|Home | About | Journals | Submit | Contact Us | Français|
Two strains of the methylotrophic yeast Pichia pastoris were used to establish cyanophycin (multi-l-arginyl-poly-l-aspartic acid [CGP]) synthesis and to explore the applicability of this industrially widely used microorganism for the production of this polyamide. Therefore, the CGP synthetase gene from the cyanobacterium Synechocystis sp. strain PCC 6308 (cphA6308) was expressed under the control of the alcohol oxidase 1 promoter, yielding CGP contents of up to 10.4% (wt/wt), with the main fraction consisting of the soluble form of the polymer. To increase the polymer contents and to obtain further insights into the structural or catalytic properties of the enzyme, site-directed mutagenesis was applied to cphA6308 and the mutated gene products were analyzed after expression in P. pastoris and Escherichia coli, respectively. CphA6308Δ1, which was truncated by one amino acid at the C terminus; point mutated CphA6308C595S; and the combined double-mutant CphA6308Δ1C595S protein were purified. They exhibited up to 2.5-fold higher enzyme activities of 4.95 U/mg, 3.20 U/mg, and 4.17 U/mg, respectively, than wild-type CphA6308 (2.01 U/mg). On the other hand, CphA proteins truncated by two (CphA6308Δ2) or three (CphA6308Δ3) amino acids at the C terminus showed similar or reduced CphA enzyme activity in comparison to CphA6308. In flask experiments, a maximum of 14.3% (wt/wt) CGP was detected after the expression of CphA6308Δ1 in P. pastoris. For stabilization of the expression plasmid, the his4 gene from Saccharomyces cerevisiae was cloned into the expression vector used and the constructs were transferred to histidine auxotrophic P. pastoris strain GS115. Parallel fermentations at a one-to-one scale revealed 26°C and 6.0 as the optimal temperature and pH, respectively, for CGP synthesis. After optimization of fermentation parameters, medium composition, and the length of the cultivation period, CGP contents could be increased from 3.2 to 13.0% (wt/wt) in cells of P. pastoris GS115 expressing CphA6308 and up to even 23.3% (wt/wt) in cells of P. pastoris GS115 expressing CphA6308Δ1.
Since the first isolation of a methylotrophic yeast, Kloeckera sp. strain 2201, in 1969 (43), the two methylotrophic yeasts Pichia pastoris and Hansenula polymorpha have become the most popular methylotrophs in industry and academia (9, 23, 24). The main benefits of these organisms for the production of recombinant proteins are their growth to cell densities as high as 130 g cell dry matter per liter (50, 57) and the availability of strong and tightly regulated promoters that result in a high product yield (13). Viral hepatitis B surface antigen, S. cerevisiae mating factor α, and S. cerevisiae invertase are only a few examples of compounds produced by recombinant P. pastoris (reviewed in reference 9).
A variety of strains were optimized for the expression of recombinant proteins (9). Protease-deficient strains such as strain KM71(H) were generated to circumvent the proteolytic degradation of recombinant proteins (17). Three different phenotypes exist that differ in the ability to utilize methanol (reviewed in reference 37). (i) Mut+ strains grow on methanol as the sole carbon and energy source at the wild-type rate. (ii) Muts strains possess a disrupted alcohol oxidase 1 (AOX1) gene and therefore rely on the weaker AOX2 gene, leading to decreased methanol utilization rates in comparison to those exhibited by Mut+ strains. (iii) Mut− strains are not able to utilize methanol as a carbon and energy source; consequently, such strains use the compound as an inducer only and are dependent on the concomitant addition of carbon sources that do not repress the AOX1 promoter (30, 31). Depending on the required product, any of these phenotypes can be optimal (37). The AOX1 promoter is totally repressed during growth on, e.g., glycerol, whereas it is strongly expressed after methanol is supplied (11). Therefore, P. pastoris fermentations are divided into two phases. (i) During growth on glycerol, high cell densities are reached; (ii) subsequent growth on methanol leads to induction of heterologous protein synthesis, resulting in a high product yield (14). Besides glycerol, several other carbon sources, such as, e.g., glucose, acetate, ethanol, or sorbitol, were used for the production of foreign proteins (30, 31). Several fermentation strategies that allow optimal cell and product yields have been established (8, 25, 28).
