Introduction

Candida tropicalis is a diploid yeast which does not undergo a sexual reproductive stage (Blandin et al. 2000). In recent years, growing attention has been devoted to C. tropicalis especially for its use in chemical and fuel production. Moreover, unique physiological features such as the possibility to use C. tropicalis for the production of long-chain α,ω-dicarboxylic acids (DCA) (Liu et al. 2004; Picataggio et al. 1992) and xylitol (Ahmad et al. 2013; Guo et al. 2013; Misra et al. 2013) have enhanced the academic and commercial value of this diploid yeast.

In yeast molecular biology, gene disruption plays an important role in physiological studies and gene function research. Having an appropriate selection marker is a key element, being necessary for the isolation of a gene-disrupted mutant. An integrative transformation system for C. tropicalis that depends on a uracil-auxotrophic mutant was reported for the first time by Haas et al. (1990). In this system, both alleles of the POX4 and POX5 genes, which encode distinct isozymes of acyl coenzyme A involved in DCA metabolism, were disrupted in a C. tropicalis strain, enhancing the yield and efficiency of DCA production (Picataggio et al. 1992). It is noteworthy that it was difficult and time consuming to isolate ura3 mutant and re-use the URA3 gene as a selection marker. In Saccharomyces cerevisiae, a cassette was developed through the introduction of direct repeats of Salmonella hisG DNA segments at both ends of the URA3 marker for the re-utilization of the selection marker (Alani et al. 1987). After cassette integration into the yeast genome, the URA3 gene was deleted by spontaneous intra-chromosomal recombination between the hisG fragment repeats. This resulted in a uracil-auxotrophic mutant that can be selected on plates containing 5-fluoroorotic acid (5-FOA), and the marker gene can be reused repeatedly. A similar approach was applied to gene disruption in C. tropicalis strains (Ahmad et al. 2013), where direct repeats of hisG sequences were replaced by other DNA sequences (glu, leu, trp, or arg) isolated from the Bacillus subtilis genome.

A mutated hygromycin resistance gene has also been used as a selection marker for C. tropicalis transformation (Hara et al. 2001). However, we had difficulty in repeating these results with our yeast host (unpublished data). Recently, a SAT1 flipper method for gene disruption developed in C. albicans (Reuß et al. 2004) was directly applied to C. tropicalis MYA3404 (Chen et al. 2014). The SAT1 flipper cassette contains a dominant nourseothricin resistance marker (CaSAT1) for the selection of integrative transformants, and the CaSAT1 marker was subsequently deleted based on the FLP-FRT recombination system with a C. albicans-adapted FLP gene. However, there are still several issues associated with the use of antibiotic-resistant genes as selection markers. One of the most common problems is that resistance genes often used as selection markers in other yeast transformation are ill suited to C. tropicalis as, for example, C. tropicalis is highly resistant to G418 (Hara et al. 2000). Moreover, codon usage bias is common in C. tropicalis strains. For example, the CTG codon, which would be mistranslated to serine instead of leucine in C. tropicalis (Santos et al. 2011; Ueda et al. 1994), adds difficulty in the genetic manipulation of C. tropicalis. A conventional Ura-blaster that contains direct repeats of hisG enables re-use of the marker gene, as already reported. However, the hisG repeats resulted in URA3 pop-out via homologous annealing during PCR (Wilson et al. 1999, 2000). In addition, the hisG repeats led to difficulties in the construction of the large disruption cassette because of low amplification efficiency of large DNA fragments.

In this study, we developed an efficient gene disruption strategy for C. tropicalis, in which the selection marker can be easily re-utilized. Using this method, two C. tropicalis carnitine acetyltransferase encoding genes (CAT) were sequentially deleted. Furthermore, we demonstrated that this technology can integrate and functionally express an exogenous gene into the chromosome at the location of the deleted gene.

Materials and methods

Strains, media, and culture conditions

C. tropicalis strains used in this study are listed in Table 1. The diploid strain of C. tropicalis ATCC 20336 was maintained on yeast peptone dextrose (YPD) medium (10 g L−1 yeast extract, 20 g L−1 glucose, 20 g L−1 peptone). C. tropicalis XZX, a derivative of C. tropicalis ATCC 20336 generated by random mutagenesis with N-methyl-N′-nitro-N-nitrosoguanidine (NTG), was used as a host strain for transformation. The mutagenesis process is described below. For genetic manipulation of C. tropicalis XZX, minimal medium (MM, 0.67 g L−1 yeast nitrogen base (YNB), 20 g L−1 glucose, 10 g L−1 [(NH4)2SO4]), supplemented medium (SM, MM supplemented with uracil at a concentration of 0.06 g L−1) and FOA-SM (SM supplemented with 5-fluoroorotic acid at a concentration of 2 g L−1) were used. All C. tropicalis strains were routinely cultured at 30 °C. Escherichia coli strain JM109 (TaKaRa, Dalian, China) was used as a host for all cloning experiments and was cultured in LB medium (5 g L−1 yeast extract, 10 g L−1 tryptone, 10 g L−1 NaCl) at 37 °C. Ampicillin was supplemented at 100 μg mL−1. When necessary, agar was added to a final concentration of 20 g L−1 for solid media.

