Gene 540 (2014) 46–53

Contents lists available at ScienceDirect

Gene journal homepage: www.elsevier.com/locate/gene

Cloning, characterization, and heterologous expression of a novel glucosyltransferase gene from sophorolipid-producing Candida bombicola Daniel K.Y. Solaiman a,⁎, Yanhong Liu b, Robert A. Moreau c, Jonathan A. Zerkowski a a b c

Biobased and Other Animal Co-Products Research Unit, Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, USA Molecular Characterization of Foodborne Pathogens Research Unit, Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, USA Sustainable Biofuels and Co-Products Research Unit, Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, USA

a r t i c l e

i n f o

Article history: Accepted 13 February 2014 Available online 22 February 2014 Keywords: Candida bombicola gtf-1 Glucosyltransferase Sophorolipid Sterol

a b s t r a c t Candida bombicola is well-studied for the production of a biosurfactant, the sophorolipids. In this paper, the cloning of a glucosyltransferase gene using polymerase-chain-reaction (PCR) technique is described. Degenerative primer-pairs were first designed based on the highly conserved amino-acid sequences of several selected yeast glucosyltransferases. Using these primers, an amplified sequence (amplicon) of 700 base-pair from C. bombicola was obtained and subsequently sequenced. Based on the sequence of this amplicon, additional target-specific PCR primers were designed for use in subsequent rounds of 3′- and 5′-extension using DNA walking technique to eventually obtain a C. bombicola genomic sequence containing an open-reading-frame putatively identified as a glucosyltransferase (gtf-1). The gene was subcloned in Saccharomyces cerevisiae for expression and functional characterization. Quantitative RT-PCR confirmed the expression of gtf-1 in the recombinant S. cerevisiae. In vitro assay with the sonicated cells of the recombinant yeast confirms the presence of glucosylation activity on sterol and hydroxy fatty acid substrates. This study reports for the first time the cloning and characterization of a broad-specificity lipid glucosylation gene from C. bombicola, and the functional activity of its gene product. Published by Elsevier B.V.

1. Introduction Sophorolipid is a glycolipid produced and secreted by several yeast species such as Candida apicola, Candida bombicola, Rhodotorula bogoriensis, Starmerella bombicola, Wickerhamiella domericqiae, and Candida batistae (see cited references in Konishi et al., 2008; Van Bogaert et al., 2007c). The molecule contains a hydrophilic disaccharide sophorose (2-O-β-D-glucopyranosyl-β-D-glucopyranose), which may or may not be enzymatically acetylated in vivo at the 6′-(C-6′) or both C-6′ and C-6″. The hydrophobic portion of the molecule is constituted of a hydroxy fatty acid, which is glycosidically linked to the C-2′ of the sophorose moiety. In this respect, sophorolipid differentiates itself from the synthetic alkylpolyglucosides, which contain an ester bond between the hydrophobic and hydrophilic moieties of the molecule. The Abbreviations: gtf-1 and Gtf-1, glucosyltransferase gene and gene-product/enzyme, respectively; PCR and qRT-PCR, general and quantitative Reverse-Transcriptase Polymerase Chain Reaction, respectively; Km, kanamycin; Cb, carbenicillin; SC-U, Synthetic Complete minimal medium without Uracil; HPLC–MS, High Performance Liquid Chromatography– Mass Spectrometry; 17-OH oleate, 17-hydroxy oleate; Ct value, Crossover threshold value (in qRT-PCR experiments). ⁎ Corresponding author at: Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, 600 E. Mermaid Lane, Wyndmoor, PA 19038, USA. E-mail address: [email protected] (D.K.Y. Solaiman).

http://dx.doi.org/10.1016/j.gene.2014.02.029 0378-1119/Published by Elsevier B.V.

carboxyl group of the hydroxy fatty acid moiety of sophorolipid may also be found esterified in vivo to the C-4″ of the sophorose, resulting in a lactone structure of the molecule. The proportion of the various forms of sophorolipid produced from a particular fermentation depends to certain degree on the yeast strain used, the fermentation substrates, and the growth conditions. It should be noted that R. bogoriensis produces sophorolipid containing mainly a 13-hydroxydocosanoic acid (13-OH-C22) moiety, while the other sophorolipid-producing yeast strains predominantly synthesize the glycolipid having a 17hydroxyoctadecenoic acid (17-OH-C18:1) as its hydrophobic component. The structural versatility of sophorolipid affords the possibility of designing specific production system to tailor make products suitably targeted for an intended application. Sophorolipid is an important microbial product with immense potential for industrial applications (Solaiman et al., 2004a). Its amphipathic structure bestows an excellent surface-active property to the molecule. Surface-tension measurements of aqueous solution of sophorolipid routinely yielded values of 30–40 mN/m (see, for example, Solaiman et al., 2004b). The antimicrobial activity and various biomedical properties of sophorolipid have been extensively documented (Hardin et al., 2007; Kim et al., 2002, 2005; Krivobok et al., 1994; Lang et al., 1989; Shah et al., 2005). The potential applications of sophorolipid in foodindustry arena, such as its use as a formulation ingredient to modulate