Besides the AOX1 promoter, several other suitable promoters are available (10), e.g., the copper-inducible CUP1 promoter from S. cerevisiae (33, 38), the inducible ICL1 promoter from the isocitrate lyase gene (8), or the constitutive GAP promoter from glyceraldehydes-3-phosphate dehydrogenase (56).
Synthesis of cyanophycin (multi-l-arginyl-poly-l-aspartic acid [CGP]) was only recently established in the yeast S. cerevisiae. Recombinant strains harboring cphA from Synechocystis sp. strain PCC 6308 but otherwise with a wild-type background accumulated CGP up to 6.9% (wt/wt) (52), whereas recombinant strains with a mutation in arginine metabolism accumulated CGP even up to 15.3% (wt/wt) of the cell dry mass (CDM) (54). All of the strains synthesized the polymer in soluble and insoluble forms, which was also observed in transgenic plants (29, 42); the soluble type of CGP was first observed in Escherichia coli expressing the cphA gene from Desulfitobacterium hafniense (59). Several cyanobacterial and heterotrophic CGP synthetase genes were expressed heterologously in the past (16, 26, 29, 52, 59). To unravel structurally or catalytically relevant residues of the enzyme, a few site-directed mutations were generated in cyanobacterial cphA genes (26, 27, 35, 53). In addition, several variations in the amino acid composition of the polymer were recently obtained; while cyanobacterial CGP or CGP synthesized by specific CphA proteins exhibiting a narrow substrate range contained aspartate and arginine only (18, 51); lysine was observed as a component replacing arginine at up to 18 mol% in recombinant strains of E. coli and S. cerevisiae harboring CphA with a broader substrate range (34, 54). Moreover, citrulline and ornithine were also detected as constituents replacing arginine in mutants of S. cerevisiae expressing CphA from Synechocystis sp. strain PCC 6308 (54). The soluble CGP contained up to 20 mol% citrulline or up to 8 mol% ornithine instead of arginine. The latter enzyme also revealed a wide substrate range in vitro comprising agmatine and canavanine besides arginine, lysine, citrulline, and ornithine (2, 58).
A multitude of technical or pharmaceutical applications are known for degradation products of CGP (44, 48, 49). Dipeptides obtained after α cleavage of the polymer by cyanophycinases are employed as high-value pharmaceuticals (45, 46). Through β cleavage of the polymer, polyaspartic acid can be obtained, which serves as a biodegradable alternative to the persistent polyacrylic acid (9). Finally, research on the synthesis of bulk chemicals such as urea or acrylonitrile from CGP has become of special interest (40, 48, 49).
In this study, the methylotrophic yeast P. pastoris was, for the first time, employed for synthesis of the polyamide CGP to analyze if this organism provides a perspective for the production of the polymer. For further optimization of polymer yields, mutated CphA proteins were generated by site-directed mutagenesis and characterized and optimal growth parameters were determined in parallel fermentations.
All of the bacteria, yeast strains, and plasmids used in this study are listed in Table Table1.1. For the growth conditions and media used, see the supplemental material.
Cultivation at the 1- to 2-liter scale was performed in parallel fermentors (Biostat Bplus—Twin 2 L MO; Sartorius). One liter of medium was inoculated with 100 ml of a well-grown culture in YPG medium (yeast extract-peptone-dextrose with 2%, wt/vol, glycerol instead of glucose). If not indicated otherwise, the pH was kept at 5.0 and the temperature was set at 30°C. The pH was adjusted automatically by addition of 4 M HCl or 4 M NaOH. When not stated differently, the mineral salts medium described by Henes and Sonnleitner (28) with 2% (wt/wt) glycerol as the carbon source was used. During fermentation, the ammonium concentration was determined by using a Quantofix kit (10 to 400 mg/liter NH4+; Macherey-Nagel, Düren, Germany); when the ammonium was depleted, 4.5 g/liter (NH4)2SO4 was fed aseptically to the cultures. Glycerol concentrations were determined by high-performance liquid chromatography (HPLC) using a LaChrom Elite apparatus (6). Cell-free culture supernatants were used as samples, and a 20-min program was run. Methanol feeding (1%) was started approximately 24 h after inoculation when glycerol was depleted. Cells were harvested in a Sorvall RC5B centrifuge for 40 min at 8,000 rpm (rotor SLA3000) and washed once with saline (0.9% NaCl) before freezing and lyophilization.
