Modified Cre-loxP Recombination in Aspergillus oryzae by Direct Introduction of Cre Recombinase for Marker Gene Rescue

Abstract

Marker rescue is an important molecular technique that enables sequential gene deletions. The Cre-loxP recombination system has been used for marker gene rescue in various organisms, including aspergilli. However, this system requires many time-consuming steps, including construction of a Cre expression plasmid, introduction of the plasmid, and Cre expression in the transformant. To circumvent this laborious process, we investigated a method wherein Cre could be directly introduced into Aspergillus oryzae protoplasts on carrier DNA such as a fragment or plasmid. In this study, we define the carrier DNA (Cre carrier) as a carrier for the Cre enzyme. A mixture of commercial Cre and nucleic acids (e.g., pUG6 plasmid) was introduced into A. oryzae protoplasts using a modified protoplast-polyethylene glycol method, resulting in the deletion of a selectable marker gene flanked by loxP sites. By using this method, we readily constructed a marker gene-rescued strain lacking ligD to optimize homologous recombination. Furthermore, we succeeded in integrative recombination at a loxP site in A. oryzae. Thus, we developed a simple method to use the Cre-loxP recombination system in A. oryzae by direct introduction of Cre into protoplasts using DNA as a carrier for the enzyme.

INTRODUCTION

Aspergillus oryzae is used extensively in the manufacture of fermented foods and commercial enzymes for food processing (16, 19, 30, 44). The complete genome sequence of this fungus is known (23, 29). The A. oryzae genome (38 Mb) contains 12,074 genes and is significantly larger than that of Aspergillus fumigatus or Aspergillus nidulans (11, 34). The functions of many of the extra genes are unknown or poorly characterized. Gene function studies often rely on methods such as gene targeting to create deletions.

The bacteriophage P1 bipartite Cre-loxP recombination system is a simple two-component system currently recognized as a powerful DNA recombination tool (26). When the Cre-loxP system was used to rescue marker genes in organisms, including Aspergillus sp., Cre was generally expressed intracellularly. In most cases, many time-consuming steps, including the construction of a Cre expression plasmid, introduction of the plasmid, and Cre expression in the transformant, are required. To circumvent such a laborious process, a method of eliminating marker genes by unselected transient transfection with a Cre expression plasmid in Epichlöe festucae, Neotyphodium sp., and A. nidulans was developed by Florea et al. (8). Furthermore, in the site-specific FLP-FRT, β-Rec-six systems, and Cre-loxP recombination systems, a flipper cassette carrying the specific sites and the recombinase gene together with a resistance marker was constructed by Kopke et al. and Hartmann et al. (17, 24). This cassette, which can regulate the expression of the recombinase gene, enables one-step marker excision.

We investigated whether Cre could be directly introduced into A. oryzae cells for excision of a marker. Nucleic acids, such as a fragment or plasmid, were found to act as carriers of Cre for direct introduction. It has been reported that cultured animal cells will take up Cre recombinase that has been fused with a basic peptide and that this enables recombination at loxP sites in the genome (36). This simple method required examining for fusion of the optimal basic peptide for Cre in host cells. In comparison, our simple method has the advantage that commercially available Cre can be used. In this study, we describe a simple marker rescue method using the Cre-loxP system with the direct introduction of Cre using a Cre carrier in A. oryzae. We constructed a marker-free A. oryzae strain lacking ligD for optimized homologous recombination. In addition, we attempted integrative recombination at a loxP site in vivo with direct introduction of Cre.

MATERIALS AND METHODS

Strains, media, and molecular biology techniques.