Table 1 C. tropicalis strains used in this study
Full size table

Auxotrophic mutant isolation

The uracil-auxotrophic mutant was derived from C. tropicalis ATCC 20336 by chemical mutagenesis with slight modifications to a double selection method (Haas et al. 1990). A colony of C. tropicalis ATCC 20336 was inoculated in YPD medium and incubated at 30 °C and 200 rpm to prepare log growth phase cells. The resulting cells were washed once with sterile H2O and re-suspended for 40 min in 10-mL YPD medium containing 1–4 mg NTG. After NTG treatment, the culture was washed twice with 50-mL PBS buffer (NaCl 8 g L−1, KCl 0.2 g L−1, Na2HPO4 1.42 g L−1, KH2PO4 0.27 g L−1, pH 7.4) and re-suspended in 50-mL YPD medium for 3 h at 30 °C. Then, cells were washed twice with MM and incubated in 50 mL MM for 6–8 h. For enrichment of auxotrophs, nystatin was added to a concentration of 50 μg mL−1 and cells were incubated with shaking for 2 h. Cells were then washed twice with 50 mL sterile PBS buffer and incubated in YPD medium for 10 h. The culture was then spread onto FOA-SM plates, and colonies that appeared 2 to 3 days later were transferred to MM and SM plates by replica plating. Colonies that re-appeared on SM agar plates, but not on MM plates, exhibited uracil auxotrophy. The URA3 genes of uracil auxotrophs were isolated and sequenced.

Construction of Tm-gda-URA3 and Tm-URA3-gda

A series of plasmids Tm-gda-URA3 containing different sized gda fragments, were constructed as follows. The 1.6 kb URA3 gene to be used as marker was amplified by PCR from the genomic DNA of C. tropicalis ATCC 20336 using primers URAU and URAR based on the URA3 sequence (GenBank accession number AB006207). The PCR product was ligated into the pMD 18-T (TaKaRa, Dalian, China) to generate plasmid Tm-URA3. gda fragments from the 3′ terminus of the URA3 gene were amplified by PCR with primers Ugda143 and Dgda (yielding a 143 bp product), Ugda245 and Dgda (yielding a 245 bp product), Ugda324 and Dgda (yielding a 324 bp product), and Ugda488 and Dgda (yielding a 488 bp product). A 325 bp segment (gda325) located before the terminator codon (TAA) in URA3 was also amplified with primers Ugda325 and Dgda325. These fragments were then ligated to the 5′ end of the URA3 gene as follows: all the gda fragments were digested with PstI and SalI and inserted into Tm-URA3 at the corresponding sites. The resulting plasmids were named Tm-gda143-URA3, Tm-gda245-URA3, Tm-gda324-URA3, Tm-gda488-URA3, and Tm-gda325-URA3. The primers used in this study are listed in Table 2.

Table 2 Primers used in this study
Full size table

In addition, a 305 bp gda segment (gda305) located in the promoter region of URA3 and a 302 bp gda segment (gda302) located after the initiation codon (ATG) of URA3 were amplified from the URA3 gene by PCR with primers Ugda305 and Dgda305, Ugda302 and Dgda302, respectively. The gda305 and gda302 fragments were ligated to the 3′ end of the URA3 gene by digestion with XbaI and EcoRI and inserted into Tm-URA3, yielding plasmids Tm-URA3-gda305 and Tm-URA3-gda302.

Construction of Ts-CAT1-gda-URA3 and Ts-CAT1-URA3-gda

The plasmid Ts-CAT1-gda-URA3 containing the cassette for CAT allele 1 disruption was constructed as follows. Based on the C. tropicalis ATCC 20336 CAT gene sequence (GenBank accession number D84549.1), a 1881 bp CAT fragment was amplified from C. tropicalis XZX genomic DNA by PCR using primers CATU and CATR (Table 2), which was termed CAT1. The CAT1 fragment was ligated into the pMD 18-T Simple vector (TaKaRa, Dalian, China) to yield plasmid Ts-CAT1. The primers rCATU and rCATR (Table 2) were then used to eliminate the middle region (approximately 1.2 kb) by inverse PCR. The PCR product (CAT1-Ts-CAT1) was then digested with PstI and XbaI and ligated to various gda-URA3 cassettes, yielding plasmids Ts-CAT1-gda143-URA3, Ts-CAT1-gda245-URA3, Ts-CAT1-gda324-URA3, Ts-CAT1-gda488-URA3, and Ts-CAT1-gda325-URA3.