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

the rheological and textural properties and as an inhibitor of biofilmformation by food pathogens, have been proposed (Nitschke and Costa, 2007). Kasture et al. (2007) recently showcased the value of sophorolipid in the field of nanotechnology by demonstrating its use in the capping of cobalt nanoparticles, thereby improving its water stability and dispersive property. The individual structural components of sophorolipid, i.e., the sophorose (Suto and Tomita, 2001) and the hydroxy fatty acid (Zerkowski and Solaiman, 2006, 2007), are also valuable specialty chemicals when separated and isolated (Rau et al., 2001). With a myriad of potential applications envisioned and the high commercial value expected of the sophorolipid, research and development efforts have largely been aimed at increasing the production yield and reducing the cost of this glycolipid. In comparison, fundamental research to delineate the metabolic pathway and the genetics of sophorolipid biosynthesis is lacking. A notable exception is a series of pioneering studies on the enzymology of sophorolipid biosynthesis of R. bogoriensis (formerly Candida bogoriensis) conducted by Esders, Light and collaborators. They first identified two glucosyl- and one acetyltransferase activities in R. bogoriensis (Esders and Light, 1972a). The two glucosyltransferase activities (i.e., glucosyltransferases 1 and 2), which resisted various chromatographic attempts to separate them, catalyzed the sequential addition of glucose units to 13-hydroxydocosanoic acid to form the final sophorolipid molecule by utilizing UDP-glucose as substrate (Breithaupt and Light, 1982; Esders and Light, 1972a). An UDP-glucose:sterol glucosyltransferase, which catalyzed the transfer of activated glucosyl group to ergosterol (an indigenous substrate) and cholesterol, was also identified in the course of their studies (Esders and Light, 1972b). These investigators further identified an acetyltransferase activity that catalyzed the transfer of an acetyl group from acetyl coenzyme A (acetyl-CoA) to a mono- or un-acetylated sophorolipid (Esders and Light, 1972a). The enzyme exhibited low reactivity toward mono-glucopyranosyl-13-hydroxydocosanote substrate. This finding, coupled with the observation that the acetyltransferase activity only peaked at a later time of fermentation, suggested that the acetylation of sophorolipid occurred mainly after the complete synthesis of the di-glucopyranosyl-13-hydroxydocosanoate. Cutler and Light (1979, 1982) reported that a low concentration of glucose substrate led to diminished glucosyltransferase activities, and a high glucose concentration was necessary for the synthesis of fatty acids having 20and 22-carbon chain length found in sophorolipid of R. bogoriensis. Although the enzymic aspect of sophorolipid biosynthesis in R. bogoriensis was well-illustrated by this earlier elegant work from Light's laboratory, the probable sophorolipid biosynthesis pathway of C. bombicola is only beginning to emerge through a series of recent molecular biological studies (Saerens et al., 2011a,b; Van Bogaert et al., 2009). In a separate effort in our laboratory to elucidate the genetic system of sophorolipid biosynthesis in C. bombicola, we report in this paper the cloning of a lipid-glucosyltransferase gene (gtf-1) from this organism. In silico analysis of its DNA sequence shows that it is structurally distinct from the other known yeast and fungal glucosyltransferases, including those identified recently in C. bombicola (Saerens et al., 2011a,b). Our work also shows that the Gtf-1 enzyme produced by a recombinant Saccharomyces cerevisiae expressing the C. bombicola gtf-1 exhibits a broad-specificity glucosyltransferase activity toward various sterols and hydroxy fatty acid. The outcome of this study adds to the knowledge-base on the glycolipid metabolism of the industrially important C. bombicola and lays down valuable foundation for metabolic manipulation to achieve commercially viable production of glycolipids. 2. Materials and methods 2.1. Organisms and growth conditions C. bombicola strains ATCC 22214 and NRRL Y-17069 were obtained from American Type Culture Collection (ATCC, Manassas, VA) and ARS Culture Collection (NCAUR, Peoria, IL), respectively. Candida cultures

47

were maintained on YM Broth (liquid) or Agar (solid) (Difco Laboratories, Detroit, MI) as appropriate, at 30 °C or less. Escherichia coli DH5α (Invitrogen, Carlsbad, CA) and α-Select (Bioline, Taunton, MA) were routinely used interchangeably for subcloning of genes and DNA (i.e., PCR fragments). E. coli cells were grown and maintained in liquid or solid (1.2–1.5% w/v of agar) Luria medium (1% w/v tryptone, 0.5% w/v yeast extract, 0.5% w/v NaCl) at 37 °C, with 250 rpm orbital-shaking for the liquid cultures. Kanamycin (Km, 35 μg/ml) and carbenicillin (Cb, 50 μg/ml) were added to the growth media as needed. 2.2. Molecular biological procedures Restriction digestion, 5′-dephosphorylation and DNA ligation reactions were performed according to the supplier's instructions using the appropriate restriction enzymes, alkaline phosphatases and T4 DNA ligase, respectively, obtained from commercial sources. Polymerase chain reactions (PCRs) were performed using commercially available thermo-tolerant DNA polymerases. PCRs were carried out in a GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA) or an Eppendorf Mastercycler Gradient PCR machine (Eppendorf, Westbury, NY) using thermal cycling programs appropriate for the specific thermotolerant DNA polymerase and the primers employed. Isolation of DNA restriction fragments or PCR products from agarose gel was carried out using commercial kits such as Zymoclean Gel DNA Recovery Kit (Zymo Research, Orange, CA) or Rapid Gel Extraction System (Marligen Biosciences, Ijamsville, MD). Plasmids in E. coli transformants were isolated by using a GenElute Miniplasmid Kit (Sigma, St. Louis, MO) or alkaline lysis method (Birnboim and Doly, 1979) for subsequent restriction analysis. The kit-purified plasmids were suitable for nucleotide-sequence determination on an Applied Biosystems 3730 DNA Analyzer or an Applied Biosystems 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA). Genomic DNA of C. bombicola was isolated using a YeaStar Genomic DNA Kit (Zymo Research). Custom oligonucleotides used as PCR and sequencing primers were purchased from Sigma Genosys (Woodlands, TX). The chromosomal DNA-walking experiments that ultimately led to the attainment of the complete sequence of the desired gene were performed using a DNA Walking SpeedUp Kit (Seegene, Rockville, MD). Computer program-based in silico sequence analyses were performed using web-based BLAST program of National Center for Biotechnology Information (Altschul et al., 1997) and a PC-based Clone Manager 9 Professional program (Scientific & Educational Software, Cary, NC). 2.3. Construction of recombinant S. cerevisiae expressing gtf-1 The coding sequence of gtf-1 was excised from a recombinant plasmid pIX-153-C by triple restriction digestion using endonucleases PmlI, SmaI and BstZ17I (all from New England Biolabs) according to the supplier's instructions. A 3.8-kb digestion fragment flanked by PmlI and BstZ17I sites and containing the gtf-1 coding sequence was purified from agarose gel following electrophoretic separation. Separately, pYES2/NT-B shuttle vector (Invitrogen) was linearized with BamHI, blunt-ended with DNA Polymerase Klenow fragment, and dephosphorylated with calf-intestinal alkaline phosphatase. This linearized vector was then ligated with the purified 3.8-kb fragment containing gtf-1 to generate recombinant plasmids pYES2/NT/gtf1 and pYES2/NT/gtf1CON (Fig. 1). The recombinant plasmids were constructed and maintained in E. coli. S. cerevisiae strain INVSc1 (Invitrogen; genotype: his3Δ1/ his3Δ1 leu2/leu2 trp1-289/trp1-289 ura3-52/ura3-52; phenotype: His−, Leu−, Trp−, Ura−) was made competent for transformation by extracellular recombinant plasmid DNA by using an S. c. EasyComp Transformation Kit (Invitrogen). The competent INVSc1 was transformed with pYES2/NT/gtf1, pYES2/NT/gtf1CON and pYES2/NT/lacZ (a control plasmid provided by Invitrogen) individually, and the positive transformant clones were selected as colonies growing on an SC-U + Glu(2%) solid (agar) medium containing (per liter) Yeast Drop-Out Minus Uracil (2 g; Sigma-Aldrich), Yeast Nitrogen Base Without Amino Acids