Standard Escherichia coli manipulations were performed as described previously (38). E. coli strain DH5α (Nippon Gene Co., Ltd., Tokyo, Japan) was used for plasmid propagation. Standard yeast genetic manipulations were performed as described by Adams et al. (1). Saccharomyces cerevisiae strain BY4741 (MATa his3Δ leu2Δ met15Δ ura3Δ) was used for in vivo plasmid construction. A. oryzae genomic DNA was isolated as described previously (6). A. oryzae NS4 (43) carrying the double selectable markers niaD and sC, derived from RIB40 (National Research Institute of Brewing Stock Culture and ATCC 42149), which was used for the genome-sequencing project (29), was the recipient strain for construction of the loxP::sC/NS4 mutant strain. The A. oryzae ligD disruptant (ΔligD::ptrA), derived from NS4, was prepared as previously described (31). These strains were grown in complete YPD medium (1% yeast extract, 2% polypeptone, 2% glucose) or in CDME medium (Czapek-Dox [CD] minimal medium [10] supplemented with 30 μg/ml l-methionine, 2 mM magnesium chloride, and 70 mM monosodium glutamate instead of magnesium sulfate and sodium nitrate as the sulfur, magnesium, and nitrogen sources, respectively) for the preparation of conidial suspensions. CDE medium (CD medium supplemented with 70 mM monosodium glutamate instead of sodium nitrate as the nitrogen source), CDM medium (CD medium supplemented with 30 μg/ml l-methionine and 2 mM magnesium chloride instead of magnesium sulfate as the sulfur and magnesium sources), and CDME medium were used in auxotrophy tests of the loxP/NS4 and ΔligD::loxP mutant strains. CDE medium was also used as the selection medium for ligD knockout derivatives of A. oryzae. CDSe medium (CDE medium supplemented with 30 μg/ml d-methionine and 2 mM magnesium chloride instead of magnesium sulfate as the sulfur and magnesium sources and 0.2 mM Na₂SeO₄ for selenate) was used as the selection medium for the sC marker excision event in A. oryzae.

Construction of the loxP::sC/NS4 mutant strain.

The plasmid pUGsCniaD carrying the loxP-A. nidulans sC-loxP cassette and A. oryzae niaD, was used to construct the loxP::sC/NS4 mutant strain, which was the target strain for marker rescue. pUGsCniaD was constructed as follows with all primers described in Table S1 in the supplemental material. An SphI site was introduced into pUG6 (15) using a QuikChange site-directed mutagenesis kit (Stratagene, Santa Clara, CA) with the primers QCSphIFw and QCSphIRv (see Table S1 in the supplemental material), resulting in pUG6Sp. The sC marker cassette was prepared from pUSC (43) digested with XbaI and SphI. The digested sC fragment was ligated into the XbaI and SphI sites of pUG6Sp, resulting in pUG6sC. Next, the niaD marker cassette was prepared from pNGA142 (41) digested with AvrII and SpeI. The pUG6sC plasmid was digested with SpeI, dephosphorylated with E. coli alkaline phosphatase (Toyobo Co., Ltd., Osaka, Japan), and ligated with the digested niaD fragment, resulting in pUGsCniaD.

A. oryzae NS4 was transformed with pUGsCniaD digested with BamHI as previously described (13) (see Fig. S1A in the supplemental material). A. oryzae transformants were screened for sulfate and nitrate prototrophy and purified by subculturing at least three times on CD agar plates. These candidates were subjected to PCR using primer sets 1 (loxPsC1Fw and loxPsC1Rv) and 2 (loxPsC2Fw and loxPsC2Rv) (see Table S1 in the supplemental material) with A. oryzae genomic DNA as the template. When digested pUGsCniaD was inserted into the targeted niaD locus, a 1-kb fragment was amplified using primer set 1 and a 3.7-kb fragment was amplified using primer set 2. The amplification of a 3-kb fragment with primer set 2 indicated that pUGsCniaD had been inserted into an ectopic locus or that the candidate was still heterokaryotic. A single correct homologous integration resulting in the insertion of digested pUGsCniaD into the resident niaD gene was confirmed by Southern blotting. The niaD probe used for hybridizations was obtained by PCR with primers niaDp-Fw and niaDp-Rv (see Table S1 in the supplemental material) using the pUGsCniaD plasmid as the template.

Direct introduction of Cre into loxP::sC/NS4 cells.

Protoplasts of loxP::sC/NS4 cells were prepared using a cocktail of enzymes as described previously (32). The protoplasts were washed three times in solution B (1.2 M sorbitol, 50 mM CaCl₂, 10 mM Tris-HCl [pH 7.5]) and adjusted to 1 × 10⁸ cells/100 μl in solution B. The protoplasts (100 μl) were mixed well with 12.5 μl of solution C (50% polyethylene glycol [PEG] 4000, 50 mM CaCl₂, 10 mM Tris-HCl [pH 7.5]), 10 μl of Cre (Clontech Laboratories, Inc., Mountain View, CA), and 3 μg of pUG6 plasmid. The facilitated protoplast-PEG transformation method (13) was performed as follows. The mixed sample was placed on ice for 30 min and transferred to a 50-ml tube, to which 1 ml of solution C and subsequently 2 ml of solution B were added. Forty microliters of soft agar (0.5% agar, 45°C) was poured into the tube containing the suspension and mixed gently. The mixed sample was immediately overlaid on several selenate plates containing 0.8 M NaCl. The plates were incubated at 30°C for 4 to 7 days.