The URA3-gda305 fragment was separated from plasmid Tm-URA3-gda305 by digestion with PstI and EcoRI, and the PCR product CAT1-Ts-CAT1 was digested with PstI and XbaI. The cloning process included two steps: firstly, the EcoRI and XbaI sticky ends were made blunt by pfu DNA polymerase (TaKaRa, Dalian, China); secondly, URA3-gda305 and CAT1-Ts-CAT1 fragments having the same cohesive end and a blunt end were ligated together by T4 DNA ligase. The resulting plasmid was termed Ts-CAT1-URA3-gda305. A similar strategy was used for construction of the plasmid Ts-CAT1-URA3-gda302.

Disruption cassettes for the first CAT allele were isolated from plasmids Ts-CAT1-gda-URA3 or Ts-CAT1-URA3-gda by PCR using primers CATU and CATR. The resulting PCR products of CAT1-gda-URA3-CAT1 and CAT1-URA3-gda-CAT1 were purified using the Axyprep PCR cleanup kit (Axygen, Suzhou, China).

Construction of Ts-CAT1-hisG-URA3-hisG

The plasmid Ts-CAT1-hisG-URA3-hisG was constructed as described previously (Ko et al. 2006) with slight modifications. Two 1.1 kb hisG fragments were amplified by PCR from the plasmid pCUB6 (Lu et al. 2008) using HisG-F1 and HisG-R1, HisG-F2 and HisG-R2 as primers. The hisG1 fragment was digested with EcoRI and BamHI, and inserted into Tm-URA3 digested with EcoRI and BamHI, yielding the plasmid Tm-URA3-hisG1. Then, the hisG2 fragment was inserted into the plasmid between the PstI and SalI sites. The resulting plasmid was named Tm-hisG2-URA3-hisG1. Finally, the hisG2-URA3-hisG1 fragment was flanked by homologous regions as in the construction of Ts-CAT1-URA3-gda305, yielding plasmid Ts-CAT1-hisG-URA3-hisG.

Construction of Ts-CAT2-gda324-URA3

A similar method was adopted for the construction of Ts-CAT2-gda324-URA3, a plasmid containing the cassette for the disruption of the second CAT allele. In short, a 906 bp CAT2 fragment was amplified by PCR from the genomic DNA of C. tropicalis XZX using primers CAT2ndU and CAT2ndR, which were designed from sequences inside the disrupted region of the CAT1 gene (Fig. 5a). The CAT2 fragment was inserted into the pMD 18-T Simple vector, which was linearized and ligated to the gda324-URA3 cassette to generate plasmid Ts-CAT2-gda324-URA3. A 2.6 kb disruption cassette for the second CAT allele CAT2-gda324-URA3-CAT2 was obtained by PCR using primers CAT2ndU and CAT2ndR.

Construction of GFP expression cassette

The open reading frame (ORF) of the wild-type green fluorescent protein-encoding gene (GFP) (wtGFP) gene was amplified from plasmid pCAMBIA1302 (Cambia, Canberra, Australia) by PCR with primers GFP-U and GFP-D. The wtGFP gene was then expressed under the control of the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) promoter and terminator derived from C. tropicalis ATCC 20336 genomic DNA by PCR with the primers PgapF, PgapR, TgapF, and TgapR (Jeon et al. 2012). The promoter-wtGFP-terminator fragment was inserted into pMD 18-T Simple to obtain plasmid Ts-P-wtGFP-T.

The plasmid Ts-P-wtGFP-T was digested with XbaI and EcoRI to obtain the P-wtGFP-T cassette, which was then inserted into plasmid Tm-gda324-URA3, following digestion with the same restriction enzymes. The resulting plasmid was designated as Tm-gda324-URA3-P-wtGFP-T. Finally, the gda324-URA3-P-wtGFP-T cassette was flanked by homologous regions as in the construction of Ts-CAT2-gda324-URA3. For this construct, only the site for the restriction enzyme XbaI was changed to that of EcoRI in the rCAT2ndR primer. The resulting plasmid Ts-CAT2-gda324-URA3-P-wtGFP-T was linearized by restriction enzyme DraII and transformed into C. tropicalis 01-3 cells to generate transformant C. tropicalis 05-1. After removing the gda324-URA3 fragment in the C. tropicalis 05-1 genome, the resulting strain C. tropicalis 06-1 was obtained.

The codon-optimized GFP (GFP CTC ) gene (GenBank accession number KU763376) was chemically synthesized based on the sequence in plasmid pCAMBIA1302 in which two CTG codons were changed to optimal leucine codons (CTC). The GFP CTC gene was then digested with HindIII and SalI and inserted into plasmid Ts-P-wtGFP-T that had previously been digested with the same restriction endonucleases to replace the wtGFP gene. The resulting plasmid was named Ts-P-GFP CTC -T. Finally, the plasmid Ts-CAT2-gda324-URA3-P-GFP CTC -T was constructed as described above.

