A novel kanamycin/G418 resistance marker for direct selection of transformants in Escherichia coli and different yeast species



We have developed a set of cloning vectors possessing a modified Tn903 kanamycin resistance gene that enables the selection of both kanamycin-resistant transformants in Escherichia coli and G418-resistant transformants in the yeasts Saccharomyces cerevisiae, Hansenula polymorpha and Pichia pastoris. Expression of this gene in yeast is controlled by the H. polymorpha glyceraldehyde-3-phosphate dehydrogenase promoter, while expression in E. coli is governed by an upstream E. coli lacZ promoter. Applicability of the vectors for gene disruption in H. polymorpha and S. cerevisiae was demonstrated by inactivation of the HpMAL1 and URA3 genes, respectively. One of the vectors possesses a H. polymorpha ARS allowing plasmid maintenance in an episomal state. The small size of the vectors (2–2.5 kb) makes them convenient for routine DNA cloning. In addition, we report a novel approach for construction of gene disruption cassettes. Copyright © 2009 John Wiley & Sons, Ltd.


Genetic manipulations often require introduction of recombinant DNA constructs into cells and subsequent selection of transformants, which is ensured by the use of an appropriate plasmid selectable marker. Yeast recipient strains usually possess auxotrophic mutations, which can be complemented by the corresponding chromosomal genes introduced into genetic constructs as selectable markers. However, heterologous genes conferring resistance to some antibiotics (e.g. G418, hygromycin and Zeocin) can also be used as plasmid selectable markers (Davies and Jimenez, 1980; Gritz and Davies, 1983; Gatignol et al., 1987). Such markers are especially helpful if the host strain does not possess appropriate auxotrophic mutations, or if these mutations are required for subsequent genetic manipulations.

As a rule, creating a genetic construct for introduction into yeast cells involves an intermediate step of plasmid amplification in Escherichia coli. The use of markers that function in both E. coli and yeast may reduce the size of plasmid and simplify its modifications. A set of vectors possessing a Zeocin resistance marker that works in both E. coli and yeast has been developed (Invitrogen). However, the high cost of Zeocin restricts extensive use of plasmids with this marker. Here, we describe a novel selectable marker providing resistance to kanamycin in E. coli and to G418 in the yeasts Saccharomyces cerevisiae, Hansenula polymorpha and Pichia pastoris.

Materials and methods

Strains, media and culture conditions

Standard YPD (2% peptone, 1% yeast extract, 2% glucose) and SD (0.67% yeast nitrogen base, 2% glucose) media were used to cultivate yeast cells. E. coli was cultivated on 2YT medium (1% yeast extract, 1.6% tryptone, 0.5% NaCl) supplemented with 50 mg/l kanamycin. H. polymorpha DL1-L (leu2) (Sohn et al., 1996), S. cerevisiae 10B-H49 (MATα kar1-1 ade2-1 leu1 SUQ5 his3 lys1) (Ter-Avanesyan et al., 1994), P. pastoris GS115 (his4) (Cregg et al., 1985), and E. coli DH5α (Woodcock et al., 1989) strains were used in this work. Yeasts were transformed according to the standard procedure (Ito et al., 1983), with modifications for H. polymorpha (Bogdanova et al., 1995) and P. pastoris (Cregg and Russell, 1998). E. coli was transformed as previously described (Inoue et al., 1990). YPD supplemented with 200 mg/l G418 was used for selection of G418-resistant transformants of H. polymorpha and P. pastoris, while the S. cerevisiae transformants were selected on 150 mg/l G418. Prior to plating onto G418-containing medium, yeast transformation mixes were pre-incubated in liquid YPD for 1 h. S. cerevisiae and P. pastoris cells were cultivated at 30 °C, while H. polymorpha was grown at 37 °C.

