Production of functionally active Penicillium chrysogenum isopenicillin N synthase in the yeast Hansenula polymorpha

Background β-Lactams like penicillin and cephalosporin are among the oldest known antibiotics used against bacterial infections. Industrially, penicillin is produced by the filamentous fungus Penicillium chrysogenum. Our goal is to introduce the entire penicillin biosynthesis pathway into the methylotrophic yeast Hansenula polymorpha. Yeast species have the advantage of being versatile, easy to handle and cultivate, and possess superior fermentation properties relative to filamentous fungi. One of the fundamental challenges is to produce functionally active enzyme in H. polymorpha. Results The P. chrysogenum pcbC gene encoding isopenicillin N synthase (IPNS) was successfully expressed in H. polymorpha, but the protein produced was unstable and inactive when the host was grown at its optimal growth temperature (37°C). Heterologously produced IPNS protein levels were enhanced when the cultivation temperature was lowered to either 25°C or 30°C. Furthermore, IPNS produced at these lower cultivation temperatures was functionally active. Localization experiments demonstrated that, like in P. chrysogenum, in H. polymorpha IPNS is located in the cytosol. Conclusion In P. chrysogenum, the enzymes involved in penicillin production are compartmentalized in the cytosol and in microbodies. In this study, we focus on the cytosolic enzyme IPNS. Our data show that high amounts of functionally active IPNS enzyme can be produced in the heterologous host during cultivation at 25°C, the optimal growth temperature for P. chrysogenum. This is a new step forward in the metabolic reprogramming of H. polymorpha to produce penicillin.

Background β-lactam antibiotics like penicillins and cephalosporins belong to one of the largest-selling classes of drugs worldwide with a production of forty-five thousand tons in the year 2000 [1]. Penicillins and cephalosporins are pro-duced by the filamentous fungi Penicillium chrysogenum and Acremonium chrysogenum, respectively, as well as some filamentous bacteria. These antibiotics possess as common structural motif the β-lactam ring [2]. Not surprisingly, the penicillin and cephalosporin biosynthetic pathways have the first two enzymatic steps in common. First, the non-ribosomal peptide synthetase δ(L-α-aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS) forms the tripeptide ACV. The formation of ACV acts as the committed step in both penicillin and cephalosporin biosynthesis. ACV is subsequently converted into isopenicillin N (IPN), which has the characteristic β-lactam backbone, by the enzyme isopenicillin N synthase (IPNS). Both ACVS and IPNS have been shown to be located in the cytosol in P. chrysogenum [3,4]. Subsequent replacement of the αamino adipoyl side chain of IPN by the more hydrophobic phenylacetyl or phenoxyacetyl moieties occurs in P. chrysogenum in the specific environment of the microbody and results in the formation of penicillin [3,5]. On the other hand, epimerization of the α-amino adipoyl moiety followed by ring expansion leads to cephalosporin biosynthesis in A. chrysogenum [6]. So far, a requirement for specific organelles for cephalosporin production in this filamentous fungus is unknown.
Isopenicillin N synthase (IPNS) belongs to a class of nonheme ferrous iron dependent oxidoreductases. During its enzymatic reaction one molecule of oxygen is completely transformed into two water molecules by removal of four hydrogen atoms from the ACV tripeptide [7]. Detailed mechanistic studies of the IPNS enzyme were carried out using Aspergillus nidulans IPNS produced in Escherichia coli [8,9]. These studies showed that the formation of the βlactam ring is carried out by an iron (IV)-oxy intermediate with the His212, Asp214, and His268 residues of A. nidulans IPNS forming the active center. Furthermore, a conserved Arg-Xxx-Ser oxidase motif is involved in binding of the ACV substrate [10].
Our objective is to introduce the penicillin biosynthesis pathway from P. chrysogenum into the methylotrophic yeast Hansenula polymorpha. Yeast species have the advantage of being versatile, easy to handle and cultivate, and possess superior fermentation properties relative to filamentous fungi. Additionally, introducing the penicillin biosynthesis pathway into this yeast species allows a better understanding of the function of microbodies in penicillin production. In the past, the microbody-localized proteins involved in penicillin biosynthesis in P. chrysogenum -isopenicillin N:acyl CoA acyltransferase (IAT) and phenylacetyl-CoA ligase (PCL) -were successfully produced in H. polymorpha in an active form [11,12]. Here we present our data on the expression of the P. chrysogenum pcbC gene encoding IPNS in H. polymorpha.

