Discovery of a rhamnose utilization pathway and rhamnose-inducible promoters in Pichia pastoris

The rhamnose utilization pathway in Pichia pastoris has not been clarified although this strain can grow well on rhamnose as a sole carbon source. In this study, four genes, PAS_chr4_0338, PAS_chr4_0339, PAS_chr4_0340, and PAS_chr4_0341, were, for the first time, predicted to be involved in rhamnose metabolism along with the previously identified gene PAS_chr1_4-0075. Moreover, expression of these genes, especially PAS_chr4_0341 and PAS_chr1_4-0075 designated as LRA4 and LRA3, was confirmed to significantly increase and clearly decrease in the presences of rhamnose and glucose, respectively. LRA4 encoding a putative L-2-keto-3-deoxyrhamnonate aldolase, was further confirmed via gene disruption and gene complementation to participate in rhamnose metabolism. Using β-galactosidase and green fluorescent protein as reporters, the promoters of LRA4 and LRA3 performed well in driving efficient production of heterologous proteins. By using food grade rhamnose instead of the toxic compound methanol as the inducer, the two promoters would be excellent candidates for driving the production of food-grade and therapeutically important recombinant proteins.


Results
Putative genes involved in L-rhamnose metabolism in P. pastoris. P. pastoris grew slightly slower on 2% (w/v) L-rhamnose than on glucose (Fig. 1A), suggesting that P. pastoris can utilize L-rhamnose as a sole carbon source. However, to date, genes involved in L-rhamnose metabolism have not been elucidated in P. pastoris except for LRA3, which was annotated as L-rhamnonate dehydratase after genome sequencing. By searching the nucleotide sequences of genes homologous to those of genes related to rhamnose metabolism in S. stipitis, the four following genes were predicted to participate in rhamnose utilization in P. pastoris: PAS_chr4_0338, PAS_ chr4_0339, PAS_chr4_0340, and LRA4, which putatively encoded L-rhamnose 1-dehydrogenase, L-rhamnono-ɤlactonase, a transcription regulator, and L-2-keto-3-deoxyrhamnonate (L-KDR) aldolase, respectively (Fig. 1B).
To further investigate the relationships between theses genes and rhamnose metabolism, P. pastoris was cultured in BMDY, BMRY, and BMDRY media. When rhamnose was used as a sole carbon source, transcription levels of all four genes and LRA3 increased. In particular, transcription levels of LRA3 and LRA4 were enhanced over 4000-fold compared to P. pastoris growing on glucose. When a mixture of rhamnose and glucose was used as the carbon source, the expression levels of these genes clearly decreased (Fig. 1C). The above results showed that expression of the five genes was induced by rhamnose and repressed by glucose, and also indicated that these genes are probably related to rhamnose metabolism. Moreover, transcriptional levels of LRA3 and LRA4 were respectively 75% and 22% of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcription levels in the presence of an initial concentration of 2% rhamnose (Fig. 1D). Additionally, the above results also imply that the LRA3 promoter (P LRA3 ) was strongly rhamnose-inducible, while the LRA4 promoter (P LRA4 ) was weakly rhamnose-inducible.
LRA4 is a homolog of L-KDR aldolase from S. stipitis according to amino acid sequence identity. In addition, LRA4 showed high homology when aligned with aldolases from Debaryomyces hansenii, Candida tenuis and Glarea lozoyensis. Furthermore, two amino acid residues (a lysine residue and a tyrosine residue) that are essential active sites in L-KDR aldolase from S. stipitis 10 are also present in LRA4 ( Fig. 2A). In order to confirm that LRA4 was associated with L-rhamnose metabolism, a LRA4-disrupted strain and a LRA4-complemented strain were developed by gene disruption and gene complementation, respectively. The wild-type strain, LRA4-disrupted strain, and LRA4-complemented strain were grown on glucose or rhamnose as a sole carbon source. As expected, all three strains grew well on glucose, but only the wild-type strain and the LRA4-complemented strain grew well on rhamnose (Fig. 2B), which suggests that LRA4 is definitely involved in L-rhamnose utilization.