48

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

Bst Z17I

Sma I

pYES2/NT B

KanR

pIX-153-C gtf-1

bla

BamHI Klenow enzyme CIAP

Pml I PmlI, SmaI, BstZ17I

gtf-1 (3803 bps)

T4 DNA ligase

GAL1

GAL1 2u ori

2u ori

pYES2/NT/gtf1

gtf1

pYES2/NT/gtf1 CON

gtf1

bla

bla CYC1

CYC1

Fig. 1. Schematic of the construction of pYES2/NT/gtf1 and pYES2/NT/gtf1CON expression vectors. Complete map features for the pYES2/NT B cloning vector are available on the supplier's (Invitrogen) website. Abbreviations: gtf-1, C. bombicola glucosyltransferase gene cloned in this study; KanR, aminoglycoside 3′-phosphotransferase conferring kanamycin resistance; bla, β-lactamase conferring ampicillin/carbenicillin resistance; CIAP, calf intestinal alkaline phosphatase; 2 μ ori, origin-of-replication of yeast 2 μ-plasmid; GAL1, yeast GAL1 promoter; CYC1, CYC1 transcription termination signal.

(6.7 g; Sigma-Aldrich), D -(+)-glucose (or dextrose, 20 g; SigmaAldrich) and Bacteriological Agar (20 g; Sigma-Aldrich) according to the protocol supplied by the vendor (Invitrogen). For long-term storage, the positive clones were grown in SC-U + Glu(2%) (without agar) at 30 °C and 250 rpm for 18 h. Sterile glycerol was added to the cultures to a 15% (v/v) final concentration. The cultures were then stored in a −80 °C freezer. 2.4. qRT-PCR analysis for transcription of gtf-1 in recombinant S. cerevisiae S. cerevisiae INVSc1 [pYES2/NT/gtf1] and INVSc1 [pYES2/NT/lacZ] were first streaked on SC-U + Glu(2%) agar solid-medium plates and grown for 2 days at 30 °C. Individual colonies were picked to inoculate 15 ml of SC-U + Raf(2%) consisted of (per liter) Yeast Drop-Out Minus Uracil (2 g; Sigma-Aldrich), Yeast Nitrogen Base Without Amino Acids (6.7 g; Sigma-Aldrich) and D-(+)-raffinose (20 g; Fluka) in 50-ml Erlenmeyer flasks. One set of cultures were grown 18–20 h at 30 °C and 200 rpm, and 1-ml of these overnight (i.e., grown for ≥16 h) cultures was further used to inoculate another set of similar medium. The two sets of cultures were returned to the incubator/shaker for another 18–20 h growth under the same conditions. Induction of the expression of the cloned genes was initiated by adding 1.5 ml of 20% (w/v) sterile D-(+)-galactose (Sigma-Aldrich) solution to each culture, and the cultures were incubated under the same conditions for a further 6 h. Cells (from the 15-ml cultures) were harvested by centrifugation

in a Sorvall RC-5B Plus centrifuge (5000 ×g; 5 min; 4 °C), and were washed once with 0.5-ml Milli-Q H2O in an Eppendorf Microcentrifuge 5415R (9300 ×g; 1 min; 25 °C). Cell pellet was thoroughly resuspended in 0.5 ml of an RNA stabilization reagent (RNAlater, Qiagen) and stored at − 80 °C. Total RNA was prepared by using an RNeasy Mini Kit (Qiagen) according to manufacturer's protocol. Primers used for qRT-PCR reactions were designed using Primer3 (v.0.4.0) software (http://frodo.wi.mit.edu/). Preliminary qRT-PCR on the housekeeping genes ALG9 (coding for a mannosyltransferase activity in protein amino acid glycosylation) and UBC6 (coding for a ubiquitin-protein ligase activity involved in endoplasmic reticulumassociated protein catabolic process) using primers described by Teste et al. (2009) showed that in our study the UBC6 gene displayed the most consistent Ct values. UBC6 was thus chosen as the internal control gene for quantification. The primers (Sigma-Aldrich) used for qRT-PCR of gtf-1 were gtf-1 B FWD (5′-GGATTCTGACGATGATGAGG3′) and gtf-1 B REV (5′-GCTTGGGTCACTCGAAAATA-3′). cDNA synthesis was performed on an Applied Biosystems GeneAmp 9600 PCR System using a SuperScript III First-Strand Synthesis SuperMix Kit (Invitrogen, Carlsbad, CA) according to the manufacturers' instructions. Reactions were prepared for each RNA sample using 1 μg of DNase I-treated RNA. Reactions without reverse transcriptase were used as negative controls. qRT-PCR was carried out on a 96-well plate using a Bio-Rad iQ5 realtime PCR system (Bio-Rad, Hercules, CA) in a 50-μl reaction volume