IPNS produced in H. polymorpha at 37°C is not stable
The P. chrysogenum pcbC gene encoding IPNS was cloned downstream of the strong, inducible H. polymorpha alcohol oxidase promoter (P AOX ) in pHIPX4 and was inte-grated at the P AOX locus in the H. polymorpha genome (see Methods). A strain, designated IPNS 4.2, containing the pcbC expression cassette was selected and cultivated at 37°C on methanol medium to induce the P AOX . Wild type H. polymorpha was used as control. To analyze the induction profile of IPNS protein, aliquots were taken at different stages of growth on methanol (Fig. 1, Panel A). Western blots were prepared using crude extracts of TCA precipitated cells and decorated with a polyclonal α-IPNS antiserum. This demonstrated (Fig. 1, Panel B) that strain IPNS 4.2 produces an α-IPNS specific protein of the expected size (approximately 37 kDa), that was absent in WT controls (data not shown). However, after the onset of the stationary growth phase IPNS levels declined rapidly (Fig. 1, Panel B). To check the stability of IPNS protein produced in H. polymorpha, cells of strain IPNS 4.2 were grown on methanol medium to mid-exponential growth phase, and then 0.5% glucose was added to the cultures, a treatment that results in complete repression of the P AOX . Subsequently, aliquots of equal volume were taken at regular time intervals and TCA precipitated. Western blot analysis showed that within 30 minutes after glucose addition the IPNS protein levels in the cells had decreased more than 50% (Fig. 2, Panel A), indicating that in H. polymorpha IPNS is not stable when produced at cultivation conditions physiological to the host.

Effect of the cultivation temperature on IPNS production in H. polymorpha
Instability of heterologous produced protein in yeast can be attributed to improper folding in the unnatural host, or to absence of activation, both of which can result in turnover of the protein [13]. Our assumption was that P. chrysogenum IPNS was not properly folding in H. polymorpha, which might be overcome by cultivating the recombinant strain at a reduced temperature. Since P. chrysogenum hyphae are usually grown at 25°C, we cultivated H. polymorpha strain IPNS 4.2 on methanol media at 30°C and 25°C again taking aliquots at different stages of growth ( Fig. 1, Panels C and E). Under these conditions, the growth rate of the H. polymorpha cells is reduced. Western blots demonstrated that IPNS protein was present in H. polymorpha extracts at all stages of growth ( Fig. 1, Panels D and F). We also analyzed the effect of the reduction in the growth temperature of strain IPNS 4.2 on the stability of IPNS. Cells grown at 30°C and 25°C on methanol media to the mid-exponential growth phase were shifted to 0.5% glucose in order to repress the P AOX and aliquots of equal volume were taken at regular time intervals after the shift. Western blot analysis demonstrated that IPNS protein was fully stable when produced at the lower growth temperatures (Fig. 2, Panels B and C).
In order to directly compare the amount of IPNS protein produced in H. polymorpha IPNS 4.2 at the different culti- (panels E and F). Aliquots were taken at different stages during cultivation at approximately equal optical densities (expressed as A 660 and indicated as 1, 2, 3, and 4 in panels A, C and E). IPNS is produced in strain IPNS 4.2 at all growth temperatures. However, during later stages of growth relatively low IPNS protein levels were observed in cells grown at 37°C, while at 30°C and 25°C enhanced levels of IPNS protein were detected. vation temperatures, equal amounts of extracts from cells exponentially grown on methanol were analyzed by Western blotting. Extracts of identically grown H. polymorpha wild type cells and P. chrysogenum cells grown on penicillin production medium supplemented with phenylacetic acid were used as controls. This revealed that the steady state levels of IPNS protein produced in H. polymorpha at 25°C are significantly enhanced relative to production at 37°C and 30°C (Fig. 3). Additionally, the amount of IPNS produced at 25°C in the heterologous host is comparable to that produced in a high penicillin producing strain of P. chrysogenum.

H. polymorpha IPNS 4.2 cells grown at reduced temperatures produce functionally active IPNS
We have determined the enzyme activity of IPNS produced in methanol grown cells of strain H. polymorpha IPNS 4.2 at different growth temperatures via a bioassay that utilizes the β-lactam sensitive bacterium Micrococcus luteus (ATCC 9341) as indicator strain [14]. Cells of the H. polymorpha host strain and P. chrysogenum served as controls. Crude extracts were prepared using glass beads as indicated in Methods and equal amounts of protein were incubated with substrate ACV. Subsequently, the resulting reaction product was used in the bioassay. When extracts prepared from the H. polymorpha strain IPNS 4.2 grown at 25°C and 30°C were used in this assay, this resulted in the formation of a zone of growth inhibition on the indicator plates (Fig. 4, Panel A). In contrast, extracts prepared from H. polymorpha WT and IPNS 4.2 grown at 37°C produced no inhibition zone. Also the use of higher amounts of extract from IPNS 4.2 cells grown at 37°C did not result in halo formation (data not shown). In order to check whether the halo formed by the reaction product of crude extracts of H. polymorpha IPNS 4.2 grown at 25°C and 30°C was caused by the synthesis of IPN, we added 50,000 IU of penicillinase to the reaction mixes and continued incubation at 25°C for an additional 30 minutes. The results (Fig. 4, Panel B) indicate that the product formed by the IPNS 4.2 extract is sensitive to penicillinase, confirming formation of the β-lactam structure. From this, we conclude that IPNS produced in H. polymorpha at 25°C and 30°C is functionally active. Cells of strain IPNS 4.2 were cultivated on methanol to induce IPNS production and prepared for immunocytochemistry using polyclonal α-IPNS antiserum. Immunogold labelling was only observed in the cytosol or on nuclear profiles (Fig. 5) as is characteristic for soluble proteins [15]. Microbodies and mitochondria invariably lacked specific labelling, which led us to conclude that IPNS produced in H. polymorpha is located in the cytosol.