In an earlier report, C. albicans codon-adapted SAT1 and FLP genes were functionally expressed in C. tropicalis (Chen et al. 2014). Based on this observation, a 717 bp yeGFP3 gene was chemically synthesized in the light of the NCBI sequence (GenBank accession number U73901.1) where all 239 codons for amino acids were optimized according to the C. albicans codon usage bias (Cormack et al. 1997). The plasmid Ts-CAT2-gda324-URA3-P-yeGFP3-T, which contains yeGFP3 expression cassette, was constructed as described above.

Transformation of C. tropicalis

Yeast transformation was performed using lithium chloride according to a previously described method (Haas et al. 1990), with slight modifications. We used the PCR products from plasmids Ts-CAT1-gda-URA3, Ts-CAT1-URA3-gda, Ts-CAT1-hisG-URA3-hisG, Ts-CAT2-gda324-URA3, or Ts-CAT2-gda324-URA3-P-yeGFP3-T. Cells from a SM pre-culture were diluted to a density of 0.08 at 600 nm (OD600) in 50-mL fresh SM. After 7–9 h of growth to an OD600 of 1.0, the cells were washed once with 2 mL of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.4). Cells were re-suspended in 1 mL 100 mM LiCl solution, and the suspension was incubated at 30 °C and 200 rpm for 1 h. A mixture containing 88 μL of cell suspension, 2 μL salmon testis DNA (10 g L−1), and 1.5 μg transforming DNA was incubated at 30 °C for 30 min. Then, 900 μL of PEG3350 solution (40 % polyethylene glycol, 100 mM LiCl) was added to each sample, followed by mixing and incubation at 30 °C for 1 h with shaking. Samples were incubated at 42 °C for 5 min and then subjected to a cold shock step in an ice bath for 3–5 min. The cells were washed once with sterile water and spread onto MM plates. The plates were incubated at 30 °C for 2–3 days until colonies appeared.

Marker gene pop-out

Clones of individual C. tropicalis strains transformed by a gene cassette were cultured in SM at 30 °C and 200 rpm until the cell concentration reached 15 (OD600), spread onto a FOA-SM plate, and cultured at 30 °C for 3–4 days. Colonies resistant to 5-FOA were inoculated onto an SM plate and into a liquid culture to extract chromosomal DNA. PCR analysis was performed to identify excision of the URA3 marker from the genome of recombinant cells.

Preparation of crude enzyme and enzyme assay

Strains were cultured for 48 h in SM and then inoculated into SM supplemented with 2 % dodecane and cultured for 48 h for induction. Crude enzyme was extracted from C. tropicalis strains as described previously (Ueda et al. 1982). CAT activity was assayed at 30 °C by the 5,5-dithiobis(2-nitrobenzoic acid) (DTNB) method (Kawamoto et al. 1978). One unit of enzyme activity was defined as the amount of enzyme required to catalyze the formation of 1 μmol of acetyl-CoA per minute at 30 °C.

Real-time quantitative PCR

The recombinant strains were cultured in 20 mL SM. Total RNA isolation and synthesis of cDNA were performed with the Yeast RNAiso Kit and PrimeScript® RT Reagent Kit with gDNA Eraser (TaKaRa, Dalian, China) as recommended by the manufacturer. Real-time quantitative PCR (RT-qPCR) was performed to measure the relative levels of messenger RNA (mRNA) from target genes. The ACT1 gene (accession no. FM864204.1) was used as the reference housekeeping gene. For RT-qPCR, a CFX96 real-time PCR instrument with a SYBR® Premix Ex Taq™ Kit (TaKaRa) and designed primers (ACT1-F and ACT1-R, CAT-F and CAT-R, GFP-F and GFP-R, yeGFP3-F and yeGFP3-R, corresponding to ACT1, CAT, wtGFP, GFP CTC , and yeGFP3, respectively) were used. The relative quantification of target gene transcription was performed using the comparative Ct method (Silva et al. 2011). All experiments were performed in triplicate, and mean values were analyzed for target gene expression.