Plasmid construction

Routine in vitro DNA manipulations were carried out as previously described (Sambrook et al., 1989). The pKAM444 plasmid was obtained in two steps (Figure 1). In the first step, the region of the pUK21 plasmid (Vieira and Messing, 1990) between the BspHI site 51 bp upstream from the aminoglycoside 3′-phosphotransferase (APH) open reading frame (ORF) and the XbaI site of the polylinker was replaced with the 0.6 kb XbaI–EcoRI fragment of pGAG418 (Sohn et al., 1999) bearing H. polymorpha glyceraldehyde-3-phosphate dehydrogenase promoter (PGAPDH). Prior to ligation, the EcoRI- and BspHI-generated termini of the insert and vector, respectively, were treated with Mung Bean Nuclease. None of the E. coli transformants obtained with the ligation mix possessed the correctly assembled fragments, indicating that the resulting selectable marker might not be efficiently expressed in E. coli. The plasmid pKNR80, which was finally selected at this step, possessed deletions at the junction of PGAPDH and APH extending for 44 bp towards the APH ORF and 12 bp towards PGAPDH, which apparently led to an improvement of this region for APH expression in E. coli. The XbaI–SpeI fragment of pKNR80 was then deleted to obtain pKAM444. To construct the pKAM540 plasmid for HpMAL1 disruption (Figure 2), the H. polymorpha chromosomal DNA was digested with Ecl136II, self-ligated and used as a template for PCR with the primers U (5′-cattc aacga gactg agcac-3′) and L (5′-aggaa ttaat attgt caaga ggg-3′). Then, the obtained PCR product was inserted between the PciI and PvuII sites of pKAM444. To obtain pKAM554, the pKNR80 plasmid was cleaved with SpeI and NcoI, treated with BAL31 and self-ligated (Figure 1). pKAM555 (Figure 3) was constructed by replacing the PciI–CfrI fragment in pKAM554 with the DNA fragment obtained by annealing of the oligonucleotides 5′-catga catgg cgcca tggga agctt ggatc cccgg gtacc gagct cgaat tcact-3′ and 5′-ggcca gtgaa ttcga gctcg gtacc cgggg atcca agctt cccat ggcgc catgt-3′. To obtain the pKAM567A and pKAM567B plasmids (Figure 3), a linker obtained by annealing of the oligonucleotides 5′-gatcc aagct tccca tggcg ccatg tcatg agtggc-3′ and 5′-gccact catga catgg cgcca tggga agctt ggatc-3′ was introduced into the PvuII site of the pKAM554 plasmid in two opposite orientations. The pKAM556 plasmid (Figure 3) was constructed by replacing the KpnI–PvuII fragment of pKAM555 with the HARS6-bearing KpnI–EcoRV fragment of the p2CHA6 plasmid (Chechenova et al., 2004). To obtain the pKAM558 plasmid, the HpLEU2-containing SalI–BamHI fragment of the pCLHX plasmid (Sohn et al., 1996) was introduced between the XhoI and BamHI sites of pKAM556. The pESC1 plasmid was constructed by insertion of the 248 bp EcoRV–StuI fragment of the S. cerevisiae URA3 gene into the PvuII site of pKAM444.

Figure 1.

Construction scheme for the pKNR80, pKAM444 and pKAM554 plasmids. Sequences originating from the pUK21 plasmid are designated by the thick line; the DNA fragments of the PGAPDH chromosomal locus are designated by the open bars; position of lacZ promoter, PGAPDH, and APH ORF are designated by dashed, dotted and solid arrows, respectively; position of E. coli replication origin is indicated by the thin line. Restriction sites that are not unique in the plasmid are underlined. Numbers 23 and 194 under the dotted lines indicate the size (bp) of BAL31 deletions obtained during the construction of the pKAM554 plasmid. Vertical dotted lines indicate sequence correspondence in parent and descendant molecules

Figure 2.

Construction scheme for the HpMAL1 disruption cassette. (I) H. polymorpha chromosomal DNA was digested with Ecl136II; (II) restriction fragments were self-ligated; and (III) were used as a template for PCR with primers U and L; (IV) PCR product was inserted into the pKAM444 plasmid; (V) the resulting pKAM540 plasmid was linearized with Ecl136II and used for H. polymorpha transformation. Recombination of the linearized plasmid with the chromosomal HpMAL1 gene resulted in replacement of its internal sequence with the sequence of pKAM444 plasmid. Double lines, H. polymorpha chromosomal sequences; single lines, pKAM444 sequence; empty bar, selectable marker; E, Ecl136II restriction sites; U and L, primers. The dashed frames mark sequences of the HpMAL1 ORF

Figure 3.

Schemes of the pKAM555, pKAM556, pKAM567A and pKAM567B plasmids. Positions of polylinkers on the schemes are indicated by solid bars; their nucleotide sequences are presented below the schemes. The pKAM567B polylinker sequence is identical to that of pKAM567A but has the reverse orientation. The T residues representing 3′ overhangs after digestion of the pKAM567A with XcmI are shown in bold

Results and discussion

pKAM444 cloning vector

The original aim of this work was to develop a convenient vector with a G418 resistance marker for H. polymorpha transformation. The marker, which was previously constructed for this purpose (Sohn et al., 1999), consisted of the PGAPDH-controlled Tn903 ORF encoding aminoglycoside 3′-phosphotransferase (APH). Plasmids equipped with the bacterial ampicillin resistance marker and containing the aforementioned PGAPDHAPH gene provided E. coli transformants selected on ampicillin-containing medium with kanamycin resistance, but the primary selection of transformants on kanamycin plates was not possible (data not shown). Another disadvantage of this marker was the presence of the HindIII, SmaI and XhoI restriction sites within its ORF. Notably, all these restriction sites had been removed from the kanamycin resistance marker in the pUK21 cloning vector (Vieira and Messing, 1990).