For cloning of cphA6308 into E. coli-P. pastoris shuttle vectors pPICHOLI-3 and pPICHOLI-C (Table (Table1),1), PCR was done with Pfx DNA polymerase (Gibco BRL) according to the manufacturer's instructions by using oligonucleotides fw-SalI and rw-NotI as sense and reverse primers, respectively (Table (Table2).2). Plasmid pET-23a::cphA6308 (Table (Table1)1) was used as the template. Subsequently, the PCR products were cloned into the SalI-NotI-treated E. coli-yeast shuttle vectors, yielding pPICHOLI-3::cphA6308 and pPICHOLI-C::cphA6308, respectively. As significant CGP accumulation was only observed using pPICHOLI-3, further cphA genes (see below) were cloned into this vector only.
Five different cphA genes were generated by PCR using specific oligonucleotides (Table (Table2).2). Three genes coding for C-terminally truncated CphA proteins were constructed: cphA6308Δ1 coding for CphA6308Δ1 with one amino acid truncated at the C terminus of the enzyme, cphA6308Δ2 coding for CphA6308Δ2 with two amino acids truncated, and cphA6308Δ3 coding for CphA6308Δ3 with three amino acids truncated. Furthermore, the gene cphA6308C595S coding for CphA6308 with the point mutation Cys595Ser (35, 53) was (i) amplified with the primers used for cphA6308 and (ii) used as the template to generate cphA6308Δ1C595S with the antisense primer used for the generation of cphA6308Δ1. All amplified and purified cphA genes were treated as described for cphA6308 and cloned into the vector pPICHOLI-3 (Table (Table1).1). All constructs were transferred to P. pastoris strains GS115 and KM71H and E. coli BL21(DE3)pLysS.
Strain GS115 is histidine auxotrophic and lacks an active his4 gene coding for a multifunctional enzyme containing phosphoribosyl-ATP pyrophosphatase, phosphoribosyl-AMP cyclohydrolase, and histidinol dehydrogenase activities; it catalyzes the 2nd, 3rd, 9th, and 10th steps of histidine biosynthesis. To enable strict stabilization of the pPICHOLI vectors in P. pastoris strain GS115, the his4 gene was amplified with chromosomal DNA from S. cerevisiae strain G175 (Table (Table1)1) as the template and primers containing NotI recognition sites (Table (Table2).2). S. cerevisiae was chosen as the his4 donor to avoid homologous recombination between the episomally encoded intact his4 gene and the his4 gene of P. pastoris GS115 with a chromosomally encoded defect. The functionality of this gene in P. pastoris was shown before (12). The sense primer binds 266 bp upstream of the his4 start codon to amplify the gene with its own promoter region. The 2,666-bp PCR product was subcloned into the cloning vector pJET1.2 (Fermentas). Plasmid pJET1.2::his4 was restricted with NotI, and his4 was purified and ligated into the vectors pPICHOLI-3::cphA6308 and pPICHOLI-3::cphA6308Δ1C595S, which were also restricted with NotI and dephosphorylated with FastAP (Fermentas) to avoid religation. The constructs obtained, pPICHOLI-3::cphA6308/his4 and pPICHOLI-3::cphA6308Δ1/his4, were transferred to P. pastoris GS115. Transformants were selected on minimal medium without histidine and supplemented with 100 μg/ml zeocin.
His-tagged CphA proteins were purified under native conditions with His SpinTrap columns as described in the manual provided (GE Healthcare). Binding buffer (pH 7.4) contained 20 mM imidazole; elution buffer (pH 7.4) had an imidazole concentration of 500 mM.
To determine plasmid stabilities, samples were taken from cultures at the indicated times; defined aliquots were spread on minimal medium without amino acids containing or not containing zeocin. The plates were incubated at 30°C for 48 h, and the CFU were determined and analyzed by comparing the CFU on plates without zeocin with the CFU on plates with zeocin.