Fermentation and analytical methods

The fermentation experiments for long-chain α,ω-dodecanedioic acid (DCA12) production were performed in a 750-mL shaking flask with 120-mL fermentation medium at 200 rpm and 30 °C. The fermentation medium for DCA12 production consisted of 6.7 g YNB, 3 g yeast extract, 75 g glucose, 3 g (NH4)2SO4, 1 g K2HPO4, 1 g KH2PO4, and 0.06 g uracil L−1 (Picataggio et al. 1992). Yeast cells were pre-cultured in 250-mL Erlenmeyer flasks in SM until the OD600 value reached 20 (approximately 36 h). This stationary phase culture was inoculated into the fermentation medium at a final concentration of 5 % (v/v), and the pH was maintained at 5.5–6.0 by addition of 5 M NaOH or 2 M H2SO4 during the growth phase (36 h). During the DCA12 production phase, the pH was maintained at 7.5–8.0. In each run, a total of 4 % dodecane was added to each flask, in which 2 % was added twice, first at 0 and then at 72 h. During the conversion phase, glucose was added at a final concentration of 5 g L−1 every 24 h to the yeast culture. The concentration of DCA12 was determined by an HP 4890D gas chromatograph with an FID-detector (Agilent, Shanghai, China) (Picataggio et al. 1992).

Fluorescence observation

C. tropicalis 06-1, 06-2, 06-3, and XZX strains were incubated in SM overnight on a rotary shaker at 200 rpm at 30 °C and then stored at 4 °C for several hours. The culture sample was washed twice with sterile PBS buffer, re-suspended in PBS buffer, and the fluorescence intensity was observed using a Nikon Eclipse 80i microscope (Nikon Corp., Tokyo, Japan).

Results

Isolation of ura3 mutant strains

Using C. tropicalis ATCC 20336 as the parent strain, 13 auxotrophic mutant strains defective in the URA3 gene were isolated by chemical mutagenesis and a double selection procedure that combined nystatin enrichment selection and 5-FOA resistance selection. Three ura3 mutant strains were chosen and designated C. tropicalis XZW, XZX, and XZB. Mutant strains and the wild-type strain were streaked on SM plates and MM plates. After 4 days of incubation, the wild-type strain was able to grow well on both SM plates and MM plates (Fig. 1). On the other hand, the three mutant strains only exhibited growth on SM plates (Fig. 1). The URA3 alleles were amplified from the three mutant strains and sequenced. The results showed single-base mutations with amino acid substitutions in all the URA3 alleles. The residue change at Gly203 to Asp was found in all three mutant URA3 alleles, thus indicating that Gly203 might reside in a particular domain or in the active site of the URA3 protein.

Fig. 1
figure 1

Phenotypes observed on an SM plate (a) and an MM plate (b) of the parent and mutant strains. A C. tropicalis ATCC20336, W C. tropicalis XZW, X C. tropicalis XZX, B C. tropicalis XZB

Full size image

Sequential disruption of the two CAT alleles in C. tropicalis XZX

Physical maps of disruption cassettes and the disruption strategy used for the two CAT alleles are represented schematically in Fig. 2a, c. The miniature Ura-blaster cassette was constructed as described in the “Materials and methods” section. For the deletion of the allele 1 of the CAT genes, the deletion cassette CAT1-gda-URA3-CAT1 was generated by PCR using Ts-CAT1-gda-URA3 as a template and CATU and CATR as primers, and the linear DNA fragment was used to transform the URA3 mutant strain C. tropicalis XZX. Uracil prototrophic colonies were isolated and gene replacement was confirmed by PCR using the primers CATU and CATLD (Fig. 2b) and sequencing. A correct transformant in which one of two CAT genes was disrupted was designated C. tropicalis 01-3 (ura3/ura3, cat::gda324-URA3/CAT) (lane 1, Fig. 2b). After the elimination of the URA3 marker from C. tropicalis 01-3, the resulting strain was designated as C. tropicalis 02-3 (ura3/ura3, cat::gda324/CAT) (lane 2, Fig. 2b).

Fig. 2
figure 2

Sequential gene disruption of CAT in C. tropicalis. a Deletion of CAT allele 1 using the miniature disruption cassettes. The gray boxes represent the homologous arms LA (left arm) and RA (right arm). The white boxes represent the gda fragments. The black arrow indicates the URA3 gene. CAT allele 1 and CAT allele 2 are represented by the blue arrow and the red arrow, respectively. b Agarose gel electrophoresis of PCR products generated with identification primers. Lanes 1, 2, and 3 indicate PCR with primers CATU and CATLD; lane 1 represents C. tropicalis 01-3, lane 2 represents C. tropicalis 02-3, and lane 3 represents C. tropicalis XZX. Lanes 4, 5, and 6 indicate PCR with primers CAT2ndU and CATR. Lane 4 represents C. tropicalis 03, lane 5 represents C. tropicalis 04, and lane 6 represents C. tropicalis XZX. Lane M represents DNA marker. c Deletion of CAT allele 2 using the miniature disruption cassettes. (Color figure online)

Full size image

Using a similar strategy, the second CAT disruption cassette (CAT2-gda324-URA3-CAT2) was transformed into C. tropicalis 02-3. The Ura+ transformants were verified by PCR using primers CAT2ndU and CATR, and a positive transformant was isolated and designated C. tropicalis 03 (ura3/ura3, cat::gda324/cat::gda324-URA3) (lane 4, Fig. 2b). After elimination of the URA3 marker from C. tropicalis 03, the resulting strain was designated C. tropicalis 04 (ura3/ura3, cat::gda324/cat::gda324) (lane 5, Fig. 2b).