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

consisted of 25 μl Power SYBR Green PCR Master Mix (ABI), 1.25 μl of each primer at 10 μM, 0.5 μl of cDNA, and nuclease-free water to make up the difference in volume. Thermal cycling parameters were 50 °C for 2 min for 1 cycle, an initial denaturation at 95 °C for 10 min for 1 cycle, followed by 35 cycles of 95 °C for 15 s and 60 °C for 1 min. Fluorescence data were collected at the 60 °C annealing step. The final step was a dissociation curve of 95 °C for 15 s, 60 °C for 1 min, 95 °C for 15 s, and 60 °C for 15 s. Results were visualized using the 7500 System SDS Software provided with the thermocycler. All qRT-PCR experiments were performed with two biological and two technical replicates in a 96-well plate. To evaluate genomic DNA contamination in the cDNA samples, “no amplification controls” (NAC, a minus-reverse transcriptase control) were included. In addition, three “no template controls” (NTC) containing all of the RT-PCR reagents except the RNA template were included in each reaction plate. To determine relative gene expression, the Ct value of the internal control gene (UBC6) was subtracted from the experimental (pYES2/NT/gtf1) and the non-gtf1 control (pYES2/NT/ lacZ) samples. The ΔCt, ΔΔCt, and the 2−fx values were calculated as previously described (Liu and Ream, 2008; Pfaffl, 2001). 2.5. Glucosyltransferase activity assay of recombinant S. cerevisiae An HPLC–MS-based method was developed to assay for glucosyltransferase activity in recombinant S. cerevisiae expressing C. bombicola gtf-1 gene cloned in the pYES2/NT/gtf1 plasmid construct. As control samples, S. cerevisiae transformed with pYES2/NT/gtf1CON and pYES2/NT/lacZ were included in the study. The growing and induction of S. cerevisiae cells were performed following the protocol described in the manual of pYES2/ NT Yeast Expression System (Invitrogen). Briefly, cells from long-term (−80 °C) storage stock cultures were streaked on a Yeast Extract Peptone Dextrose (YPD) agar solid medium (1% w/v yeast extract, 2% w/v peptone, 2% w/v dextrose (D-(+)-glucose), 2% w/v agar) and grown at 30 °C for 1–2 days. A single colony was picked to inoculate 5 ml of SC-U + Glu(2%) in a 25-ml Erlenmeyer flask. The culture was grown overnight at 30 °C and 200 rpm. A volume of the overnight culture, which was based on achieving a theoretical absorbance at 600 nm (A600 nm) = 0.4 in the fresh culture, was added to 50 ml of SC-U + Raf(2%) medium in a 125-ml Erlenmeyer. After an overnight incubation (30 °C, 200 rpm), A600 nm values of the cultures were recorded. Five milliliters of sterile 20% (w/v) D-(+)-galactose was added to each culture to induce expression of the cloned gene, and cells were incubated for 4 h at 30 °C, 200 rpm. Cells were then harvested by centrifugation (Sorvall RC-5B Plus centrifuge; 5000 ×g; 10 min; 4 °C), washed with 5-ml cold water (centrifugation as in the preceding step), and resuspended in a volume of a Breaking Buffer (50 mM sodium phosphate at pH 7.4; 1 mM NaEDTA; 5% v/v glycerol; 1 mM phenylmethylsulfonylfluoride (PMSF) which was added immediately before use) in 15-ml Corex centrifugation tube to give a theoretical A600 nm of 200. Cell suspension could be stored at −20 °C overnight and still retained glucosyltransferase activity. Cells were broken by sonication (Sonicator model W-385 equipped with a microtip probe; Heat-Systems, Inc., Farmingdale, NY) for 10 × 15 s (with 30 s sample cooling on ice-water bath) at maximum probetolerant power setting. The sonicated cell suspension (SCS) was used in the reaction mixtures to assay for glycosylation activity. The reaction mixture for glycosylation assay was largely based on that described by Saerens et al. (2011b). The following procedure was used: ten (10) μl of a 50 mM-lipid substrate stock solution (which was prepared in one of the following solvents: DMF (dimethylformamide), DMSO (dimethylsulfonyloxide), THF (tetrahydrofuran), ethylacetate, or ethanol) and 10 μl of a 50 mM-UDP-glucose solution (Sigma Aldrich) were added to 30 μl of 50 mM potassium phosphate buffer (pH 7.5) in a 13 × 100 mm test tube on ice. Two hundred (200) microliter of the sonicated cell suspension was then added to the mixture of the two substrates. The final reaction mixture (250 μl) was incubated at 30 °C for 3 h with frequent manual mixing (approximately every 10–15 min). Reaction was stopped by the addition of 200 μl of 2 N HCl, followed by