Expression of heterologous genes driven by P LRA3 and P LRA4 in P. pastoris. Previously, LRA4 and LRA3 exhibited relatively high transcription activities in the presence of rhamnose. To investigate whether P LRA3 and P LRA4 were subject to strict rhamnose regulation, the effects of several carbon sources on expression of lacB under the control of P LRA3 and P LRA4 were monitored. Expression of lacB driven by P LRA3 and P LRA4 was only observed when using rhamnose as a sole carbon source. In contrast, under control of the well-characterized Relative transcription profiles of these genes in P. pastoris grown in different media. The relative expression value of each gene in BMDY was assigned as 1 and the GAPDH gene was used as a reference (C); the relative expression value of the GAPDH gene was used as a reference and was assigned as 100% (D). Each test was conducted in triplicate and the results are presented as means ± SEM of three replicates (n = 3).
Scientific RepoRts | 6:27352 | DOI: 10.1038/srep27352 GADPH promoter (P GAP ) lacB expression was observed when methanol, glucose, mannitol, sorbitol, and glycerol were used as sole carbon sources, in addition to rhamnose (Fig. 3A). The above results indicate that P LRA3 and P LRA4 were strict rhamnose-dependent promoters. Additionally, a dose-effect relationship occurred between rhamnose concentration (0.25-2.0%) and production of β -galactosidase driven by P LRA3 and P LRA4 (Fig. 3B), meaning that production of a target protein could be controlled in a rhamnose-dose dependent manner.
To validate the efficiency of the rhamnose-inducible P LRA3 and P LRA4 in target gene expression, P GAP and the two identified rhamnose-inducible promoters were subjected to a comparative protein production study using β -galactosidase and green fluorescent protein (GFP) as reporters. Vectors pGHLRA3α , pGHLRA4α , and pGH-GAPα based on P LRA3 , P LRA4 , and P GAP , respectively, were developed, with pGHLRA3α illustrated as an example  in Fig. 4A. One reporter gene, lacB, encoding a β -galactosidase, was ligated into the above vectors via the SnaBI and NotI restriction sites to generate pGHLRA3α LacB, pGHLRA4α LacB, and pGHGAPα LacB, in which lacB expression was controlled by P LRA3 , P LRA4 , and P GAP , respectively, and was then integrated into the gas1 locus of the P. pastoris chromosome after transformation. The β -galactosidase activities in culture supernatants of the three recombinant P. pastoris strains grown in BMRY and BMDY were monitored. It was observed that β -galactosidase activities in supernatants of the different cultures in BMRY increased with incubation duration and reached a maximum at 72 h: 23 U/ml, 18 U/ml, and 6 U/ml for recombinant strains harboring pGHLRA3α LacB, pGH-LRA4α LacB, and pGHGAPα LacB, respectively (Fig. 4B). The target protein productions under the control of P LRA3 and P LRA4 in the presence of rhamnose were approximately 80% and 26% of that under P GAP , respectively, which was consistent with the transcription levels determined by real-time PCR. Production of β -galactosidase in culture supernatants of the different recombinant strains was assayed via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 4C). To investigate whether the two promoters also drove efficient expression of other genes, expression of another reporter gene, GFP (gfp), was also monitored. For intracellular expression of gfp, three new vectors, pGHLRA3, pGHLRA4, and pGHGAP, were developed without a protein secretion signal based on vectors pGHLRA3α , pGHLRA4α , and pGHGAPα . gfp was ligated into vectors pGHLRA3, pGHLRA4, and pGHGAP via the EcoRI and NotI restriction sites, and was integrated into the gas1 locus of the P. pastoris chromosome. Green fluorescence was monitored after the recombinant strains were grown in BMRY. Fluorescence was observed in all three recombinant strains, in which gfp expression was regulated by P GAP , P LRA3 , and P LRA4 . Mean green fluorescence intensity was 235 (P LRA4 ), 563 (P LRA3 ) and 2249 (P GAP ) respectively, which indicated that the strengths of the three promoters were ranked as follows: P GAP (high), P LRA3 (middle), and P LRA4 (weak) (Fig. 4D). According to the above results, P LRA3 would be more advantageous for inducible expression of target genes than P LRA4 .