RNA was isolated from yeast cells grown under induced conditions as described in the manual provided by MoBiTec. RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany) as described by the manufacturer. RT-PCR was performed as described by Qiagen, using the OneStep RT-PCR kit. For a detailed description, see the supplemental material.
Protein and CGP (52) concentrations were determined as described by Bradford (5) and Lowry et al. (36). Soluble cell fractions were used for determination of cellular protein concentrations and were obtained as described above. For determination of CGP concentrations, the polymer was solubilized in buffer (Tris/HCl, pH 7.2) or 0.1 M HCl, according to its solubility behavior.
The CGP synthetase enzyme assay followed the procedure described by Aboulmagd et al. (2). Soluble cell fractions and purified CphA proteins were used to determine enzyme activities. Scintillation counting was carried out with a model LS 6500 scintillation counter (Beckman Instruments GmbH, Munich, Germany).
For isolation of soluble and insoluble CGP on a small scale, yeast cells were disrupted as described above and the soluble and insoluble forms of CGP were isolated from the cell debris by different procedures as described by Steinle et al. (54).
The amino acid constituents of CGP isolated from transgenic yeast strains were determined by HPLC (1, 54). Calibration was done with samples from an amino acid reference kit (Kollektion AS-10 from Serva Feinbiochemica, Heidelberg, Germany).
For the methods used for cell harvesting, cell disruption, electrophoresis, immunological analysis, and transfer of DNA and for general DNA techniques, see the supplemental material.
For heterologous expression in P. pastoris, the cphA gene from Synechocystis sp. strain PCC 6308 was chosen, as it was successfully used for heterologous expression in the yeast S. cerevisiae (52, 54). For expression, two different promoters were chosen to compare them for suitability for efficient CGP synthesis in P. pastoris; cphA6308 was cloned into the expression vectors pPICHOLI-C with the copper-inducible CUP1 promoter and pPICHOLI-3 with the methanol-inducible AOX1 promoter. In S. cerevisiae, use of the CUP1 promoter resulted in significant CGP synthesis (52, 54). As expression strains, two strains differing in methanol sensitivity were used: strain GS115, exhibiting a Mut+ phenotype, and strain KM71H, exhibiting a Muts phenotype. Both strains were transformed with pPICHOLI-3::cphA6308 and pPICHOLI-C::cphA6308, respectively, and cells were grown in flasks under inducible conditions. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of different cell extracts revealed significant CGP synthesis only in strains harboring vector pPICHOLI-3::cphA6308, although RT-PCR employing RNA from the recombinant strains revealed transcription of cphA in any of the four strains. Therefore, all further experiments were only carried out with cphA genes under the control of the AOX1 promoter.
Previous studies revealed that CphA proteins with increased specific activities could be obtained through site-directed mutagenesis (26, 27). Therefore, three mutated cphA genes coding for CphA proteins with truncated C termini (CphA6308Δ1, CphA6308Δ2, CphA6308Δ3) lacking one, two, or three amino acids, respectively, were generated by PCR and cloned into the vector pPICHOLI-3 (see Materials and Methods). Moreover, CphA6308C595S harboring the point mutation Cys595Ser and CphA6308Δ1C595S harboring the same point mutation and possessing the C terminus truncated by one amino acid were amplified by PCR and cloned into pPICHOLI-3. The hybrid vectors generated were transferred to P. pastoris strains GS115 and KM71H, respectively. All CphA proteins were purified by His SpinTrap columns, and their purity was verified by SDS-PAGE and Western blotting using anti-His antibodies indicating that the purified enzymes are CphA proteins. A specific immunoreaction was obtained with a protein band of ca. 100 kDa, corresponding to the theoretical size of 98.23 kDa calculated for six-His-tagged CphA6308 (www.expasy.org) (see Fig. S1 in the supplemental material). The enzyme activities of the purified CphA proteins were determined (Table (Table3).3). The specific enzyme activities of the purified CphA proteins obtained showed that CphA6308Δ1 exhibited about twofold higher activity (4.95 U/mg) than wild-type CphA6308 (2.01 U/mg), whereas CphA6308Δ2 (1.48 U/mg) and CphA6308Δ3 (0.94 U/mg) showed lower specific activities than wild-type CphA. Also, CphA6308C595S revealed increased enzyme activity (3.20 U/mg) in comparison to that of wild-type CphA but lower activity in comparison to that of CphA6308Δ1. CphA6308Δ1C595S exhibited slightly higher activity (4.17 U/mg) than CphA6308C595S.