Conclusion
In this study, we demonstrate that P. chrysogenum IPNS, produced in the methylotrophic yeast H. polymorpha, is stable and functionally active when cells of the heterologous host are cultured at 25°C. In contrast, IPNS protein produced at 37°C, the ambient growth temperature of H. polymorpha is unstable and inactive. Our data support that the instability of IPNS at the higher temperature is caused by poor/slow folding of the protein. Cultivation of cells at the lower temperature is thought to slow down protein synthesis, which allows more time for proper folding. Indeed, the growth rate of H. polymorpha at 25°C is reduced as compared to 37°C (Fig. 1, compare Panels A and E). Structural analysis of A. nidulans IPNS [8] indicated that the protein contains β-sheets which form an active hydrophobic core that is shielded from the hydrophilic environment by α-helical structures. Previously, it was shown in E. coli that the solubility of IPNS improved significantly by lowering the cultivation temperature [16]. Thus, poor folding at high temperatures may be an intrinsic feature of IPNS.
Previously, we succeeded in producing IAT and PCL in H. polymorpha and to target these proteins to the correct subcellular location, the microbody [11,12]. Both proteins showed enzymatic activity in H. polymorpha cells cultured at 37°C, implying proper folding at the physiological growth temperature of the heterologous host. Nevertheless, since both proteins originate from P. chrysogenum, optimal enzyme activities are expected at 25°C. The observation that active IPNS is only produced in H. polymorpha cells cultured at 25°C implies that penicillin production in this heterologous host may only be successful at reduced growth temperatures. Our next step on the way to insert the entire penicillin biosynthesis pathway into H. polymorpha will be the production of functionally active ACVS in the heterologous host. This undoubtedly is the most challenging step, since yeast genomes do not encode non-ribosomal peptide syn-  (well 2) and 25°C (well 3). As controls, H. polymorpha WT cells, cultivated on methanol at 25°C (well 5), and P. chrysogenum cells, cultivated for 40 h on penicillin production media supplemented with phenylacetic acid (well 1) were used. Cell free extracts (40 μg of protein each) were incubated for 20 min at 25°C in a buffer containing ACV to allow production of IPN. Panel A: Reactions were terminated by methanol addition. After protein precipitation, an equal volume of reaction product was added to each well of a bioassay plate on which M. luteus had been plated as indicator strain. The results indicate that extracts prepared from IPNS 4.2 cells cultivated at 25°C and 30°C contain active IPNS enzyme, resulting in the formation of a halo. In contrast, extracts from IPNS 4.2 cells grown at 37°C show no activity. Panel B: As control, the reaction products were further incubated in the presence of penicillinase for an additional 30 min at 25°C. After that, the reaction mixtures were terminated by methanol addition and the reaction products analyzed on a bioassay plate as described in A. The absence of halos on the plate demonstrates that the IPNS 4.2 extracts indeed allow formation of the β-lactam ring.
thetases like ACVS, while in bacteria and filamentous fungi these proteins are widespread.

Microorganisms and growth conditions
H. polymorpha strains used are derivatives of NCYC495 ade11.1 leu1.1 [17] and were grown at 37°C, 30°C or 25°C in either (i) rich complex media (YPD) containing 1% yeast extract, 1% peptone and 1% glucose, (ii) selective media containing 0.67% yeast nitrogen base without amino acids (DIFCO) supplemented with 0.5% glucose (YND), or (iii) mineral medium (MM) as described by Van Dijken et al. [18], supplemented with 0.25% ammonium sulphate using 0.5% glucose or 0.5% methanol as carbon source. For growth on plates, 2% granulated agar was added to the media. Whenever necessary, media were supplemented with 30 μg/ml leucine and 20 μg/ml adenine. For biochemical analysis, selected strains were precultured at least three times in MM containing glucose and subsequently shifted to MM containing methanol to induce expression of genes under the control of the P AOX .
A high penicillin producing P. chrysogenum strain (DSM anti-infectives, Delft, The Netherlands) was used as a control and was grown for 48 hours on a defined penicillin production medium supplemented with 0.05% phenylacetic acid [19].