Colony PCR in A. oryzae.

Colony PCR was performed using the modified Thomson and Henry method (42). The template was prepared as follows. Small amounts of mycelia and spores from the candidates were suspended in 50 μl buffer A (10 mM EDTA, 1 M KCl, 100 mM Tris-HCl [pH 9.5]) and vortexed for 3 min. The samples were heated twice in a microwave oven for 30 s, vortexed for 3 min, and placed at 4°C until use. The PCR amplification utilized 1 μl of sample supernatant as the template, specific primers, and the PCR amplification enzyme kit KOD FX (Toyobo), which is based on a novel KOD DNA polymerase from Thermococcus kodakaraensis KOD1 (40). Amplification consisted of 35 cycles with a temperature profile specified by the manufacturer.

Construction of the ΔligD::loxPsC mutant strain and marker rescue.

The plasmid ΔligD::loxPsC/pYes2, carrying the loxP-A. nidulans sC-loxP cassette used for ligD disruptions, was constructed according to the methods of Oldenburg et al. (35) and Colot et al. (7). The 5′ and 3′ fragments of ligD were obtained by PCR with primers 5ligDloxPFw and 5ligDloxPRv and primers 3ligDloxPFw and 3ligDloxPRv (see Table S1 in the supplemental material) with A. oryzae NS4 genomic DNA as the template. The 5ligDloxPFw primer incorporated an AatII site (underlined in Table S1 in the supplemental material) to mutate an initiation codon of ligD. The loxP-sC-loxP cassette was prepared from pUG6sC digested with PshBI. The yeast vector pYES2 (Invitrogen Co., Tokyo, Japan) was digested with EcoRI and BamHI. These four DNA fragments were assembled in S. cerevisiae using the endogenous homologous recombination system, resulting in ΔligD::loxPsC/pYES2.

A. oryzae ΔligD::ptrA (31) was transformed with ΔligD::loxPsC/pYES2 digested with BamHI and MluI as previously described (13). A. oryzae transformants in which the selectable marker sC flanked by loxP sites in the same orientation replaced ptrA were screened for sulfate prototrophy and purified by subculturing at least three times on CDE agar plates. These candidates were subjected to colony PCR using primer set 1 (ligDloxPsCFw and ligDloxPsCRv; see Table S1 in the supplemental material) with A. oryzae genomic DNA as the template. Colony PCR was performed as described above. A single correct homologous integration resulting in replacement with ΔligD::loxPsC was confirmed by Southern blotting. The probe used for hybridization was obtained by PCR with primers 3ligDloxPFw and 3ligDloxPRv (see Table S1 in the supplemental material) using the ΔligD::loxPsC/pYes2 plasmid as the template.

To rescue the sC marker gene in the ΔligD::loxPsC mutant strain, ΔligD::loxPsC protoplasts (1 × 10⁷ cells/100 μl of solution B) were mixed well with 12.5 μl of solution C, 5 μl of Cre (1,000 U/ml; New England BioLabs Inc., MA), and 3 μg of pUSC (44). The subsequent steps were as described above.

In vivo loxP targeting using direct introduction of Cre.

Protoplasts were prepared from a loxP/NS4 mutant strain from which the sC marker had been rescued. The protoplasts (1 × 10⁷ cells/100 μl of solution B) were mixed well with 12.5 μl of solution C, 5 μl of Cre (1,000 U/ml; New England BioLabs), and 10 μg of pUG6sC. Subsequent steps were as described above.

Candidates with the selectable marker sC introduced into the loxP site were screened for sulfate prototrophy and confirmed by colony PCR using primers confirm-Fw1 and confirm-Rv2 (see Table S1 in the supplemental material). The candidates were purified by subculturing at least three times on CDE agar plates and confirmed by Southern blotting. The niaD probe used for hybridizations was obtained as described above.