Soluble and insoluble CGP polymers of all strains were analyzed by SDS-PAGE to investigate if the mutated CphA proteins lead to differing molecular weight distributions in comparison to CphA6308 (Fig. (Fig.1).1). For this, 50 mg of dry cells of recombinant P. pastoris strain GS115 expressing the different cphA genes was processed as for CGP isolation. Twenty microliters of the soluble cell fractions treated with proteinase K and 20 μl of the acidic fractions obtained after resuspension of the cell debris in 0.1 M HCl were analyzed by SDS-PAGE. Equal amounts of the respective fractions were applied to the gels to display possible differences in the amounts of CGP synthesized. For all strains, except the strain expressing CphA6308Δ3, no differences in the amount of CGP could be concluded from this analysis. For the latter strain, significant differences in the molecular mass distribution of the polymer in comparison to CGP synthesized by CphA6308 were observed: soluble CGP exhibited a maximal mass of 26 kDa, while the masses of the soluble CGPs synthesized by other CphA proteins ranged from 19 to 40 kDa. Additionally, no insoluble CGP could be detected in cells harboring cphA6308Δ3. The other strains synthesized insoluble CGP with a mass distribution between 24 and 35 kDa; thus, this form of the polymer exhibited lower mass distributions than the soluble form of CGP.
Cultivation experiments in flasks were done to determine and compare the CGP contents of P. pastoris strains GS115 and KM71H harboring vector pPICHOLI-3 with the different cphA genes. The cells were grown in 50 ml BMMY medium (39a) supplemented with zeocin. After 24 h, 1% methanol was fed to strain GS115 and 0.5% methanol was fed to strain KM71H; feeding was repeated at the same concentrations for 4 days. Table Table44 shows the cell densities obtained in grams per liter and CGP yields in percent (wt/wt). The amino acid compositions of the soluble and insoluble types of the polymer were determined by HPLC. Insoluble CGP, independent of the CphA or strain used, consisted of a maximum of 3.5 mol% lysine, besides arginine and aspartate. In contrast, soluble CGP exhibited higher molar fractions of lysine (Table (Table3),3), ranging from 4.7 to 6.8 mol% lysine in CGP isolated from transgenic P. pastoris GS115 strains and up to 13.5 mol% lysine in CGP isolated from transgenic P. pastoris KM71H strains.
The CGP contents varied, depending on the CphA expressed, and were comparable for strains GS115 and KM71H. CGP accumulated mainly in the soluble form, which accounted for 85 to 98% of the total polymer. Strains expressing wild-type CphA6308 accumulated CGP to 8.4 to 10.4% (wt/wt) of the CDM, while the contents increased to a maximum of 14.3% (wt/wt) after expression of the mutated enzyme CphA6308Δ1, CphA6308C595S, or CphA6308Δ1C595S. After expression of CphA6308Δ2, CGP contents were comparable to those of strains harboring CphA6308 and cells expressing CphA6308Δ3 synthesized only about 50% of the amount that cells expressing wild-type CphA did (Table (Table44).
As P. pastoris strain GS115 enables strict stabilization of the plasmid due to its histidine auxotrophy, it was used in parallel fermentations in small stirred tank reactors for optimization of the CGP contents of the cells. Therefore, the his4 gene was additionally cloned into the vector pPICHOLI-3::cphA6308 and plasmid stabilities were determined in flask experiments first and then in fermentations (see below). After growth in flasks, 85% or 88% of the recombinant cells grown without zeocin in minimal medium harbored the plasmid, while 98% of the cells expressing the his4 gene retained the plasmid episomally.