Miscellaneous DNA techniques
All DNA manipulations were carried out according to standard methods [20]. H. polymorpha cells were transformed by electroporation [21]. DNA modifying enzymes were used as recommended by the supplier (Roche, Almere, The Netherlands). Pwo polymerase was used for preparative polymerase chain reactions (PCR). Oligonucleotides were synthesized by Life Technologies (Breda, The Netherlands). DNA sequencing reactions were performed at BaseClear (Leiden, The Netherlands). For DNA sequence analysis, the Clone Manager 5 program (Scientific and Educational Software, Durham, USA) was used.

Construction of H. polymorpha IPNS 4.2
For the construction of a pcbC expression plasmid a Hin-dIII site was introduced at the 5' end and a SalI site at the 3'end of P. chrysogenum pcbC by PCR with the primers ipns-F (5' ACTAAGCTTATGGCTTCCACCCCCAAG 3') and ipns-R (5' ATCGTCGACTCATGTCTGGCCGTTCTTGTTG 3') using DNA from a P. chrysogenum cDNA library as tem-plate [22]. The resulting 1002 bp DNA fragment was digested with HindIII and SalI and cloned between the HindIII and SalI sites of pHIPX4 [23] resulting in plasmid pHIPX4-pcbC. The nucleotide sequence of the cloned PCR fragment was confirmed by sequencing (data not shown). Subsequently, pHIPX4-pcbC was linearized with SphI in the P AOX region and transformed into H. polymorpha NCYC495 ade11.1 leu1.1 cells. Fast growing leucine prototrophic transformants were selected, which are expected to carry multiple copies of the expression cassette [24]. Correct integration was confirmed by colony PCR using the genome-specific primer Aox-F (5' TCACAC-CGTAACGCTTTATCGCC 3'), which corresponds to a region directly upstream of the AOX promoter and the plasmid-specific primer pcbC-R (5' ATCTGGTCCTGGT-GCTCCTTG 3'), which corresponds to the 5'end of the pcbC coding sequence. This will result in the formation of a product of 1898 bp exclusively in correct integrants (data not shown), one of which was designated IPNS 4.2.

Biochemical methods
Crude extracts of H. polymorpha cells were prepared with glass beads basically as described previously [25]. Crude extracts of P. chrysogenum hyphae were prepared from 150 mg of lyophilized mycelium [26]. For Western blots, extracts of H. polymorpha and P. chrysogenum cells were prepared using the TCA method [27]. Protein concentrations were determined using the Bio-Rad Protein Assay system with bovine serum albumin as standard. SDS-PAGE and Western blotting were performed by established procedures. For Western blots H. polymorpha cells of the indicated strains were cultivated on methanol-containing media to induce the expression of the AOX promoter. At the indicated time intervals, aliquots of cells were taken and TCA precipitated. Extracts were prepared for Western blotting and decorated with polyclonal IPNSspecific antibodies [4]. Equal amounts of protein were loaded per lane, unless indicated otherwise.

Bioassay for isopenicillin N synthase enzyme activity
Isopenicillin N synthase activity was determined in crude cell extracts prepared with extraction buffer (50 mM Tris-HCl, pH 8.0, 0.3 mM DTT, 1 mM PMSF) instead of phosphate buffer. Cell extracts corresponding to a fixed amount of protein (10 to 50 μg) were incubated at 25°C for 10 min in a final volume of 200 μl with 0.75 mM reduced Bis-ACV (Bachem Fein Chemicalien AG, Switzerland), 0.43 mM of FeSO 4 .7H 2 O as cofactor and 14.1 mM L-ascorbic acid as additive. After incubation, the reaction was terminated by addition of an equal volume of methanol to precipitate the proteins. After centrifugation, equal volume fractions of the supernatants were loaded in a well on an bioassay plate on which the indicator strain Micrococcus luteus ATCC 9341 had been plated. Plates were incubated overnight at 30°C. When penicillinase was used to degrade β-lactams in the reaction mixture, this was added (50,000 IU per reaction) prior to termination of the reaction with methanol.

Morphological analysis
Intact cells were prepared for immunocytochemistry as described previously [25]. Ultrathin sections of unicrylembedded cells were labeled using polyclonal antiserum raised in rabbit against IPNS and goat-anti-rabbit antibodies conjugated to gold according to the instruction of manufacturers (Amersham UK).