RESULTS

Generation of the loxP::sC/NS4 mutant strain.

To investigate whether Cre could be directly introduced into A. oryzae cells and can function in the nucleus, we constructed the loxP::sC/NS4 mutant strain by using plasmid pUGsCniaD harboring the A. nidulans sC marker gene between two loxP sites in the same orientation and the A. oryzae niaD marker (see Fig. S1A in the supplemental material). pUGsCniaD digested with BamHI was inserted into the niaD locus of the A. oryzae NS4 strain. The loxP::sC/NS4 transformants were screened for nitrate and sulfate assimilation and identified by PCR using primer sets 1 (loxPsC1Fw and loxPsC1Rv) and 2 (loxPsC2Fw and loxPsC2Rv). The loxP::sC/NS4 candidate was further confirmed by Southern blotting (see Fig. S1B in the supplemental material), which demonstrated the successful integration of the loxP-sC-loxP cassette at the resident niaD locus.

Selectable marker rescue by direct introduction of Cre into A. oryzae cells.

To introduce Cre into loxP::sC/NS4 cells, we prepared protoplasts of the loxP::sC/NS4 mutant strain. We tried to introduce Cre alone or along with a Cre carrier consisting of the pUG6 plasmid containing loxP sites. Because the sC gene encoding ATP sulfurylase is a bidirectionally selectable gene conferring sulfate assimilation and its absence results in resistance to selenate (4), the protoplasts were inoculated onto selenate plates to screen for candidate strains with the sC marker rescue.

Using Cre with pUG6, we obtained more than 200 candidates. In contrast, using Cre alone, we obtained only a few candidates. The rescued sC marker candidates were confirmed by PCR using the appropriate primer set (Fig. 1A and B). The amplification of a 0.8-kb fragment indicated that the sC marker was rescued from the loxP::sC/NS4 mutant strain (Fig. 1B), whereas amplification of a 4.1-kb fragment indicated that the sC marker was not rescued (Fig. 1A). The sC marker gene rescue using Cre and pUG6 was observed in the candidates shown in lanes 5 to 9 (Fig. 1C). These transformants were designated the loxP/NS4 mutant strain. These results indicated that Cre could be directly introduced into A. oryzae using nucleic acids such as pUG6. These results also suggested that pUG6 harboring loxP sites acted as a carrier of Cre.

Fig 1.

Confirmation of sC marker rescue from the loxP::sC/NS4 mutant strain. (A) Schematic representation of the inserted loxP-sC-loxP cassette in the niaD locus of the loxP::sC/NS4 mutant strain. (B) Schematic representation after sC marker rescue from the loxP::sC/NS4 mutant strain. The arrows indicate the primers used for colony PCR to confirm the sC marker rescue. (C) Agarose gel electrophoresis of PCR-amplified DNA fragments from the sC marker region. Candidates transformed with Cre alone are shown in lanes 1 to 4. Candidates transformed with Cre and pUG6 are shown in lanes 5 to 9. A negative control is shown in lane 10.

To confirm whether the sC marker gene was eliminated from the loxP/NS4 mutant strain, we examined sulfate assimilation by the loxP/NS4 mutant strain. The strain did not show sulfate assimilation using sulfur as the source (Fig. 2). The result indicated that the sC marker gene was excised by Cre-mediated recombination and that the resulting loxP/NS4 mutant strain was transformable using the sC marker again.

Fig 2.

Auxotrophy of A. oryzae loxP/NS4 mutant strain. Wild-type (NS4), loxP::sC/NS4, and loxP/NS4 cells (1 × 10³) were inoculated on each plate and incubated at 30°C for 4 days. The media were CD (A), CDM (B), and CDME (C), as described in Materials and Methods. The inoculated position of each cell type is shown in panel D.

Effects of nucleic acids species as Cre carriers on Cre-loxP-mediated marker rescue.