Cells of recombinant P. pastoris GS115 strains were analyzed light microscopically before and after induction to investigate if CGP inclusions are visible (see Fig. S2 in the supplemental material). As cells of the strain harboring CphA6308Δ1 synthesized higher amounts of the polymer, it was analyzed in parallel. All cells, independent of induction, exhibited light-scattering inclusions.
As the constructed vectors are also suitable for expression of cphA in E. coli BL21(DE3) strains under the control of the T7 promoter, the constructs harboring the different cphA genes were transferred to E. coli BL21(DE3)pLysS (Table (Table1)1) to compare the characteristics of CGP synthesis in recombinant yeast and bacteria. Induced cells were analyzed microscopically, and CGP was isolated to examine if CGP accumulation was similar to that of the P. pastoris strains. Unlike yeast, all cells accumulated the polymer exclusively in the insoluble form. Therefore, CGP accumulation could be visualized by light microscopy (see Fig. S2 in the supplemental material). When the strains harboring cphA6308, cphA6308Δ1, and cphA6308Δ3 were compared, a behavior similar to that of the recombinant yeast strains became obvious: Light-scattering inclusions were visible in some cells harboring cphA6308, while most cells expressing cphA6308Δ1 exhibited inclusions and almost no cells harboring cphA6308Δ3 showed inclusions. The microscopic analysis could be corroborated by determination of CGP contents: E. coli expressing CphA6308 accumulated 8.2% (wt/wt) CGP, whereas twice as much polymer was isolated from cells expressing CphA6308Δ1 (Table (Table44).
For optimization of CGP synthesis, different parameters were varied by using up to six stirred tank bioreactors in parallel, each with a maximal medium capacity of 2 liters. If not indicated otherwise, transgenic P. pastoris strain GS115 harboring pPICHOLI-3::cphA6308/his4 was cultivated in 1 liter of medium at 30°C and a pH of 5.0. One percent (vol/vol) methanol was added after depletion of glycerol, and feeding was repeated when required. CGP was isolated from 0.3 g of dry cells as described in Materials and Methods and analyzed by HPLC. No significant variations in amino acid composition were observed in the CGPs in comparison to the analysis of CGPs isolated from flask cultivation experiments. Interestingly, after all cultivations only soluble CGP could be isolated. All of the cell densities and cell CGP contents obtained are depicted in Fig. Fig.2.2. First, three different media, i.e., mineral salts media established by (i) Henes and Sonnleitner (28) and (ii) Invitrogen and (iii) the complex BMMY medium recommended by MoBiTec, were compared. Cells of P. pastoris strain GS115 harboring pPICHOLI-3::cphA6308 were cultivated in the presence of 100 μg/ml zeocin. As shown in Fig. Fig.2A,2A, the highest CGP contents of 4.4% were observed in cells grown in the medium described by Henes and Sonnleitner (28), compared to 3.2% CGP in cells grown in the other media. Therefore, this medium was used for further fermentations.
A subsequent experiment was carried out to determine the optimal length of the cultivation period. For this, cells of P. pastoris strain GS115 harboring pPICHOLI-3::cphA6308 were cultivated for 11 days and samples were withdrawn from the cultivation vessel daily to determine the CGP contents (Fig. (Fig.2B).2B). After 6 days of induction, the maximal polymer yield was observed; thereafter, CGP accumulation remained constant and therefore all further cultivations were performed with an induction period of 6 days.
In the next step, the cell densities and CGP contents of P. pastoris strain GS115 possessing or not possessing the episomally encoded his4 gene were compared after growth in the presence or absence of zeocin (Fig. (Fig.2C).2C). Higher product yields of 8.2% (wt/wt) were detected in cells encoding a functional his4 gene. Furthermore, explicit higher cell densities were obtained for this strain. Addition of zeocin resulted in slight stabilization of the plasmid in cells not expressing his4 and did not show an effect on the other strains.