49

vigorous vortex mixing for approximately 30–60 s. The reaction mixture was extracted with 1–2 ml 1:1 v/v ethylether/ethylacetate mixture by vigorous vortex mixing. The tube was left undisturbed on lab bench for phase separation to occur. A Hamilton syringe was used to carefully remove the organic phase (upper layer) into a clean test tube (12 × 75 mm). The solvent was then removed by passing nitrogen stream over the solution. The dry residue was subjected to HPLC–MS analysis as described in the following paragraph. The lipid substrates tested in this study were cholesterol (Alfa Aesar, Ward Hill, MA), ergosterol (Alfa Aesar), stigmasterol (MP Biomedicals, Solon, OH), β-sitosterol (MP Biomedicals), and 17-hydroxy oleate (17-OH oleate, prepared as described by Zerkowski and Solaiman (2006)). HPLC–MS was conducted with an Agilent 1100 HPLC with Evaporative Light Scattering Detection (via an Agilent 1100 ELSD) and quadrupole mass spectrometry system. The column was a Prevail C18 3 μ column (2.1 × 150 mm, Alltech Associates, Deerfield, IL) operated at a flow rate of 0.2 ml/min. A gradient mobile phase was used with A) water and B) methanol and 0.2% acetic acid. The gradient timetable was: 0 min, 20/80, A/B; 40 min, 0/100; 50 min, 0/100; 51 min, 20/80; and 60 min 20/80. An electrospray ionization chamber was used, with the following parameters: 200–1200 m/z, fragmentor 5 V, drying gas 10, nebulizer pressure 20, drying gas temperature 300 °C, and capillary voltage 4000. In this system the lipid substrates, cholesterol and 17-OH oleate, eluted at 49.5 and 10.0 min, respectively; and the glucosyltransferase products, cholesterol glucoside and 17-OH oleate glucoside, eluted at 35.3 min and 5.8 min, respectively. 3. Results and discussion 3.1. PCR cloning and in silico characterization of C. bombicola gtf-1 The strategy of PCR cloning described previously (Solaiman, 2000; Solaiman and Ashby, 2005) was adopted for the cloning of C. bombicola gtf-1. Two primers, CL-V-114L1 (5′-TGCCATGGACTA GAACTAGAGCTTAYCCNCAYGC-3′) and CL-V-114R2 (5′-GGAATACC AGCTCTCATAGTAGCACCNGTNGTNCC-3′), were designed based on the conserved regions of the amino acid sequences of UDP-glucose:sterol glucosyltransferases of S. cerevisiae (gi6323218), P. pastoris (gi4768597) and C. albicans (gi4768599). Initially, a ~700 bp PCR product, termed VIII 88-1 fragment, was successfully obtained from C. bombicola genomic DNA using CL-V-114L1 and CL-V-114R2 primers. Using the strategy involving chromosomal DNA walking technique described earlier (Solaiman, 2000; Solaiman and Ashby, 2005), the VIII 88-1 fragment successfully served as a “bridgehead” for the eventual cloning and sequence determination of the entire 3.8-kb C. bombicola gtf-1 (GenBank Accession No. FJ231291). The C + G content of gtf-1 gene is 47.4%, which is close to the values observed with the other C. bombicola genes, such as glyceraldehyde 3′-phosphate dehydrogenase (gi:167888515; C + G = 56.7%), a putative ADP-ribosylation factor-like protein 1 (gi:167888515; 52.4%), a putative NADPH cytochrome P450 reductase (gi:126253759; 52.6%), an orotidine-5′-monophosphate decarboxylase (gi:116256774; 50.5%), and UDP-glucosyltransferases ugtA1 (gi:315057132; 49.8%) and ugtB1 (gi:315057134; 49.4%) (Van Bogaert et al., 2007a,b, 2008). BLASTP analysis of the translated sequence of gtf-1 (termed Gtf-1) detected three conserved domains. A Pleckstrin homology (PH) domain was identified at the amino-(N\) terminal; this kind of domain is usually associated with membrane-targeting of protein or specific lipid binding by enzyme. In the middle of the putative Gtf-1 is a GRAM conserved domain found in various glucosyltransferases, lipid phosphatases and some membraneassociated proteins. The most prominent feature, found in the carboxyl(C\)terminal half of the putative C. bombicola Gtf-1, is the identification of a UDP-glucoronosyl- and UDP-glucosyltransferase-associated conserved domain. The identification of these three conserved domains strongly indicates that Gtf-1 is a membrane-bound enzyme possessing lipid binding and glucosyltransferase activities, which are properties needed for sophorolipid synthesis.

50

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

BLASTP results also identified two closely matched orthologs of the Gtf-1. One is a putative Yarrowia lipolytica UDP-glucose:sterol glucosyltransferase (product YALI0D18403p of GenBank sequence gi:199425822), and the second is an experimentally verified sterol 3-β-glucosyltransferase of Y. lipolytica (gi:73619418). The Nterminal regions of the three proteins have 38–40% identities (53% positives), while the C-terminal regions have a match of 54% identities (70% positives). These results imply that the lipid or membranebinding properties, which is attributed to the N-terminal PH domain, of Gtf-1 and the Y. lipolytica proteins are only marginally similar, while their glucosyltransferase activities inherent to the Cterminal UDP-glucosyltransferase domain are even more closely linked. The BLASTP results also showed sequence matches between Gtf-1 and other glucosyltransferases, such as the Neurospora crassa UDP-glucose:sterol glucosyltransferase (gi:164429290) and sterol 3-βglucosyltransferase (gi:73619414), a hypothetical protein PGUG_04770 of Pichia guilliermondii (gi:190348249), and a putative Aspergillus nidulans sterol 3-β-glucosyltransferase (gi:67536862). In these later cases, however, only the C-terminal halves of the proteins were matched at 50–52% identities (66–68% positives). The N-terminal portions of these proteins did not show homology. Taken together, the results indicate that Gtf-1 is a glucosyltransferase that has a lipid- or membrane-binding property distantly similar to the Y. lipolytica glucosyltransferases.