We investigated which type of nucleic acids could act as a Cre carrier (Table 1). The loxP::sC/NS4 mutant strain was used as the host. Elimination of the sC marker gene from loxP::sC/NS4 was confirmed by colony PCR as described in Materials and Methods. The results indicated that plasmids, with or without the loxP sites, could act as Cre carriers (pUG6 and pUC18 in Table 1). DNA fragments or single-stranded DNA could also act as Cre carriers (pUG6 fragment, pUC18 fragment, and single-stranded DNA in Table 1). However, the results suggested that short DNAs could not act as Cre carriers (oligonucleotide DNA in Table 1). Although we obtained more than 200 selenate-resistant colonies using Cre with plasmids, DNA fragments, and single-strand DNA, we also obtained a few colonies using Cre with oligonucleotide DNA, Cre alone, or a pUG6 plasmid alone (Table 1). The selenate-resistant colonies obtained by Cre with oligonucleotide DNA, Cre alone, or a pUG6 plasmid alone were thought to be due to mutations in the sC gene or other mutations that can confer selenate resistance.

Table 1.

Cre carrier examination for Cre introduction

Cre carrier	No. of colonies with sC marker removed	No. of selenate-resistant colonies
		Tested	Obtained
pUG6 plasmid	14	14	>200
pUG6 fragmenta	14	14	>200
pUC18 plasmid	9	9	>200
pUC18 fragmentb	10	10	>200
Single-stranded DNAc	13	13	>200
Oligonucleotide DNAd	0	8	14
Nonee	0	10	14
pUG6 plasmidf	0	13	13

^aObtained by treatment with restriction enzymes XhoI and XbaI.

^bObtained by treatment with restriction enzyme XbaI.

^cFrom salmon sperm.

^d34-bp loxP sequence.

^eOnly Cre enzyme.

^fOnly Cre carrier.

Markerless ligD gene disruption in A. oryzae.

For the application of the Cre-loxP system using direct introduction of Cre, we attempted to construct a ligD disruptant from which the marker gene was rescued. ligD is involved in the final step of DNA nonhomologous end joining (20). The deletion of ligD from A. oryzae greatly increases the efficiency of gene targeting compared with that in the wild type (31). To construct an unmarked ligD disruptant, we first constructed the ΔligD::loxPsC mutant strain from the ΔligD::ptrA mutant strain (31). The ΔligD::loxPsC mutant strain harbored an sC marker gene between two loxP sites in the same orientation (Fig. 3A). The sC marker gene was rescued using the Cre-loxP system with direct introduction of Cre using the pUSC plasmid (43) as the Cre carrier. Because more than 200 candidates were obtained in the previous experiment, we used only 1/10 of the number of protoplasts for sC marker rescue from the ΔligD::loxPsC mutant strain.

Fig 3.

Generation of a marker-free ΔligD::loxP mutant strain. (A) Strategy for the construction of a ligD disruptant (ΔligD::loxP) from which the sC marker gene was rescued by the direct introduction of Cre. The black bars indicate the hybridization positions of the probe used to confirm sC marker rescue by Southern blotting. Csp45I restriction sites are indicated by the letter C. (B) Southern blotting of the genomic DNA of transformants. Each lane contained 20 μg of restriction enzyme-digested genomic DNA of the parent strain (ΔligD::ptrA), the ΔligD::loxPsC mutant strain, and ΔligD::loxP transformants 1, 2, and 3 cut with Csp45I. (C) Auxotrophy of the A. oryzae ΔligD::loxP mutant strain. ΔligD::ptrA, ΔligD::loxPsC, and ΔligD::loxP mutant cells (1 × 10³) were inoculated onto CDE, CDME, and CDME plates supplemented with 0.1 μg/ml pyrithiamine and incubated at 30°C for 4 days. The inoculated positions are shown in the lower right part of the panel.

The ΔligD::loxP marker rescued transformants were screened by selenate selection and the colony PCR. The ΔligD::loxP candidates were confirmed by Southern blotting. The analysis revealed the expected hybridization signals at 3.9 kb, 7.3 kb, and 5.6 kb in digested genomic DNA isolated from ΔligD::loxP candidates 1 to 3 and the ΔligD::loxPsC and ΔligD::ptrA mutant strains, respectively (Fig. 3B). These results showed the successful rescue of the sC marker gene from the ΔligD::loxPsC mutant strain.

To determine whether the sC and ptrA marker genes function in the ΔligD::loxP mutant strain, we examined the assimilation of sulfate and resistance to pyrithiamine in the ΔligD::loxP mutant strain. The ΔligD::loxP mutant strain could not assimilate sulfate as a sulfur source and showed pyrithiamine sensitivity (Fig. 3C). The result indicates that the sC and ptrA marker genes are available for the ΔligD::loxP mutant strain.