To overcome the disadvantages of the previous marker, the pKAM444 plasmid was constructed (Figure 1). This plasmid possessed pUK21 APH ORF fused to PGAPDH for its efficient expression in yeast. Expression in E. coli was ensured by the lacZ promoter (data not shown), located upstream of PGAPDH, and by a favourable sequence at the junction of PGAPDH and APH ORF randomly selected during the plasmid construction (see Materials and methods).

To confirm the ability of pKAM444 to confer G418 resistance in yeast, H. polymorpha DL1-L and P. pastoris GS155 were transformed with this plasmid, which was linearized by PvuII prior to transformation. In H. polymorpha, linearization of transforming DNA stimulates its integration into random genomic sites (van Dijk et al., 2000) and one could expect a similar effect in the related yeast, P. pastoris. Indeed, transformations of both species produced a number of G418-resistant clones. The presence of the plasmid sequence in transformants was confirmed by PCR analysis (data not shown). Efficient transformation of P. pastoris with the plasmid that did not contain any its DNA sequences indicates the random integration of the plasmid.

In contrast to H. polymorpha and P. pastoris, integration of transforming DNA into the S. cerevisiae genome occurs mainly via homologous recombination. To study whether the obtained marker could be used for transformation of S. cerevisiae, the integrative vector pESC1 was constructed by insertion of an internal fragment of the S. cerevisiae URA3 ORF into the pKAM444 plasmid. Integration of the pESC1 plasmid by a single cross-over into the chromosomal URA3 locus should generate two defective copies of this gene. To stimulate recombination, prior to the transformation this plasmid was cleaved by ScaI within the URA3 sequence. All transformants of the 10B-H49 strain obtained on G418 plates were unable to grow on uracil omission medium, indicating that the chromosomal URA3 gene was disrupted.

Application of the pKAM444 plasmid for gene replacement

To demonstrate applicability of the obtained select- able marker for gene disruption in H. polymorpha, plasmid pKAM540 (see Materials and methods), bearing the disruption cassette for the maltase-encoding gene HpMAL1 (Liiv et al., 2001), was constructed. The routine approach for the disruption of chromosomal genes by replacement with a selectable marker includes transformation of cells with a disruption cassette—a linear DNA fragment containing a selectable marker flanked by sequences of a target locus. The classical procedure for obtaining such gene disruption cassettes consists of two cloning steps: (a) cloning of a target locus sequence in an E. coli vector; and (b) replacement of the internal part of the gene being disrupted with a selectable marker. Prior to yeast transformation, the disruption cassette is excised from the plasmid by restriction enzymes. Here, an alternative approach including only one cloning step was used to construct the HpMAL1 disruption cassette (Figure 2). In this approach, the PCR product obtained by reverse amplification of the flanking (RAF) homology sequences is cloned into a vector equipped with a yeast selectable marker. The entire plasmid then serves as a disruption cassette. This allows the use of a single selectable marker for transformation of both E. coli and yeast. To disrupt the HpMAL1 gene, the strain DL1-L was transformed with the linearized pKAM540 plasmid (Figure 2V). Approximately 60% of the H. polymorpha transformants selected on G418-containing medium were unable to grow on a medium containing maltose as a sole carbon source, indicating disruption of the MAL1 gene.

pKAM554, pKAM555, pKAM556, pKAM567A and pKAM567B vectors

The fragment of the PGAPDH locus in the pKAM444 plasmid begins from the −578 position, counting from the GAPDH ORF (Sohn et al., 1999), and contains several restriction sites, which could be undesirable for further use of the plasmid as a cloning vector. It has been previously shown that even the 146 bp PGAPDH fragment lacking the restriction sites mentioned above provides APH expression at levels sufficient for high resistance of the H. polymorpha transformants to G418 (Sohn et al., 1999). Based on this, the region bearing the undesirable restriction sites was removed during construction of the pKAM554 plasmid (Figure 1).