Because the temperature and pH of the medium often play a crucial role in obtaining high product yields, temperatures ranging from 22°C to 32°C and pH values ranging from 3.0 to 7.0 were used (Fig. 2D and E). The parameters were set after the first induction with methanol. As shown in Fig. Fig.2D,2D, the highest CGP contents of 10.4% (wt/wt) were detected in cells cultivated at 26°C, while cells grown at 30°C accumulated 9.0% CGP; the lowest contents of 4.7% CGP were observed in cells grown at 22°C. Additionally, the CDMs obtained were higher after cultivation at 26°C (13.4 g/liter) than after cultivation at 30°C (12.5 g/liter). In further cultivations, the temperature was therefore set at 26°C during the induction phase. Through variation of the pH, further increases in CGP content were attempted; the lowest CGP amounts were observed after cultivation at pH 4.0, while the highest CGP content of 12.1% (wt/wt) was detected in cells grown at pH 6.0 or 7.0. As the highest cell density was obtained after growth at pH 6.0, this pH was set in all further cultivations.
To further increase the CGP contents of the cells, P. pastoris GS115 harboring cphA6308Δ1 was cultivated in comparison to the same strain harboring wild-type cphA6308. Additionally, to increase cell densities, cells harboring cphA6308Δ1 were fed twice with 2% glycerol before induction with methanol. The cell densities and CGP contents obtained are depicted in Fig. Fig.2F.2F. Cells expressing cphA6308Δ1 accumulated the polyamide up to 23.3% (wt/wt), compared to the 11.9% (wt/wt) CGP accumulated by cells harboring cphA6308. Furthermore, the cell density reached 60 g/liter (CDM) after glycerol-fed batch culture before induction with methanol, compared to the 17.3 and 17.9 g/liter observed for cells grown without additional glycerol feeding.
This report describes the first synthesis of CGP in a methylotrophic yeast. Besides P. pastoris, S. cerevisiae, accumulating CGP up to 15.3% (wt/wt) of its CDM, is the only yeast used for synthesis of the polyamide. As reported before, methylotrophic yeasts offer advantages over S. cerevisiae for the production of some proteins (22). This study indicated that P. pastoris represents a suitable candidate for synthesis of the polymer. A CGP content of 23.3% (wt/wt) is so far the highest content ever reported in eukaryotes. However, to make production of CGP a realistic industrial process, several parameters need to be optimized to enhance CGP accumulation. Therefore, different approaches might be useful. A first step was taken during this study through the generation and application of CphA6308Δ1, which exhibited 2.5-fold higher specific activity than CphA6308. Furthermore, metabolically engineered P. pastoris strains might result in increased product yields, as demonstrated previously for recombinant S. cerevisiae or bacteria (19, 54). Also, optimization of fermentation strategies, as shown for Acinetobacter baylyi and recombinant strains of Pseudomonas putida or Ralstonia eutropha (16, 18), might result in a significant increase in CGP accumulation. This study demonstrated that variation of only a few simple parameters can result in a significant increase in product yield (Fig. (Fig.2).2). However, the maximum achievable CGP content seems not to be reached by far, which becomes clear after comparison of the CGP contents of P. pastoris and E. coli, respectively. While expression of CphA6308Δ1 compared to CphA6308 resulted in an improved CGP accumulation of 30% in P. pastoris, E. coli synthesized twice as much CGP (Table (Table4).4). From this it was concluded that cells of P. pastoris lack specific substrates necessary for synthesis of the polymer. Another aspect for further optimization of CGP synthesis could be a detailed analysis of recombinant Muts or Mut− strains. Here, supply of mixed carbon sources might be beneficial to obtain high cell densities and a high product yield (15, 30, 31). As in S. cerevisiae, it was shown in P. pastoris that selection of the promoter controlling the expression of cphA is significant; while high CGP contents were observed in strains of S. cerevisiae expressing cphA under the control of the CUP1 promoter but not by using the GAL1 promoter (52), significant CGP contents were only observed in P. pastoris using the strong AOX1 promoter. The type of CGP synthesized by P. pastoris, concerning its solubility behavior and its molecular mass distribution, is comparable to that produced in S. cerevisiae; in both yeasts, mostly the soluble type of the polymer is accumulated in flask cultures (52, 54) and no insoluble CGP is detected after cultivation in fermentors (54a). Additionally, complete stabilization of the episomal plasmid encoding cphA was achieved during the present study through provision of a gene (his4) essential to the organism. Stabilization of cphA-encoding plasmids dependent on the addition of antibiotics provides problems during the cultivation of recombinant bacteria. The strategy of construction of an addiction system, as also used here, represents an effective method (35, 55).