3.2. Construction of S. cerevisiae clones for gtf-1 expression Once the entire nucleotide sequence of gtf-1 had been assembled via the combined PCR and genomic DNA walking techniques detailed in the preceding sections, the gtf-1 gene was then amplified from C. bombicola genome using routine PCR method, cloned into a pT7Blue-3 vector (Invitrogen) to yield pIX-153-C (Fig. 1), and maintained in E. coli. We had PCR-amplified the gtf-1 gene (3.8 kb) in sections to minimize mis-incorporation of nucleotides during the amplification process. Nucleotide-sequence determination was performed with pIX-153-C to confirm the identity of the cloned gtf-1 (GenBank Acc.# FJ231291). The complete gtf-1 gene on pIX-153-C was then excised and spliced into a S. cerevisiae expression vector pYES2/NT B to obtain pYES2/NT/gtf1 and pYES2/NT/gtf1CON as detailed in Materials and methods (M&M) section (Fig. 1). The nucleotide sequence and the orientation of gtf1 in the two plasmids, pYES2/NT/gtf1 and pYES2/NT/gtf1 CON, were experimentally verified by sequence determination. These plasmids, and pYES2/NT/ lacZ control vector, were used to transform S. cerevisiae INVSc1 as described in the M&M. All three plasmids transformed INVSc1 with a high transformation frequency, resulting in the appearance of numerous transformants on the SC-U + Glu(2%) selection agar-medium plates. In contrast, a mock transformation of INVSc1 without an added plasmid did not produce any colony on the selection plate, indicating that the clones from the plasmid-transformed S. cerevisiae were genuine recombinants containing pYES2/NT-based plasmids expressing URA3 gene needed by INVSc1 to grow on the selection medium. We randomly picked and separately grew 5 colonies (in SC-U + Glu(2%) medium) from each of the three S. cerevisiae transformants (i.e., pYES2/NT/gtf1, pYES2/ NT/gtf1CON, and pYES2/NT/lacZ) for use in the preparation of 15% v/v glycerol stock cultures for long-term storage in a −80 °C freezer.

3.3. qRT-PCR assay of gtf-1 expression in recombinant S. cerevisiae The expression of gtf-1 cloned into S. cerevisiae via the yeast vector pYES2/NT B was determined using qRT-PCR method. We used S. cerevisiae harboring the expression vector pYES2/NT/lacZ as a negative control. The Ct value used as a measurement for gene expression was calculated as described in M&M section. As shown in Table 1, the Ct value of gtf-1 in pYES2/NT/gtf1 was 15, whereas the Ct value of lacZ in pYES2/NT/lacZ was over 34. The large differences between the Ct values of the experimental sample (pYES2/NT/gtf1) and the negative control (pYES2/NT/lacZ) indicated that the gtf-1 gene was highly expressed. The Ct value of the internal control gene (UBC6; see M&M for rationale) remained constant between the two constructs, indicating that the differences in Ct between gtf-1 and lacZ were not due to the presence of unequal amounts of RNA. Results from the real-time PCR assays thus summarily demonstrated that the gtf-1 gene was highly expressed in the recombinant S. cerevisiae [pYES2/NT/gtf1]. 3.4. In vitro assay of glucosyltransferase activity in recombinant S. cerevisiae We carried out in vitro enzyme activity assay to verify the functional property of the gene-product Gtf-1 expressed in the recombinant S. cerevisiae [pYES2/NT/gtf1] as demonstrated in the qRT-PCR study. In this construct, the expression of gtf-1 is regulated by a yeast GAL1 promoter that is inducible by the addition of galactose (Fig. 1). Furthermore, the expression plasmid was constructed in such a way that a 6xHis tag is fused to the amino-terminal of the putative gene-product, allowing for metal–ion affinity column-based purification if desired. Recombinant S. cerevisiae [pYES2/NT/lacZ] served as a non-(gtf-1)-expressing control strain in this study. The sugar substrate used in our in vitro assay was UDP-glucose, and the lipid substrates tested were 17-OH oleate, cholesterol, and β-sitosterol. These selections were chosen to represent the general lipid classes of hydroxy fatty acids/esters, animal sterols, and plant sterols. We had tested various solvents (see M&M section) in the preparation of the lipid stock solutions at 50 mM concentration. We found that THF completely dissolved the added lipid; however, it inhibited the enzyme activity. Among all other tested solvents (i.e., DMF, DMSO, ethanol, and ethyl acetate), we observed that DMF is the best-suited solvent to prepare the lipid stock solutions with complete (or near complete) dissolution of substrate at room temperature. When cell extract of the recombinant S. cerevisiae [pYES2/NT/gtf1] was incubated with cholesterol, the presence of glucosyltransferase enzyme activity was confirmed by the appearance of cholesterol glucoside at 35.3 min as evidenced by the appearance of a peak with a 571 m/z ion which corresponds to molecular weight of cholesterol glucoside (548.4) + sodium (23) (Fig. 2). The 571 m/z ion (shown as an Extract Ion Chromatogram, EIC, 571 EIC m/z, in Fig. 2) was absent when the cell extract of the control S. cerevisiae [pYES2/NT/lacZ] strain was incubated with cholesterol, confirming that the Gtf-1 gene-product expressed in the S. cerevisiae [pYES2/NT/gtf1] is a glucosyltransferase enzyme. Similar experiments using β-sitosterol as the lipid substrate gave the same results (data not shown), suggesting that Gtf-1 has a broad substrate-specificity toward sterols.

Table 1 The gtf-1 gene expression in yeast, calculated by ΔΔCt method. Constructs

pYES2/NT/lacZ pYES2/NT/gtf1

Average Ct UBC6 (internal control gene)

gtf-1

27.8 ± 0.1 28.0 ± 0.4

34.6 ± 0.2 15.1 ± 0.1

ΔCt (gtf-1-UBC6)

ΔΔCt (gtf-1-lacZ)

Fold difference relative to lacZ

6.8 ± 0.2 −12.9 ± 0.4

0 −19.7 ± 0.4

1 851,708

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

51

Fig. 2. HPLC–MS of the products of incubation of recombinant S. cerevisiae [pYES2/NT/gtf1] extract with cholesterol. The 571 m/z ion was only present when the extract of the recombinant S. cerevisiae expressing the cloned gtf-1 was incubated with cholesterol, confirming that the gene-product (Gtf-1) is a glucosyltransferase enzyme active toward a sterol substrate.

