Such a method has been described in European Patent publication EP 0 448 180 A2.

In particular, it is stated in this publication that the production of secondary metabolites by an organism can be modulated by the combined actions of altering the number and/or size of organelles (more in particular microbodies) and the localisation of enzymes in relation to organelles. This statement is exemplified by the observation that the number of organelles can be influence by the level of expression of any protein present in said organelle. For example, the number of microbodies in ßlactam producing Penicillium chrysogenum was shown to be related to the level of expression of the enzyme acyl transferase-an enzyme localise in the microbodies.

The present invention is based on the finding that next to influencing the expression level of any protein present in said organelles, the homeostasis of organelles can also be modulated by influencing the expression of at least one gene which is directly related to organelle homeostasis.

Therefor, the present invention provides a method for modulating the production of secondary metabolites in the fungus Penicillium chrysogenum by modulating a least one gene which is directly related to organelle homeostasis in this fungus.

Modulating of genes as used herein comprises influencing the level of expression and or the targeting and or the primary sequence of the genes and encompasses classical selection and screening methods, herein also identified as classical genotypic modification, and more modern genetic modification techniques varying from relatively straightforward physical or chemical mutagenic treatment followed by several rounds of screening and selection to intricate recombinant technologies that can modify the level of expression, the targeting or the primary sequence of genes directly.

Homeostasis herein describes the subtle balance between organelle increase or biogenesis and organelle decrease, degradation or death. Hence, genes directly related to organelle homeostasis comprise genes who function essentially in regulating the biogenesis of organelles or in degradation or death of organelles. Accordingly, in a preferred embodiment of the present invention the organelle homeostasis is modulated by genetically engineering at least one gene encoding a protein which is responsible for organelle biogenesis (the expression and/or activity of which is to be increased) and or at least one gene encoding a protein which responsible for organelle degeneration (the expression and/or activity of which is to be decreased).

Preferably, the organelles of choice are the microbodies (peroxisomes, glyoxysomes).

Microbodies have a simple architecture and are surrounded by a single membrane. A specific feature of microbodies is that they are inducible; upon induction, the organelles multiply by growth and subsequent fission. However, microbodies lack DNA and a protein-synthesising machinery. All peroxisomal proteins identified so far are encoded by nuclear genes and with few exceptions are synthesised on free polysomes at their mature size. Hence, the essential topogenic information resides in the structure of the polypeptide.

Such genetic engineering of genes related to organelle homeostasis may involve effecting either the level of expression of the instant genes or a genetic modification of the gene itself so as to result in a protein with a higher activity in organelle biogenesis or in a lower activity in organelle degradation.

Hitherto, no proteins or genes coding for proteins related to organelle homeostasis in Penicillium chrysogenum were known. Attempts to identify such genes in Penicillium chrysogenum based on sequence homology, hence by applying polynucleotide probes derived from known sequences of proteins or genes with similar functions in other organisms failed.

Surprisingly, a novel protein and the polynucleotide encoding such protein in Penicillium chrysogenum could be identified. The present invention is therefor also related to a novel protein (Pex11 p) active in the biogenesis of organelles, in particular in the biogenesis of microbodies, and which is represented by the amino acid sequence SEQ ID NO: 2.

The present invention is also related to isolated polynucleotides encoding such protein Pex11 p, and in particular to the polynucleotide pex11 which is represented by SEQ ID NO. 1.

Such a polynucleotide can be incorporated in an expression vector suitable for transformation of the Penicillium chrysogenum and which effects upon transformation expression of the particular polynucleotide incorporated therein.

Expression vectors as well as methods for constructing recombinant expression vectors and transformation techniques useful according to the present invention are well known in the art.

The present invention further provides Penicillium chrysogenum comprising a modulated gene or functional fragment thereof related to organelle biogenesis, and more in particular Penicillium chrysogenum comprising a modulated gene encoding the protein Pex11 p as exemplified by SEQ ID NO: 1 or a functional fragment thereof.

In another example, the invention provides Penicillium chrysogenum comprising a modulated gene or functional fragment thereof related to organelle degradation.

Decreasing organelle degradation serves the same purpose as increasing organelle biogenesis; at a given moment in time, a fungal cell as provided by the invention has more organelles available by or through which a secondary metabolite can be produced, thus optimising the yield of said substance. For example, degradation can be regulated by modifying genes involved in peroxisome degradation, for example wherein said gene is a vps15 gene or a gene functionally related thereto, or wherein said gene is a pdd1 gene or gene functionally related thereto. Microbody degradation can be regulated or influenced at various levels, for example at the level of signalling of individual cells for degradation followed by vacuolisation of the organelles destined for degradation and degradation of organelle content by vacuolar hydrolases.