Characterization of putative P LRA3 . P LRA3 , a newly discovered strong rhamnose-inducible promoter, could potentially be widely used to drive production of proteins of interest in P. pastoris. To identify the minimal promoter region, 5′ -deletion constructs of P LRA3 were derived from pGHLRA3α LacB and were inserted into pGHLRA3α LacB between the AscI and SnaBI sites, generating pGHLRA3(258)α LacB, pGHLRA3(210) α LacB, pGHLRA3(140)α LacB, pGHLRA3(120)α LacB, pGHLRA3(100)α LacB, pGHLRA3(85)α LacB, pGH-LRA3(48)α LacB, and pGHLRA3(0)α LacB. These deletion constructs were transformed into P. pastoris and the β -galactosidase activities in culture supernatants of the different recombinant strains were determined. The results showed that a promoter harboring a 5′ flanking region greater than 210 bp upstream of the ATG of the initiating methionine of LRA3 was sufficient for transcriptional activation (Fig. 5A), which suggests that a 210-bp DNA fragment harbors all the necessary elements to express recombinant proteins in P. pastoris. The promoter with a 5′ flanking region containing the 140 bp upstream of the translation start site was not as efficient, indicating that elements in the DNA region between −140 bp and −210 bp were essential to efficient expression of target genes. Additionally, some important regulatory elements regulating gene expression occur between −100 bp and −140 bp, as the deletion of this region resulted in an obvious decrease in β -galactosidase activity (Fig. 5A).
The efficiency of P LRA3(210) and P LRA3 was further determined using β -galactosidase as a reporter gene and P AOX1 as a positive control promoter. The β -galactosidase activities in culture supernatants of recombinant strains harboring pGHLRA3α LacB, pGHLRA3(210)α LacB and pPIC9LacB increased during the induction. At 96 h, activity and production of β -galactosidase in culture supernatants of strain harboring pGHLRA3α LacB and pGHLRA3(210)α LacB were almost same, and were less than half of that of strain harboring pPIC9LacB (Fig. 5B,C). The results further indicated that P LRA3 was a middle strong rhamnose-inducible promoter. The nucleotide sequence of P LRA3 was shown and the region between −140 to −210 bp was underlined (Fig. 5D).

Discussion
Identification of genes related to rhamnose metabolism in P. pastoris allowed us to further elucidate the rhamnose metabolism pathway and to exploit inducible promoters for controllable expression of target genes in P. pastoris. In this study, a set of genes involved in rhamnose metabolism in P. pastoris was first predicted according to the nucleotide sequence identity to genes in S. stipitis. Then, LRA4 was further confirmed to correlate with rhamnose metabolism based on the result that LRA4 -disrupted P. pastoris did not grow on rhamnose as a sole carbon source, while the wild-type strain and the LRA4-complemented strain did. To further confirm the physiological role of this gene, the biochemical characterization of the enzyme encoded by LRA4 was examined in another study. Similarly, the physiological roles of the other genes we identified should also be verified via gene disruption and complementation and biochemical function assays in future studies. Additionally, it was noted that the genes encoding the enzymes for rhamnose metabolism in P. pastoris did not cluster. Three genes (PAS_chr4_0338, PAS_chr4_0339, and LRA4) and the regulator gene (PAS_chr4_0340) were organized in a gene cluster, while LRA3 was independent. This is not the same organization found in S. stipitis, in which a set of genes correlated with rhamnose catabolism clustered in a rhamnose operon. The difference reflects different evolutionary processes in rhamnose metabolism between S. stipitis and P. pastoris. P AOX1 , which is strictly and strongly activated by methanol, was the promoter most widely used for production of heterologous proteins in P. pastoris. However, strong promoters that produce heterologous proteins in high quantities are unsuitable for certain situations, such as production of active products that are highly dependent on correct folding or processing in the secretory pathway 18 or the expression of various lethal or toxic genes; in these cases weaker promoters have been reported to be more feasible 19,20 . Also, in addition to using the toxic compound methanol as the inducer, P AOX1 -based fermentation includes the following disadvantages: 1) relatively long fermentation duration and 2) sophisticated feeding strategies including but not limited to, i) initial culture of recombinants in