More importantly, when the S. cerevisiae [pYES2/NT/gtf1] extract was incubated with 17-OH oleate, the presence of glucosyltransferase enzyme activity was confirmed by the appearance of 17-OH oleate glucoside at 5.8 min as evidenced by the appearance of a peak with a 483 m/z ion which corresponds to molecular weight of 17-OH oleate glucoside (460.6) + sodium (23) (Fig. 3). The 483 m/z ion (shown as EIC 483 m/z in Fig. 3) was absent when the extract of control S. cerevisiae [pYES2/NT/lacZ] was incubated with 17-OH oleate. These results further suggested that the cloned C. bombicola gtf-1 gene encodes a glucosyltransferase enzyme that is active toward sterols and hydroxy fatty acids. The finding of the functional property of Gtf-1 – the gene-product of the cloned C. bombicola gtf-1 gene – has many significant implications.

First of all, C. bombicola is an industrial yeast important for the production of sophorolipids. The results of our study strongly suggested for the first time the involvement of Gtf-1 gene-product in the first step of the biosynthesis pathway of sophorolipids by its demonstrated glycosyltransferase activity on 17-OH oleate. Saerens et al. (2011a) had earlier reported the identification of a glucosyltransferase gene UGTA1 in C. bombicola. Through knocking-out the UGTA1 in C. bombicola, it was shown that sophorolipid synthesis was not detected (Saerens et al., 2011a). The enzymatic function of the UGTA1 gene, however, was not confirmed by a direct biochemical assay in the same way as the Gtf-1 gene-product in our present study. Thus the inability of UGTA1 knockout mutant to synthesize sophorolipid could suggest, but could not be conclusively attributed to the absence of glucosyltransferase activity in

52

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53

Fig. 3. HPLC–MS of the products of incubation of recombinant S. cerevisiae [pYES2/NT/gtf1] extract with 17-OH oleate. The 483 m/z ion was only present when the extract of the recombinant S. cerevisiae expressing the cloned gtf-1 was incubated with 17-OH oleate, confirming that the gene-product (Gtf-1) is a glucosyltransferase enzyme with an activity toward a hydroxy fatty acid substrate.

the first step of the biosynthesis pathway of sophorolipids. In addition to furthering the understanding of sophorolipid biosynthesis pathway, our finding that Gtf-1 has a broad spectrum of substrate specificity to include sterols renders this gene-product valuable for potential use in the commercial production of the phytosterol-glucosides with nutriceutical importance (Bouic, 2002; Fernandes and Cabral, 2007; Gabay et al., 2010; Moreau et al., 2002; Quílez et al., 2003). Furthermore, we now have a recombinant yeast strain expressing Gtf-1 glucosyltransferase having a broad spectrum of substrate specificity. By genetically modifying the gtf-1 gene, it is envisioned that mutants would be generated to enable the biosynthesis of new glycolipid biosurfactants and phytosterol-glucoside nutriceuticals having more desired or potent properties.

Conflict of interest Authors declare no conflict of interest.

Acknowledgments The authors thank Nicole Cross, Bun-Hong Lai, Richard Cook and Amy Ream for providing technical assistance. Nucleic acid sequence determination was performed by David S. Needleman, Ph.D. and Sue Lawlor of Eastern Regional Research Center Core Technologies' Integrated Biomolecular Resources group. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

References Altschul, S.F., et al., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25, 3389–3402. Birnboim, H.C., Doly, J., 1979. A rapid alkaline procedure for screening recombinant plasmid DNA. Nucleic Acids Research 7, 1513–1523. Bouic, P.J.D., 2002. Sterols and sterolins: new drugs for the immune system? Drug Discovery Today 7, 775–778. Breithaupt, T.B., Light, R.J., 1982. Affinity chromatography and further characterization of the glucosyltransferases involved in hydroxydocosanoic acid sophoroside production in Candida bogoriensis. The Journal of Biological Chemistry 257, 9622–9628. Cutler, A.J., Light, R.J., 1979. Regulation of hydroxydocosanoic acid sophoroside production in Candida bogoriensis by the levels of glucose and yeast extract in the growth medium. The Journal of Biological Chemistry 254, 1944–1950. Cutler, A.J., Light, R.J., 1982. Effect of glucose concentration in the growth medium on the synthesis of fatty acids by the yeast Candida bogoriensis. Canadian Journal of Microbiology 28, 223–230. Esders, T.W., Light, R.J., 1972a. Glucosyl- and acetyltransferases involved in the biosynthesis of glycolipids from Candida bogoriensis. The Journal of Biological Chemistry 247, 1375–1386. Esders, T.W., Light, R.J., 1972b. Occurrence of a uridine diphosphate glucose: sterol glucosyltransferase in Candida bogoriensis. The Journal of Biological Chemistry 247, 7494–7497. Fernandes, P., Cabral, J.M.S., 2007. Phytosterols: applications and recovery methods. Bioresource Technology 98, 2335–2350. Gabay, O., et al., 2010. Stigmasterol: a phytosterol with potential anti-osteoarthritic properties. Osteoarthritis & Cartilage 18, 106–116. Hardin, R., et al., 2007. Sophorolipids improve sepsis survival: effects of dosing and derivatives. The Journal of Surgical Research 142, 314–319. Kasture, M., et al., 2007. Multiutility sophorolipids as nanoparticle capping agents: synthesis of stable and water dispersible Co nanoparticles. Langmuir 23, 11409–11412. Kim, K., Yoo, D., Kim, Y., Lee, B., Shin, D., Kim, E.-K., 2002. Characteristics of sophorolipid as an antimicrobial agent. Journal of Microbiology and Biotechnology 12, 235–241. Kim, H.-S., Kim, Y.-B., Lee, B.-S., Kim, E.-K., 2005. Sophorolipid production by Candida bombicola ATCC 22214 from a corn-oil processing byproduct. Journal of Microbiology and Biotechnology 15, 55–58. Konishi, M., Fukuoka, T., Morita, T., Imura, T., Kitamoto, D., 2008. Production of new types of sophorolipids by Candida batistae. Journal of Oleo Science 57, 359–369.