Characterization of CAT-disrupted engineered strains

When incubated in SM, C. tropicalis XZX and C. tropicalis 02-3 demonstrated a similar growth rate while the growth rate and the final biomass of C. tropicalis 04 was much lower when compared to C. tropicalis XZX and C. tropicalis 02-3 (Fig. 3a). Furthermore, C. tropicalis 04 was unable to grow using dodecane as a sole carbon source (Fig. 3b). These results indicated that the CAT gene is essential for dodecane metabolism in C. tropicalis and the pathway for glucose metabolism might also be affected by the disruption of the CAT alleles.

Fig. 3
figure 3

Physiological and biochemical confirmation of CAT disruption. a Growth curve of C. tropicalis strains in SM. b Growth curve of different C. tropicalis strains using dodecane as a sole carbon source. c Assay for CAT activities in CAT-disrupted engineered strains and the parental strain. d Transcription levels of the CAT gene in various C. tropicalis strains. e Comparison of DCA12 production for three C. tropicalis strains. The data points represent the average of three independent experiments (mean ± standard deviations)

Full size image

The CAT enzymatic activity of C. tropicalis 02-3, C. tropicalis 04, and C. tropicalis XZX was measured as reported in the “Materials and methods” section (Fig. 3c). Our results showed that the CAT activity of C. tropicalis 02-3 decreased by 80 %. In addition, there was a trace amount of CAT activity in C. tropicalis 04 (Fig. 3c). A previous study indicated that the DTNB method for the CAT assay may be affected by acetyl coenzyme A hydrolase in C. tropicalis (Zhang et al. 2005), which might lead to residual CAT activity. Therefore, RT-qPCR was conducted to further confirm the deletion of the two CAT alleles (Fig. 3d). In C. tropicalis 04, CAT mRNA could not be detected by RT-qPCR. The results indicated that the two CAT alleles were deleted successfully in C. tropicalis 04.

In addition, effects of deleting the CAT alleles on DCA12 accumulation were investigated. Fig. 3e shows that C. tropicalis XZX and 02-3 could accumulate a small amount of DCA12, whereas C. tropicalis 04 could produce higher amounts of DCA12 and a final concentration of DCA12 of 12.9 g L−1 was reached with this clone. These results indicated that the deletion of both CAT alleles of C. tropicalis can lead to significantly higher DCA12 production rates.

Effect of different gda elements on the URA3 removal efficiency

To evaluate the effect of different gda elements on the URA3 marker removal efficiency, a series of disruption cassettes for the CAT allele 1 were constructed (Fig. 4a) and transformed into C. tropicalis XZX. Transformants designated C. tropicalis 01-1, C. tropicalis 01-2, C. tropicalis 01-3, C. tropicalis 01-4, C. tropicalis 01-5, C. tropicalis 01-6, C. tropicalis 01-7, and C. tropicalis 01-8 were cultured in SM at 200 rpm and 30 °C for 36 h. The viable cell numbers were counted by spreading cultures onto YPD plates. Similarly, the same culture volumes were directly spread onto FOA-SM plates. The elimination of URA3 in clones grown on FOA-SM plates was confirmed by PCR using primers CATU and CATLD (Fig. 4b–f) and sequencing. The frequencies of URA3 gene excision of various disruption cassettes were compared and are reported in Table 3. In particular, Table 3 shows that gda143 is sufficiently long to promote URA3 gene excision. The gda324 fragment in the disruption cassette CAT1-gda-URA3-CAT1 was the most efficient in the excision of the URA3 marker when compared to the other gda fragments in other cassettes. An excision frequency of approximately 10−8 is in the same order as the long hisG fragments.