In vivo loxP targeting using direct introduction of Cre.

Cre catalyzes not only the loxP-mediated excision event but also the insertion event, although the excision reaction is kinetically favored over the insertion reaction (39). Cre-loxP site-specific insertion has been used to construct a plasmid in vitro (28). Therefore, we examined whether the Cre-loxP system with direct introduction of Cre would function for loxP-mediated insertion in A. oryzae (Fig. 4A). We used the pUG6sC plasmid, including the loxP sites and sC marker gene, and the loxP/NS4 mutant strain as the host, with the pUG6sC plasmid also functioning as the Cre carrier. Candidates with the sC gene inserted into the loxP site were confirmed by colony PCR using the appropriate primer set (Fig. 4B). Amplification of a 4.1-kb fragment indicated that the sC marker had been inserted into the loxP site of the loxP/NS4 genome, whereas amplification of a 0.8-kb fragment indicated that the sC marker had not been inserted. Insertion of the sC marker gene was observed in some candidates (lanes 1, 2, 3, 5, 9, and 10 in Fig. 4B) using pUG6sC and Cre (Fig. 4B). These candidates were subcultured at least three times on CD agar plates to obtain homokaryotic strains and designated loxP::sC/loxP/NS4. The loxP::sC/loxP/NS4 mutant strains were further confirmed by Southern blotting (Fig. 4C). The analysis revealed the expected hybridization signals at 7.8 and 5.8 kb in digested genomic DNAs isolated from the loxP::sC/loxP/NS4 mutant strains. On the other contrary, 5.8- and 5.9-kb bands and a 4.4-kb band were detected in digested genomic DNAs isolated from the loxP/NS4 and NS4 strains, respectively. These results suggest that the sC gene was correctly inserted into the loxP site using direct introduction of Cre.

Fig 4.

Generation of the A. oryzae loxP::sC/loxP/NS4 mutant strain from the loxP/NS4 mutant by direct introduction of Cre with pUG6sC. (A) Strategy for in vivo loxP targeting using direct introduction of Cre with pUG6sC. The arrows indicate the colony PCR primers used to confirm the integration of the sC marker cassette into the loxP site of the loxP/NS4 mutant strain. The black bars indicate the hybridization positions of the niaD probe used to confirm loxP targeting by Southern blotting. HindIII restriction sites are indicated by the letter H. (B) Agarose gel electrophoresis of PCR-amplified DNA fragments indicating the integration of the sC marker cassette. Candidates from the Cre and pUG6sC experiments are shown in lanes 1 to 12. Candidates treated with pUG6sC alone are shown in lanes 13 to 19. (C) Southern blotting of the genomic DNA from transformants. Each lane contained 20 μg of HindIII-digested genomic DNA of the loxP/NS4 mutant strain (lane 1), the loxP::sC/loxP/NS4 mutant strain (lanes 2, 3, and 4), or wild-type strain NS4 (lane 5).

DISCUSSION

We developed a simple marker rescue method using the Cre-loxP system with direct introduction of Cre into the cells using a Cre carrier. This simple method consists of two primary steps (Fig. 5). The first step is to introduce a commercial Cre with nucleic acid as the Cre carrier into protoplasts of the target strain harboring a marker gene between two loxP sites oriented in the same direction. The second step is to screen for marker rescue strains using selection plates. Compared with the conventional method (9, 25), our simple method reduces the number of steps in the Cre-loxP system (Fig. 5).

Fig 5.

Schematic comparison of the conventional Cre-loxP recombination method with direct introduction of Cre. The left panel shows the technical steps for a previously reported marker rescue in A. fumigatus (34). The right panel shows our method of marker rescue by direct introduction of Cre.

Previously reported methods, which use unselected transfection with a Cre expression plasmid and a flipper cassette carrying the specific sites and the recombinase gene together with a resistance marker, were developed for convenient marker gene rescue with recombinases in filamentous fungi (8, 17, 24). These methods successfully circumvent the laborious conventional Cre-loxP recombination method. Our method of marker gene rescue has two major advantages over the above-described approach. First, it is not necessary to construct a Cre expression cassette and optimize the induced expression of Cre. This advantage may be effective for fungi in which the optimal controllable promoter for Cre expression does not exist, if our method is applicable in these fungi as well. In A. oryzae, to our knowledge, the optimal controllable promoters for gene expression have not been reported. Second, the selectable markers are eliminated from the target strain with high efficiency. As shown in Fig. 1 and Table 1, we have confirmed the elimination of the marker gene in many randomly selected transformants. Thus, our method may serve as the simple Cre-loxP system method similar to the above-mentioned methods (8, 17, 24).

