not yield a product in PCR amplification (Fig. 3b). The nucleotide sequences of the oriT^Q region from the 59 transformants further confirmed that their junctions were preserved (Fig. 3c and d). These results indicated that the recircularization of transferred vectors is independent of the recipient yeast DNA repair system.

We constructed and compared two types of vectors that had the same shuttle vector backbone, pEXP-AD502, but differed in having either oriT^Q/mob genes or oriT^P (Fig. 2b and d). Vectors with the oriT^P regions in clockwise (cw) or anticlockwise (acw) orientation were generated. The TKC efficiency of plasmid pEXP-AD502 oriT^P (acw) was similar to that of its mirror vector, pEXP-AD502 oriT^P (cw), but it was 3.4-fold higher than that of pEXP-AD502 oriT^Q (cw). Similarly, TKC efficiency in vector pEXP-AD502 oriT^P (cw) was higher than that of pEXP-AD502 oriT^Q (cw) by 5.5-fold (Fig. 4a). To validate the TKC transfer, the oriT regions of transferred vectors were PCR amplified and sequenced. For each of the three vectors, 32 clones were analyzed, or 96 clones in total. All amplicons showed expected sizes (1,390 or 1,348 bp; Fig. 4b), and their junction sequences were preserved (Fig. 4c and d). From these data, we concluded that while the shuttle vectors tested here showed similar accuracy of TKC transfer, the oriT^P-based vectors were more efficient than oriT^Q-based vectors in terms of the number of transconjugants obtained.

An ability to simultaneously introduce multiple vectors via a single reaction would be a key advantage of genetic manipulation. The simultaneous TKC transfer of two vectors, pEXP-AD502 oriT^P (cw) and pDBLeu oriT^P (cw), from E. coli HB101 to S. cerevisiae MaV203 was investigated using two different methods (Fig. 2). In the first method (redundant transfer), both vectors were first introduced into the same bacterial donor strain that was used for TKC transfer to yeast (Fig. 5a). Successful transfer of the vectors was examined by growth on selective media to identify exconjugants carrying the TRP1 and/or LEU2 marker genes for tryptophan and leucine auxotrophy, respectively. Double transformants (TRP1⁺ LEU2⁺) were detected at an efficiency of 1.6 � 10⁻⁶ transformants/recipient cell, which is less than one-sixth the efficiency of transformants possessing at least one of the two vectors (Fig. 5b). In a second method (redundant conjugation), the TKC reaction was performed between a single recipient yeast strain and two independent donor strains, each carrying one of the two vectors (Fig. 5c). However, no double transformant was detectable at the experimental scale (<7.7 � 10⁻⁸ transformants/recipient cell) (Fig. 5d). Thus, the efficiency of the redundant transfer method was at least 20-fold higher than that of the redundant conjugation method.

From the viewpoints of biosafety and size minimization of transferred DNA, it is advisable to remove antibiotic resistance genes and the origin of vegetative replication (oriV) from the transferred DNA. A vector designed to test the excision of the TRP1 marker was flanked by two direct repeats of the oriT^P region in the same orientation (Fig. 6a). Among 100 transformants, which were selected by leucine prototrophy, 17 transformants lacked tryptophan prototrophy, indicating excision of TRP1. The transferred vectors were extracted from 10 of the transformants with leucine prototrophy but without tryptophan prototrophy (Leu⁺/Trp⁻) and used as templates for PCR amplification to detect the oriT and its flanking region. However, we could not design appropriate PCR primers which worked well for detection of their status of excision. Thus, the same plasmids were used to transform E. coli. Restriction digestion analysis of the plasmids recovered from E. coli transformants confirmed the absence of TRP1 DNA (Fig. 6b), and sequencing results indicated that the recircularization had occurred precisely between the two nick sites in all of the plasmids (Fig. 6c and data not shown). These results indicate that both of the oriT genes were recognized and had been excised along with the intervening TRP1 gene.