Introduction

Yarrowia lipolytica is one of the most studied unconventional yeasts and is considered to be a generally recognized as safe (GRAS) organism, which has already been used for some biotechnological applications, including citric acid production and heterologous protein expression (Bankar et al. 2009; Beckerich et al.1998) and eicosapentaenoic acid production (Xue et al. 2013). Y. lipolytica can accumulate large amount of lipids either by de novo lipid synthesis or via the uptake of lipids and alkanes, and is considered to be a suitable host to produce valuable, lipid-derived compounds (Nicaud 2012). Among the oleaginous yeasts, Y. lipolytica is the only one for which genomic and efficient molecular genetic tools are currently available (Beopoulos et al. 2009). However, metabolic engineering of this strain is still time-consuming for sequential gene integrations (Celinska and Grajek 2013; Chuang et al. 2010; Xue et al. 2013). The assembly of large, recombinant DNAs encoding entire biochemical pathways represents a significant challenge. A DNA assembler, which allows the assembly of multiple fragments in a single step via in vivo homologous recombination, has been successfully applied in Saccharomyces cerevisiae (Gibson 2009; Shao et al. 2009), Xanthophyllomyces dendrorhous (Contreras et al. 2013) and Kluyveromyces marxianus (Heo et al. 2013). However, in vivo homologous recombination has not been used to assemble multiple-gene biochemical pathways in a one-step fashion in a chromosome of Y. lipolytica.

As a proof of concept, we report, for the first time, the use of a DNA assembler method to rapidly assemble β-carotene biosynthesis pathways, with sizes of approx. 11 kb, as well as their integration into a chromosome in Y. lipolytica.

Materials and methods

Strains, plasmids, reagents and medium

Strains and plasmids used in this study are listed in Supplementary Table 1. Yarrowia lipolytica ATCC 201249 was cultivated on yeast/peptone/dextrose (YPD) medium (10 g yeast extract/l, 20 g tryptone/l, and 20 g glucose/l). Synthetic complete minus uracil (SC-ura) medium (20 g glucose/l, 1.7 g yeast nitrogen base without amino acids/l, 5 g (NH4)2SO4/l, supplemented with 50 mg leucine and 50 mg lysine/l) was used to select integrants containing the assembled biochemical pathways. For plates, agar (2 %) was added. Growth was at 30 °C.

DNA manipulation

One set of DNAs, comprising the carB and carRP genes from Mucor circinelloides (Velayos et al. 2000a,b), and the GGS1 gene (Matthäus et al. 2013) and the selective marker ura3 (accession number: AJ306421.1) from the Y. lipolytica NRRL Y-1095 genome, was synthesized de novo (Crolla and Kennedy 2001; Wang et al. 2013). The other set comprised the crtE, crtI and crtYB genes from X. dendrorhous ATCC24202 (Verwaal et al. 2007), and also the selective marker ura3. Y. lipolytica promoters (TEF1p, EXP1p, FBAp and GPDp), terminators (xpr2t, mig1t and lip2t) and homologous fragments for integration into target regions (rDNAu and rDNAd) were obtained individually from the genomic DNA of Y. lipolytica strain NRRL Y-1095, and the terminator cyc1t was obtained from S. cerevisiae strain BY4742 (Blazeck et al. 2011). All primers used in this study are listed in Supplementary Table 3. The fragments were assembled by overlap extension PCR (OE-PCR) into the individual gene expression cassettes, and DNA mixtures of the individual cassettes were prepared for transformation using a previously described procedure (Shao et al. 2009). The PCRs were performed in 50 μl, as described in the instructions for the KOD-Plus-Neo kit (Toyobo, Japan).

Yeast transformation and genomic DNA extraction

Transformation of Y. lipolytica was performed using the Zymogen Frozen-EZ yeast transformation kit II (Zymo Research Corporation) (Blazeck et al. 2011). Genomic DNA (gDNA) was extracted from Y. lipolytica using the Axygen genomic DNA purification kit (Axygen, China).