D.K.Y. Solaiman et al. / Gene 540 (2014) 46–53 Krivobok, S., Guiraud, P., Seigle-Murandi, F., Steiman, R., 1994. Production and toxicity assessment of sophorosides from Torulopsis bombicola. Journal of Agricultural and Food Chemistry 42, 1247–1250. Lang, S., Katsiwela, E., Wagner, F., 1989. Antimicrobial effects of biosurfactants. Fett Wissenschaft Technologie 91, 363–366. Liu, Y., Ream, A., 2008. Gene expression profiling of Listeria monocytogenes strain F2365 during growth in ultrahigh-temperature-processed skim milk. Applied and Environmental Microbiology 74, 6859–6866. Moreau, R.A., Whitaker, B.D., Hicks, K.B., 2002. Phytosterols, phytostanols, and their conjugates in foods: structural diversity, quantitative analysis, and health-promoting uses. Progress in Lipid Research 41, 457–500. Nitschke, M., Costa, S.G.V.A.O., 2007. Biosurfactants in food industry. Trends in Food Science & Technology 18, 252–259. Pfaffl, M.W., 2001. A new mathematical model for relative quantification in real-time RT–PCR. Nucleic Acids Research 29, e45. Quílez, J., García-Lorda, P., Salas-Salvadó, J., 2003. Potential uses and benefits of phytosterols in diet: present situation and future directions. Clinical Nutrition 22, 343–351. Rau, U., Hammen, S., Heckmann, R., Wray, V., Lang, S., 2001. Sophorolipids: a source for novel compounds. Industrial Crops and Products 13, 85–92. Saerens, K.M.J., Roelants, S.L.K.W., Van Bogaert, I.N.A., Soetaert, W., 2011a. Identification of the UDP-glucosyltransferase gene UGTA1 responsible for the first glucosylation step in the sophorolipid biosynthetic pathway of Candida bombicola ATCC 22214. FEMS Yeast Research 11, 123–132. Saerens, K.M.J., Zhang, J., Saey, L., Van Bogaert, I.N.A., Soetaert, W., 2011b. Cloning and functional characterization of the UDP-glucosyltransferase UgtB1 involved in sophorolipid production by Candida bombicola and creation of a glucolipid-producing yeast strain. Yeast 28, 279–292. Shah, V., et al., 2005. Sophorolipids, microbial glycolipids with anti-human immunodeficiency virus and sperm-immobilizing activities. Antimicrobial Agents and Chemotherapy 49, 4093–4100. Solaiman, D.K.Y., 2000. PCR cloning of Pseudomonas resinovorans polyhydroxyalkanoate biosynthesis genes and expression in Escherichia coli. Biotechnology Letters 22, 789–794. Solaiman, D.K.Y., Ashby, R.D., 2005. Genetic characterization of the poly(hydroxyalkanoate) synthases of various Pseudomonas oleovorans strains. Current Microbiology 50, 329–333.

53

Solaiman, D.K.Y., Ashby, R.D., Foglia, T.A., Nuñez, A., Marmer, W.N., 2004a. Sophorolipids— emerging microbial biosurfactants. Informs 15, 270–272. Solaiman, D.K.Y., Ashby, R.D., Nuñez, A., Foglia, T.A., 2004b. Production of sophorolipids by Candida bombicola grown on soy molasses as substrate. Biotechnology Letters 26, 1241–1245. Suto, M., Tomita, F., 2001. Induction and catabolite repression mechanisms of cellulase in fungi. Journal of Bioscience and Bioengineering 92, 305–311. Teste, M.A., Duquenne, M., François, J.M., Parrou, J.L., 2009. Validation of reference genes for quantitative expression analysis by real-time RT-PCR in Saccharomyces cerevisiae. BMC Molecular Biology 10, 99–113. Van Bogaert, I.N.A., De Maeseneire, S.L., De Schamphelaire, W., Develter, D., Soetaert, W., Vandamme, E.J., 2007a. Cloning, characterization and functionality of the orotidine-5phosphate decarboxylase gene (URA3) of the glycolipid-producing yeast Candida bombicola. Yeast 24, 201–208. Van Bogaert, I.N.A., Develter, D., Soetaert, W., Vandamme, E.J., 2007b. Cloning and characterization of the NADPH cytochrome P450 reductase gene (CPR) from Candida bombicola. FEMS Yeast Research 7, 922–928. Van Bogaert, I.N.A., Saerens, K., De Muynck, C., Develter, D., Soetaert, W., Vandamme, E.J., 2007c. Microbial production and application of sophorolipids. Applied Microbiology and Biotechnology 76, 23–34. Van Bogaert, I.N.A., De Maeseneire, S.L., Develter, D., Soetaert, W., Vandamme, E.J., 2008. Cloning and characterisation of the glyceraldehyde 3-phosphate dehydrogenase gene of Candida bombicola and use of its promoter. Journal of Industrial Microbiology & Biotechnology 35, 1085–1092. Van Bogaert, I.N.A., Sabirova, J., Develter, D., Soetaert, W., Vandamme, E.J., 2009. Knocking out the MFE-2 gene of Candida bombicola leads to improved medium-chain sophorolipid production. FEMS Yeast Research 9, 610–617. Zerkowski, J.A., Solaiman, D.K.Y., 2006. Synthesis of polyfunctional fatty amines from sophorolipid-derived 17-hydroxy oleic acid. Journal of the American Oil Chemists' Society 83, 621–628. Zerkowski, J.A., Solaiman, D.K.Y., 2007. Polyhydroxy fatty acids derived from sophorolipids. Journal of the American Oil Chemists' Society 84, 463–471.