buffered glycerol medium, followed by ii) harvesting and resuspending cells in methanol medium. To broaden the applications of P. pastoris and simplify the process of fermentation, alternative promoters that do not require methanol are in demand, and various promoters have been described 21 . Five rhamnose-inducible genes were discovered in our study, and among them P LRA3 and P LRA4 were strictly activated in the presence of rhamnose and almost completely repressed by glucose. P LRA3 in the presence of rhamnose was confirmed to be comparable to the typical strong promoter, P GAP , which has been commonly used to drive efficient expression of target genes in P. pastoris. Due to using rhamnose instead of the flammable and toxic compound methanol as an inducer, P LRA3 could be used for substantial production of recombinant proteins in the food industry and production of therapeutically important recombinant proteins for commercial applications. As for P LRA4 , it was much weaker than P LRA3 and is a better candidate for certain academic purposes, for example, controllably expressing potentially lethal or toxic genes that affect cell viability. Few promoters isolated from genes related to sugar metabolism have been reported in P. pastoris and, to our knowledge, the sorbitol dehydrogenase promoter, a constitutively expressed promoter with a transcription strength similar to P GAP , is the only one that had been intensively investigated prior to our study. Compared to the sorbitol dehydrogenase promoter, P LRA3 is disadvantageous in its transcription activity, however P LRA3 is advantageous for controllable expression of heterologous genes, including some genes in which the products might be toxic to the host cell.

Materials and Methods
Bacterial strains and media. E. coli Trans1-T1 (TransGen, Beijing, China) and P. pastoris GS115 (Invitrogen, Carlsbad, CA, USA) were used as the hosts for gene cloning and expression, respectively. The pEasy-Blunt Simple Cloning Vector was purchased from TransGen.
The primers used for PCR are listed in Supplementary Table S1.
Total RNA preparation and real-time PCR. P. pastoris GS115 was grown in BMDY, BMRY, or BMDRY to OD 600 1.2-1.4, and then cells were harvested by centrifugation (12,000 g × 5 min) at 4 °C and immediately stored at −80 °C. Total RNA from P. pastoris cells was extracted with Trizol reagent (Invitrogen) followed by digestion with DNase I (RNase-free; MBI Fermentas, St. Leon-Roth, Germany) to remove trace DNA. cDNA was synthesized from each RNA sample (1 μ g) using the RevertAid H Minus First Strand cDNA Synthesis Kit (K1631; MBI Fermentas). Real-time PCR assays were performed in the IQ5 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA) using the SuperReal PreMix SYBR Green I Kit (Toyobo, Osaka, Japan). Primers used for real-time PCR are listed in Table S1. The GAPDH gene was used as a reference. All samples were analyzed in triplicate and the data are presented as means ± standard deviation (SD).  F1 and gas1-R1, gas1-F2 and gas1-R2, respectively). HIS4, a selectable marker for isolating Pichia recombinant strains, was amplified from plasmid pPIC9 using the primers his4-F and his4-R. The three abovementioned DNA fragments were ligated into a DNA fragment named "GH" via overlap-extension PCR with the primers gas1-LF and gas1-RR. In addition, a DNA fragment designated "OA" containing the origin of replication derived from pBR322 and an ampicillin resistance gene was amplified from plasmid pPIC9 with primers OA-F and OA-R. Finally, the two DNA fragments, GH and OA, were ligated into a circular molecule by homologous recombination using the CloneEZ Kit (GenScript, Nanjing, China), generating the recombination vector pGH01. pGH01 was integrated into the gas1 locus of the P. pastoris chromosome via homologous recombination.

Construction of a homologous recombination vector in
LRA4 disruption and complementation. DNA fragments upstream and downstream of LRA4, which were used as homologous arms for integration of target genes into the LRA4 locus, were amplified from genomic DNA of P. pastoris GS115. The expression cassette of a zeocin resistance gene was obtained from the plasmid pPICzα . The three PCR fragments were fused by overlap-extension PCR to generate a gene disruption cassette for disruption of LRA4 in P. pastoris. From the resultant DNA fragments, 500 ng were transformed into P. pastoris GS115 via electroporation. Transformants were first screened on yeast extract-peptone-dextrose (YPD) agar plates containing 100 μ g/ml zeocin and then verified by PCR.