Genotype confirmation

PCR analysis, restriction digestion of amplified fragments and Real–time qPCR were used to do genotype confirmation.

Confirmation of the correct DNA assembly was assessed by PCR amplification of each gene cassette from the genomic DNAs. The PCR products obtained were separated in agarose gels and purified. They were then subjected to digestion by the corresponding restriction endonucleases, and the correct restriction digest pattern was assessed via electrophoresis on 1 % agarose gels.

Real-time qPCR amplification and analysis were performed using a MyiQ2 Two Colour Real-Time PCR Detection System with iQ5 optical system software version 2.1 (Bio-Rad, USA). The host strain ATCC201249 was selected as the negative control. The GGS1, promoter EXP1 and GPD1 driven heterologous carB and carRP, and the residual ura3 fragment in the genome were chosen as reference copy (primers listed in Supplementary Table 3).

Functional analysis of the assembled pathways

Colonies of yeast harboring correctly assembled DNAs were picked into 4 ml YPD liquid medium in 20 ml test tubes and incubated overnight at 30 °C, with shaking at 230 rpm. These were used to inoculate 20 ml fresh media in 150 ml flasks. Cells were grown at 30 °C, with shaking at 230 rpm, for 96 h. Cells that were harvested and dried at 80 °C for 16 h were used to calculate dry cell weight (DCW). Cells were harvested by centrifugation at 4,000×g for 4 min, resuspended in 0.5 ml dimethyl sulfoxide, and incubated at 55 °C for 10 min and then at 45 °C for 15 min, after which an equal volume of acetone was added. Samples were then centrifuged at 13,000×g for 5 min and the supernatants containing β-carotene were transferred to a new tube. Cell extracts were analyzed by HPLC, as described by Shao et al. (2009). A sample (5 μl) was loaded onto an Agilent Diamonsil SB-C18 column and analyzed at 472 nm after elution with acetonitrile (A) and tetrahydrofuran (B). The elution program is described in Supplementary Table 3.

Results and discussion

Design of the Y. lipolytica DNA assembler

As proof of concept, we designed and constructed two sets of multiple-genes: β-carotene biosynthetic pathways, named as MC or XD, as shown in Fig. 1a and Supplementary Fig. 1. The accumulation of β-carotene resulting from the successful assembly and integration of the biosynthetic pathways turns transformants orange or red, which enables their identification. An rDNA sequence was previously used as the target locus for integration in the chromosome of Y. lipolytica (Nicaud 2012), and there are more than 200 rDNA repeat units in various chromosomes of Y. lipolytica, as well as δ sites in S. cerevisiae, that are available for integration (Heo et al. 2013; Verbeke et al. 2013). Compared to S. cerevisiae, in which the efficiency of homologous integration of exogenous DNA is very high, even with short (50 bp) homologous DNA flanking sequences, homologous recombination in Y. lipolytica only occurs at acceptable rates (80 %) with 0.5–1 kb 5′- and 3′-flanking regions of the target sequence (Verbeke et al. 2013). An rDNA site was selected as the target locus for integration into the chromosome, and ~0.6 kb 5′- and 3′-flanking regions were used as homologous flanking sequences. The two sets of β-carotene biosynthetic pathways (~11 kb) were divided into four cassettes, namely rDNAu-TEF1p-GGS1-xpr2t, EXP1p-carB-mig1t, GPDp-carRP-lip2t, and URA3-rDNAd, MC pathway or rDNAu-TEF1p-crtE-xpr2t, GPDp-crtI-lip2t, FBAp-crtYB-lip2t, and URA3-rDNAd, XD pathway, with overlaps between two successive cassettes of ~65 bp in MC or ~40 bp in XD.