An interesting aspect analyzed during this study was the comparison of CGP synthesis in yeast and in E. coli expressing identical cphA genes. The assumption that synthesis of soluble CGP is not dependent on the host or on the cphA source (21) could be partially confuted. Here, it was shown that synthesis of soluble CGP is dependent on the host organism, not on the CphA source. However, this cannot explain the occurrence of soluble CGP in E. coli observed by Ziegler et al. (59).
Nowadays, an interesting aspect is the amino acid composition of the synthesized polymer, as synthesis of dipeptides derived from CGP could find applications in pharmacy (45), and the production of a variety of these compounds with different compositions should be considered. Therefore, the comparably high fractions of lysine incorporated into CGP by P. pastoris make this organism a possible candidate for the production of lysine-rich CGP and an interesting and special candidate for industrial production of the polymer in comparison to other organisms. In addition, yeasts are especially promising hosts in this regard as they synthesize mainly the soluble form of the polymer, which exhibits a broader composition range in comparison to insoluble CGP (54). To further increase incorporated fractions of lysine, engineering of strains would be necessary, i.e., through directed changes in lysine metabolism, for example, by generating lysine-overproducing mutants (20). Furthermore, P. pastoris might represent a suitable candidate for the production of further CGP variants, as described previously in S. cerevisiae (54). Therefore, corresponding mutants defective in arginine metabolism, already described by Nett et al. in 2005 (41), should be used and analyzed. Production of CGP variants is of special interest for the synthesis of structurally different β dipeptides employed as pharmaceutical agents (46).
The conferred increased specific activities of CphA6308Δ1, CphA6308C595S, and CphA6308Δ1C595S, in comparison to that of wild-type CphA6308, were corroborated by increased CGP contents observed in recombinant strains of P. pastoris and E. coli. Residue C595 is part of the putative ATP-binding region in CphA; however, a mutation of this residue did not lead to loss of activity, indicating that C595 is not directly involved in the binding of ATP. After the mutation of residues K261 and K497, which are parts of ATP-binding motifs, complete loss of enzyme activity occurred (3, 4, 53). Furthermore, residue C595 seems not to be involved in the formation of a disulfide bond, as in this case a loss of activity would be expected due to misfolding of the enzyme. Increased CphA activity resulting from C-terminal truncation was previously observed by applying CphA from N. ellipsosporum NE1 consisting of 901 amino acids. (26, 27). Hai et al. (26) postulated that a CphA protein composed of 872 amino acids, corresponding to CphA6308Δ2, is optimal concerning high enzyme activity. However, in this study it was shown that truncation to a length of 873 amino acids led to the highest enzyme activities. The amino acid composition was not affected by the use of the different CphA proteins, leading to the assumption that the mutated or deleted residues do not play a role in substrate binding during catalysis, although the postulated amino acid substrate binding region includes the deleted residues (3, 53). It can be postulated that the last three C-terminal residues are not directly involved in the formation of the active site of CphA. Thus, detection and characterization of the putative substrate binding sites require further intensive research, which might open new ways for expansion of the substrate specificity of CphA. In conclusion, generation of CphA proteins with increased enzyme specific activities provides an important tool for increasing the CGP yields of various recombinant bacteria, yeast, or higher eukaryotes.
We thank H.-J. Galla (Institut für Biochemie, WWU Münster) for provision of P. pastoris strain KM71H and vectors pPICHOLI-3 and pPICHOLI-C. For provision of P. pastoris strain GS115, we thank G. Kunze (Leibniz-Institut für Pflanzengenetik und Kulturpflanzenforschung, Gatersleben). We are grateful for provision of S. cerevisiae strain G175 by M. Gustavsson, E. Wiberg, P. Stolt, and S. Stymne (Scandinavian Biotechnology Research, Alnarp, Sweden).
This project was partially supported by a grant (EOSLT02034) provided by SenterNovem (Utrecht, the Netherlands).
Published ahead of print on 28 December 2009.
†Supplemental material for this article may be found at http://aem.asm.org/.