One problem with using DNA as a Cre carrier for Cre is that it may integrate into the genome of the target strain by nonhomologous recombination. We propose the following solutions to this problem. First, the amount of DNA used as a Cre carrier should be less than one-third of that used for general transformation in aspergilli. Second, using the same gene for marker rescue and as the Cre carrier can minimize the possibility of nonhomologous recombination because when the Cre carrier is integrated into the target strain genome by nonhomologous recombination, the transformant is unable to grow on the selection medium. Therefore, we have used the pUSC plasmid harboring the sC marker gene as the Cre carrier in the sC marker gene rescue of the ΔligD::loxPsC mutant strain (Fig. 3). Technically, Southern analysis would be needed for cases where it is important to ensure that nonhomologous integration of a part of the Cre carrier had not occurred. In fact, we confirmed that the nonhomologous integration of the Cre carrier had not occurred in the ΔligD::loxP mutant strain by Southern analysis using the sC and vector probes, which were prepared from pUSC (data not shown).

We have shown that pUG6, which possesses loxP sequences, as well as other plasmids, DNA fragments, and single-stranded DNA, can act as a Cre carrier (Table 1; Fig. 3). Initially, we predicted that the DNA used as a Cre carrier should have loxP sequences because Cre binds loxP sites. However, we have found that even Cre carriers without loxP sequences can carry Cre into the cells. The structure of Cre has been clarified by Gopaul et al. (14). Because Cre appears to bind to DNA easily, it may bind other DNA sequences in addition to the two 13-bp recombinase-binding elements arranged as inverted repeats in the loxP site (18). Our hypothesis may be supported by the loxP mutant analysis data reported by Lee et al. (27).

The mechanism by which the Cre carrier facilitates introduction of the Cre enzyme is unknown. Using fluorescence microscopy and scanning electron microscopy of S. cerevisiae, Murata et al. recently reported that DNA targeted for transformation enters to the cell through endocytotic membrane invagination (37). Moreover, PEG mediates the binding of negatively charged DNA to the negatively charged cell surface (5, 12, 21). From these reports, we propose the following hypotheses: (i) Cre binds to DNA (the Cre carrier), (ii) the DNA bound to Cre connects with the membranes of the target strain protoplasts through the mediation of PEG, and (iii) Cre and DNA enter the cell through endocytotic membrane invagination. Therefore, a Cre carrier is required to introduce Cre into the cell. However, even in S. cerevisiae, how transforming DNA reaches the nucleus and enters through the nuclear pore is unknown (22). Therefore, elucidation of the mechanism of Cre introduction requires further investigation.

We have demonstrated integrative recombination into a loxP site in vivo with the direct introduction of Cre, as well as marker gene rescue (Fig. 5). loxP targeting following marker gene rescue has not been applied in other marker rescue methods, such as direct repeat recombination (33). However, the efficiency of loxP targeting was lower than that of marker gene rescue. We hypothesize that this result is due to the extremely weak insertion activity of Cre compared with its excision activity (39). Therefore, loxP-targeting efficiency may be improved by using mutant loxP that is specific for the insertion reaction (2, 3). However, ku and lig4 disruptants, which improve the efficiency of gene targeting, may restrict the use of loxP targeting because these disruptants allowed accurate and free integration of the target gene. Nevertheless, the application may be useful for fungi of which ku and lig4 disruptants have not yet been obtained.

In conclusion, we have developed a simple marker rescue method using the Cre-loxP system with direct introduction of Cre into cells using a Cre carrier. The ΔligD mutant strain, from which the sC marker gene was rescued, was conveniently generated. To date, some effective and simple methods of marker gene rescue have been developed using site-specific recombinases such as Cre and FLP (8, 17, 24). We believe that our simple method provides a rapid and efficient approach for targeted gene excision in filamentous fungi, including A. oryzae.