Fig. 1
figure 1

Construction of a heterologous β-carotene biosynthetic pathway from Mucor circinelloides in Y. lipolytica, namely MC a The assembled β-carotene biosynthesis pathway with its corresponding verification primers b transformants on the selection plate c PCR analysis of the assembled, chromosomal β-carotene biosynthesis pathway using primers C1-F/C1-R, C2-F/C2-R, C3-F/C3-R, and C4-F/C4-Ralso demonstrated in Table S3 d Physical characterization of PCR analysis through restriction digestion. C1 was digested with PstI and XbaI; the correct clones should exhibit six bands, with sizes of 195, 242, 300, 330, 624 and 767 bp. C2 was digested with HindIII/NdeI; the correct clones should exhibit three bands, with sizes of 660, 1093 and 1896 bp. C3 was digested with HindIII/SacI; the correct clones should exhibit five bands, with sizes of 443, 500, 630, 837 and 1105 bp. C4 was digested with SmaI/SalI/PstI; the correct clones should exhibit four bands, with sizes of 347, 551, 708 and 1159 bp e HPLC analysis of the cell extracts from CIBTS1176 carrying the combined pathway. The host strain ATCC201249 was used as a negative control

Full size image

Confirmation of the assembled β-carotene biosynthetic pathways

Prepared DNA mixtures were transformed into Y. lipolytica strain ATCC 201249. Around 92 colonies or 240 colonies appeared after 60 h on SC-ura medium for the MC or XD pathways, respectively. One-third of the MC and one-sixth of the XD transformants were orange/red (Table 1). Ten red colonies and 10 white colonies were randomly picked for amplification using four set of primers outside or inside the assembled fragment using junction PCRs as indicated in Fig. 1a and Supplementary Table 3, and the resulting 4 bands have overlaps of the assembling overlap region. the correct assemblies exhibited all four expected bands obtained following PCR (Fig. 1c and Supplementary Fig. 1c), and these four fragments were further confirmed by PstI/XbaI, HindIII/NdeI, HindIII/SacI, and SmaI/SalI/PstI digestion, as indicated in Fig. 1d. Of the MC or XD transformants, 60 or 70 %, respectively, were correctly assembled, while none of the white colonies were correct (Table 1). The total efficiency of one-step assembly of these four-gene pathways was 20.9 % for MC and 11.7 % for XD (Table 1). Higher efficiency can be further improved by extending the length of the overlap region between two cassettes as previously reported in S. cerevisiae (Shao et al. 2009), or deletion of KU70 in Y. lipolytica host (Verbeke et al. 2013). However, results of qPCR (supplementary information) indicate that CIBTS1176 assembled gene from MC which showed the deepest color underwent unexpected additional integrations of partial cassettes, including GGS1, carRP, and ura3 into the genome (supplementary information). This was caused by the high rate of non-homologous end-joining (NHEJ) in Y. lipolitica (Kretzschmar et al. 2013).

Table 1 Probability of obtaining the correct colony after integration of multiple genes into Y. lipolytica using the DNA assembler
Full size table

From the HPLC results, a peak appeared at 11.2 min from the correctly assembled strains, consistent with the elution time of authentic β-carotene, whereas no such peak was observed for the cell extracts from the wild-type strain ATCC 201209 (Supplementary Figs. 1e, d). Among all the engineered strains, the highest producing strain, CIBTS1176, yielded 2.22 ± 0.34 mg of β-carotene per g DCW of MC, while all the XD transformants produced less than 1 mg of β-carotene/g DCW. As an example, strain CIBTS1187, shown in Supplementary Fig. 1d, produced 0.84 ± 0.05 mg β-carotene/g DCW. Additionally, to our knowledge, this was the first report of β-carotene production in engineered strains of Y. lipolytica, and the production can be further optimized with additional metabolic engineering tools or our DNA assembler.

While sequential gene integration in Y. lipolytica requires n weeks for n genes, as described previously (Celinska and Grajek 2013; Matthäus et al. 2013), our finding enables the integration of multiple genes (four genes of ~11 kb) within one week. Furthermore, the method demonstrated in this study was also used to transform multiple genes into other Y. lipolytica strains, including ATCC 76861 and ATCC MYA2613 (data not shown). Thus, it seems that simultaneous gene integration is an efficient method for multiple-gene transformations in Y. lipolytica. Thus, this work contributes to the development of genetic tools that allow efficient genomic modifications in Yarrowia.