To complement LRA4, the complementation plasmid pGH/LRA4 was developed as follows. A DNA fragment harboring full length LRA4 was cloned from genomic DNA of P. pastoris GS115 and then ligated into the pGH01 vector via the AscI and PacI restriction sites to generate the pGH/LRA4 plasmid. pGH/LRA4 was introduced into P. pastoris GS115 with LRA4-disruption via electroporation after linearization by the restriction enzyme SwaI. Transformants were screened on MD medium and then verified by PCR and DNA sequencing.
Expression of exogenous genes under P LRA3 and P LRA4 in P. pastoris. DNA fragments containing the putative promoter regions of LRA3, LRA4, and GAPDH were obtained from the genomic DNA of P. pastoris. The DNA fragment comprising the protein secretion signal, multiple cloning site, and transcription termination was cloned from plasmid pPIC9. The two DNA fragments were fused by overlap-PCR to generate the exogenous gene expression cassette in which exogenous gene expression was driven by P LRA3 , P LRA4 and P GAP , and then the cassettes were respectively ligated into vector pGH01 via the AscI and PacI restriction sites to generate expression vectors pGHLRA3α , pGHLRA4α , and pGHGAPα . Additionally, three vectors, pGHLRA3, pGHLRA4, and pGHGAP, were developed without the protein secretion signal in the same manner as pGHLRA3α , pGHLRA4α , and pGHGAPα .
The lacB gene (GenBank accession number: AJ431643.1), which was cloned from the Aspergillus candidus chromosome and encodes a β -galactosidase, was ligated into pGHLRA3α , pGHLRA4α , and pGHGAPα and pPIC9 via the SnaBI and NotI restriction sites. The gfp gene (GenBank accession number: JQ733033.1) encoding green fluorescent protein was synthesized by GenScript Biotechnology Co. Ltd. and was ligated into pGHLRA3, pGHLRA4, and pGHGAP via the EcoRI and NotI restriction sites.
To evaluate the strength of P LRA3 , P LRA3(210) and P AOX1 , a single colony of different recombinant Pichia cells was inoculated into in 50-ml YPD medium at 28 °C for 48 h in shaking flasks. The cells with lacB expression under the control of P LRA3 and P LRA3(210) were harvested and then grown in 50-ml BMRY medium, and the cells with lacB expression under the control of P AOX1 were harvested and then grown in 50-ml BMMY medium with addition of methanol to a final concentration of 1.0% [v/v] at a interval of 24 h. β -galactosidase activity and production in culture supernatant were monitored at intervals.
The β -galactosidase activity was assayed as described by Nie et al. 15 , with minor modification. The reaction system containing 200 μ l of enzyme solution and 800 μ l of 0.25% (w/v) ortho-nitrophenyl-β -galactoside (oNPG) in pre-incubated (60 °C for 5 min) phosphate-citrate buffer (50 mM, pH 5.2) was incubated at 60 °C for 15 min, followed by sequential addition of 1 ml of 10% trichloroacetic acid and 2 ml of 1 M Na 2 CO 3 . The absorbance at 420 nm was measured. One unit of β -galactosidase was defined as the amount of enzyme that released 1 μ mol of o-nitrophenol per minute under standard conditions (pH5.2, 60 °C, 15 min). gfp expression and assays of green fluorescence intensity. A single colony of recombinant Pichia cells was inoculated into in YPD medium and gown overnight at 28 °C with vigorous shaking (200 rpm). The cultures were inoculated into fresh BMRY medium at 1% (v/v) and grown for 48 h at 28 °C with vigorous shaking (200 rpm). Recombinant Pichia cells were then collected, washed three times, and resuspended in 0.9% NaCl. Green fluorescence intensity in recombinant cells was determined using a laser scanning confocal microscope (Nikon, Japan) with 2s per image to collect a 1024 × 1024 pixel image.