Fig. 4
figure 4

Effect of different gda fragments on URA3 marker excision. a Physical map of different disruption cassettes. The gray boxes represent the homologous arms LA (left arm) and RA (right arm). The white boxes represent the gda fragments. The black arrow indicates the URA3 gene. b PCR confirmation of URA3 marker excision in C. tropicalis 01-1 and C. tropicalis 01-2 strains. Lanes on the left of marker represent four different C. tropicalis 01-1-type samples cultured in FOA-SM; lanes on the right of the marker represent two different C. tropicalis 01-2-type samples cultured in FOA-SM. c PCR confirmation of URA3 marker excision in C. tropicalis 01-3 and C. tropicalis 01-4 strains. Lanes on the left of marker represent eight different C. tropicalis 01-3-type samples cultured in FOA-SM; lanes on the right of the marker represent six different C. tropicalis 01-4-type samples cultured in FOA-SM. d PCR confirmation of URA3 marker excision in C. tropicalis 01-6 and C. tropicalis 01-7 strains. Lanes on the left of marker represent 12 different C. tropicalis 01-6-type samples cultured in FOA-SM; lanes on the right of the marker represent 11 different C. tropicalis 01-7-type samples cultured in FOA-SM. e PCR confirmation of URA3 marker excision in C. tropicalis 01-5. Lanes 1–12 represent 12 different C. tropicalis 01-5-type samples cultured in FOA-SM. f PCR confirmation of URA3 marker excision in C. tropicalis 01-8. Lanes 1–12 represent 12 different C. tropicalis 01-8-type samples cultured in FOA-SM. Lane M indicates DNA markers. XZX represents C. tropicalis XZX. All PCR reactions were carried out using primers CATU and CATLD (locations are shown in Fig. 5a)

Full size image
Table 3 Frequency of URA3 gene pop-out based on different gda fragments
Full size table

Integrative expression of a foreign gene combined with target gene disruption

C. tropicalis 02-3 was transformed with the yeGFP3 expression cassette (Fig. 5a), which was amplified from Ts-CAT2-gda324-URA3-P-yeGFP3-T by PCR using primers CAT2ndU and CAT2ndR and was selected on the basis of uracil prototrophy. Ura+ transformants were verified by PCR using the identification primers CAT2ndU and CATR (Fig. 5b). Thus, we confirmed that both CAT alleles were destroyed and the yeGFP3 expression cassette was specifically integrated into the CAT locus. A transformant that had the expected structure was designated as C. tropicalis 05-3 (lane 1, Fig. 5b). The strain that lost URA3 was designated as C. tropicalis 06-3 (lane 2, Fig. 5b). Similarly, the recombinant strains C. tropicalis 06-1 expressing the wtGFP gene and C. tropicalis 06-2 expressing the GFP CTC gene were also constructed.

Fig. 5
figure 5

Construction of a GFP expressing strain. a Physical map of the GFP expression cassette and integration of the cassette into the CAT locus of C. tropicalis 02-3. The gray boxes represent the homologous arms. The white boxes represent the gda fragments. The black arrow indicates the URA3 gene. The green arrow represents the ORF of yeGFP3. The red arrow represents the GAPDH promoter, while the red box represents the GAPDH terminator. b Agarose gel electrophoresis of PCR products generated with identification primers. Lanes 1, 2, and 3 indicate PCR with primers CAT2ndU and CATR; lane 1 represents C. tropicalis 05-3, lane 2 represents C. tropicalis 06-3, lane 3 represents C. tropicalis XZX

Full size image

To examine the expression of different GFP genes in transformant cells, C. tropicalis 06-1, 06-2, 06-3, and the control strain XZX were cultured and the fluorescence intensity of cells was observed by using a fluorescence microscope (Fig. 6a). High levels of green fluorescence were observed in C. tropicalis 06-3, and only a slight amount of fluorescence was observed in C. tropicalis 06-2 after a long exposure time. However, no fluorescence was observed in C. tropicalis 06-1 and the control strain XZX. Thus, yeGFP3 is functionally expressed in C. tropicalis. Furthermore, RT-qPCR analysis showed that yeGFP3 mRNA levels increased significantly when compared to a transformant containing GFP CTC (Fig. 6b), while no GFP mRNA was detected in C. tropicalis 06-1.

Fig. 6
figure 6

Confirmation of different GFP gene expression in the strain C. tropicalis 02-3. a Phase contrast and UV fluorescence images of C. tropicalis XZX, 06-1, 06-2, and 06-3 strains. (b) Transcription levels of the different GFP genes in C. tropicalis; wtGFP represents a transformant with the wild-type GFP gene; GFP CTC represents a transformant with the CTG codons optimized GFP gene; and yeGFP3 represents a transformant with the yeGFP3 gene

Full size image

Discussion

C. tropicalis represents an attractive yeast strain involved in an increasing number of applications. However, it has several characteristics that hamper strain improvement through genetic engineering. A common approach for sequential gene disruption in C. tropicalis is based on the use of ura3 auxotrophic mutant hosts. In previous reports, gene disruption in C. tropicalis and C. albicans has been performed with the hisG-URA3-hisG (“Ura-blaster”) cassette, which can be re-used in successive transformations of a single strain after homologous excision of URA3 (Fonzi and Irwin 1993; Ko et al. 2006). However, the large hisG repeats can lead to difficulties in the amplification of the disruption cassette and yeast integrative transformation. Here, we developed an efficient and simple sequential gene disruption method for stable unlabeled gene insertion into C. tropicalis chromosomes. Moreover, foreign genes were successfully expressed in C. tropicalis using this method.