Like pKAM444, the pKAM554 plasmid possessed only PciI, CfrI and PvuII sites suitable for cloning of DNA fragments. To make this vector more convenient, a linker possessing several other restriction sites was introduced between the PciI and CfrI sites of pKAM554 (Figure 1) and the obtained plasmid was designated pKAM555 (Figure 3). The pKAM556 plasmid (Figure 3) was obtained from pKAM555 by the insertion of HARS6 for autonomous replication in H. polymorpha cells. Importantly, two additional unique restriction sites (ApaI and XhoI) were introduced into the polylinker together with HASR6. Along with the restriction sites in the polylinker, AgeI and NdeI within HARS6 (Figure 3) could also be used as cloning sites if functional HARS6 is not required.

The growth rates of E. coli transformants with the obtained plasmids on kanamycin-containing medium and plasmid DNA yields from equal volumes of overnight cultures were even increased compared to that of transformants with the original vector pUK21 (data not shown), indicating that plasmid maintenance in E. coli cells was not compromised.

According to our experience, insertion of certain DNA fragments (e.g. the XhoII–XhoII fragment of the HpLEU2 locus) into the pUK21 polylinker inhibits growth of corresponding bacterial transformants on kanamycin-containing medium. In contrast, insertion of this HpLEU2-containing fragment into the polylinker of pKAM556 did not presently affect growth of E. coli transformants or the plasmid DNA yield (data not shown).

The pKAM556 plasmid was equipped with a HARS to allow its use as an episomal H. polymorpha vector. To confirm the ability of this plasmid to replicate autonomously, the H. polymorpha DL1-L strain (leu2) was transformed with the pKAM556-based pKAM558 plasmid bearing the HpLEU2 marker, which was required to facilitate monitoring of plasmid mitotic stability. Transformation frequencies with this plasmid upon selection for G418 resistance and for leucin prototrophy were the same. The Leu+ phenotype of nine tested G418-resistant transformants was mitotically unstable, indicating the episomal state of the plasmid. Along with the pKAM555 and pKAM556 plasmids, two additional derivatives of the pKAM554 plasmid, pKAM567A and pKAM567B, possessing another set of cloning sites (Figure 3), were constructed. Cleavage of these plasmids by XcmI at two sites in the polylinker produces T overhangs at the 3′ ends of the vector (Figure 3), which allows the cloning of Taq-generated PCR products (Testori et al., 1994).

Importantly, the pKAM554 plasmid and its derivatives contained a significantly shorter PGAPDH fragment than that in pKAM444. The ability of the shortened promoter to ensure APH expression in S. cerevisiae at a level sufficient for primary selection of transformants was confirmed (data not shown).

Concluding remarks

In this study a selectable marker that allows primary selection of transformants in E. coli and yeast was developed and used to construct the pKAM444, pKAM554, pKAM555, pKAM556, pKAM567A and pKAM567B vectors (Figures 2, 3) bearing different sets of cloning sites. The specially designed XcmI sites (Testori et al., 1994) in the pKAM567A and pKAM567B vectors allow their use for cloning of PCR products with A overhangs at 3′ ends. In contrast to the other vectors, pKAM556 contains a HARS for plasmid maintenance in H. polymorpha in the episomal state. The pKAM444 and pKAM554 plasmids contain only three cloning sites (CfrI, PvuII and PciI) that can be useful when an excess of restriction sites is undesirable, e.g. for construction of plasmids destined for the integration into a host genome after linearization within the insert.

Importantly, the HpMAL1 disruption cassette was constructed by the RAF-based approach, presuming integration of the entire plasmid into the target locus via double crossing-over (Figure 2). This approach has certain advantages over the classical approach, since it includes only one cloning step and allows the use of a single selectable marker for plasmid maintenance in E. coli and gene replacement in yeast.

In terms of convenience, RAF-based construction of cassettes for gene replacement can compete with purely PCR-based methods (reviewed by Wendland, 2003) if long flanking homology sequences are required for gene disruption (Wach, 1996; Gonzalez et al., 1999). Like the PCR-based methods, the RAF approach does not depend on restriction sites within the target gene. This is useful for the construction of precise junctions of the flanking homology sequences with the internal part of a disruption cassette. Moreover, this method allows replacement of a target locus with a long heterologous sequence (e.g. a large selectable marker or an expression cassette), which is difficult to amplify by PCR. Described plasmids are available from M. Agaphonov (moagaphonov@cardio.ru) for research purposes upon request.


This work was supported by Grant No. 09-04-01261 of the Russian Foundation for Basic Research.