The disruption cassette, which consists of a functional C. tropicalis URA3 gene flanked by 0.3 kb gda direct repeats derived from regions upstream or downstream of the URA3 gene and of homologous arms to the target gene, was constructed and introduced into the yeast genome by integrative transformation. Using this strategy, a pair of CAT alleles was successfully disrupted. In addition, in further experiments, two copies of POX4 and POX5 alleles, encoding distinct isozymes of the acyl-CoA oxidase that catalyzes the first reaction in the ß-oxidation pathway, were also knocked out (data not shown). This new strategy is highly efficient and time saving as compared to the previously described process (Picataggio et al. 1991). Another advantage of the miniature Ura-blaster is that the recombination efficiency can be increased by almost one order of magnitude when compared to the traditional Ura-blaster. Overall, the results presented in this paper demonstrated that the miniature Ura-blaster can be used in a highly efficient way to destroy genes of C. tropicalis with a single marker gene.

To evaluate the effect of the gda fragment on excision of the URA3 gene from the deletion cassette, different sized gda fragments were employed and evaluated in this gene disruption system. Our results showed that 143 base pairs are enough for marker recycling, whereas gda324 with 324 base pairs has the highest excision frequency of the URA3 gene (Table 3). Moreover, gda fragments from different locations of URA3 (gda305, gda302, gda324) were cloned and investigated for their effects on the excision of URA3 marker (Table 3; Fig. 3). The frequencies of the URA3 gene excision were similar for different gda fragments isolated from the 5′ or the 3′ terminus of the URA3 gene. This indicated that the gda sequence position does not significantly affect URA3 marker removal. A traditional Ura-blaster (Alani et al. 1987) disruption cassette (CAT1-hisG-URA3-hisG-CAT1) for CAT allele 1 disruption was also constructed to be compared to the frequency of URA3 gene excision. When hisG was used as repeat fragment, the frequency of URA3 excision was similar with gda324; however, it was much lower than the rate reported previously for S. cerevisiae (10−4) (Alani et al. 1987) and C. albicans (10−5–10−6) (Fonzi and Irwin 1993). This difference might be due to the complicated genetic background of C. tropicalis.

Expression of heterologous proteins is of great academic value and of commercial significance. However, it has proven difficult to functionally express a heterologous protein in Candida species and other fungi (Gustafsson et al. 2004; Tanaka et al. 2012). In the present study, wtGFP, GFP CTC , and yeGFP3 were used as reporter genes under the control of the GAPDH promoter and integrated into the chromosome of C. tropicalis 02-3. The recombinant strains were confirmed and cultured in SM for fluorescence detection. No fluorescence was observed in transformants expressing wtGFP. When wild-type red fluorescent gene (RFP) was integrated in C. tropicalis 02-3 chromosome using a similar strategy, also no fluorescence was observed (data not shown). Moreover, mRNA levels of wtGFP could not be detected by RT-qPCR, which might be explained by a nonstop mRNA degradation mechanism (Frischmeyer et al. 2002; van Hoof et al. 2002).

For transformants expressing GFP CTC , only scant fluorescence was observed. Also, transcriptional analysis showed lower mRNA levels of the GFP CTC gene. As mentioned by Cormack et al. (1997), the CTG codons in GFP are not the only reason for the lack of expression of this gene in C. albicans. A second problematic codon may also be present in wtGFP. Therefore, we synthesized yeGFP3 in which all the codons were optimal for expression in C. albicans and integrated it into a C. tropicalis chromosome. As for other C. albicans codon-optimized genes that have been shown to function in C. tropicalis (Chen et al. 2014), the C. albicans codon-optimized yeGFP3 gene also functioned in C. tropicalis and strong fluorescence was observed in all transformed cells (Fig. 6a). Furthermore, RT-qPCR showed that the mRNA levels of yeGFP3 significantly increased when compared to GFP CTC (Fig. 6b).

Although yeGFP3 is functionally expressed in C. tropicalis, the mechanism behind codon optimization is unclear and the effect on transcription levels has not yet been elucidated. To the best of our knowledge, the mechanism behind codon optimization on mRNA levels has been reported in many microorganisms such as in E. coli (Kudla et al. 2009), Aspergillus oryzae (Tanaka et al. 2012), and S. cerevisiae (Doma and Parker 2006; Tsuboi et al. 2012). However, little is known regarding C. tropicalis. It is meaningful in the future to give insight into the mechanism of transcription and translation in C. tropicalis.

In summary, we developed an efficient sequential gene disruption strategy using a gda segment as gene disruption auxiliary sequence for C. tropicalis. The gda sequence employed in this system shortened the length of the disruption cassette and increased the efficiency of integration into the target locus. The two CAT alleles were deleted successfully using this approach. Moreover, the yeGFP3 gene was functionally expressed in C. tropicalis by integration through homologous recombination at the target locus, using this effective strategy.