According to the present invention, the production of secondary metabolites by P. chrysogenum can be modulated. Such secondary metabolites comprise/ ?-tactam compounds such as penem and cephem compounds. More in particular, these micro-

organisms produce penicillin or core structures thereof (e. g. 6-APA, 6-amino penicillanic acid) or cephalosporin or compounds containing core structures thereof (e. g. 7-ADCA, 7- aminodesacetoxy cephalosporanic acid and 7-ACA, 7-amino cephalosporanic acid) Penicillinproduction is lower in microbody-deficient strains A batch of conidiospores from Penicillium strain Wisconsin54-1255 was mutagenised using nitrochinoline oxide. A selected batch of mutagenised spores (survival 1-10 %) was replia- plated on a complex rich medium and on a medium containing oleic acid (0.1 %) as the sole carbon source. Mutants that were unable to utilise oleic acid were further tested for growth on other fatty acids: lauric acid, and hexanoic acid. From these tests, five mutants were selected that were generally deficient in fatty acid catabolism. These mutants were subsequently grown in a medium supporting the proliferation of peroxisomes. After 5 days of growth in this medium, samples were prepared for Electron Microscopy. Coupes were cut for Thin Section EM. After analysis of the slides, 3 mutants appeared to be deficient in peroxisome formation. These mutants were subsequently grown in a medium supporting penicillin formation. The microbody-deficient strains were heavily affected in penicillin formation and produced only 10-30 % of the penicillin level of the host strain, Wisconsin54- 1255. (Figure 5).

It is concluded that for an adequate penicillin production the peroxisomal volume should be maintained at the highest possible level.

Peroxisome volume fraction is proportional to penicillin production.

Strain Wisconsin54-1255 was cultivated in shake flasks on minimal medium containing lactose as carbon source and phenoxy acetic acid as precursor for penicillinV. The culture was incubated for seven days at 25 degrees Celsius and 280 rpm. Each day a sample was analysed for penicillin production, carbon consumption and the peroxisomal volume fraction of cells. This way the specific penicillin production rate could be determined. To determine the peroxisomal fraction cells were fixated in KMn04. Slides for electron microscopy were labelle with antibody against, AT which is exclusively localised in peroxisomes (buller, 1991). By screening a lot of individual cells the volume of peroxisomes could be estimated.

The penicillin production rate showed a positive relationship with the peroxisomal volume fraction (Figure 6).

This strengthens the conclusion from the previous paragraph that penicillin production is directly caused by an increased volume of peroxisomes.

Peroxisomal volume is higher in strains with increased penicillin production.

The Penicillium strain Wisconsin54-1255 and a strain with a increased penicillin production (DS04825-Gist-brocades, The Netherlands) were cultivated in a medium containing lactose and phenoxy acetic acid. After 5 days of growth at 25 degrees Celsius these cultures were harvested. Part of the samples were prepared for electron microscopy, while the remaining were used to determine the penicillin production rate. The slides for EM were labelle with AT antibody to stain the peroxisomes. Several individual cells were scored to calculate the peroxisomal volume fraction. The cells of strain DS04825 contained a substantial higher volume fraction of peroxisomes, which coincided with an increased penicillin production rate (Figure 7).

This supports the conclusion that in order to obtain strains with higher penicillin production Penicillium chrysogenum can be modified so as to produce higher volumes of peroxisomes.

Vacuolar degradation of peroxisomes during decreasing penicillin production rates Penicillium chrysogenum was cultivated in shake flasks on minimal medium containing lactose as carbon source to create fl-lactam producing conditions. The culture was incubated for seven days at 25 °C and 280 rpm. Each day a sample was analysed for penicillin production and carbon consumption to calculate the specific ßlactam production rate. At the same time points samples were fixated in KMn04 and labelle with AT antibody which is exclusively localised in peroxisomes (Muller, 1991) to monitor the peroxisomes. By screening a lot of individual cells using electron microscopy it was obvious that during the latter stages of the fermentation, when the production rate is decreasing, the vacuoles of P. chrysogenum fuse with peroxisomes (see Fig 9). Via this mechanism the number and volume of peroxisomes decreases fast, which is reflected by a lower penicillin production rate. In conclusion, the decrease in penicillin production rate will be minimised by maintaining peroxisome homeostasis (e. g. prevent the degradation of peroxisomes).

The invention is further illustrated by the following examples.

Isolation of the Penicillium chrysogenum pex11 gene and effect of deletion and over- expression of Pex11 p on peroxisome function.

Materials & Methods Strains A high yielding Penicillium chrysogenum strain, e. g. DS04825 or P2, was cultured in penicillin production medium (Theilgaard et al., 1997') Escherichia coli DH5a (Sambrook et a/., 19892) was grown in LB medium supplemented with the appropriate antibiotics.

Genetic modification Recombinant DNA manipulations were as described by Sambrook et a/. (1989).

DNA sequencing was performed at the sequencing facility of BASECLEAR (Leiden, The Netherlands). Polymerase chain reaction-mediated DNA amplification was performed using Pwo polymerase or the EXPANDT" High Fidelity PCR system according to the instructions of the supplier (Boehringer, Mannheim, Germany).

Purification of peroxisomal membrane proteins from P. chrysogenum P. chrysogenum hyphe were protoplasted by incubation for 1.5 h in KC buffer (6 % KCI, 0.2 % citric acid) containing 2 mg/ml Novozym. Protoplasts were harvested, washed with a 1: 1 mixture of KC buffer and buffer A (1.2 mol/I sorbitol, 5 mmol/I MES, 1 mmol/I MgCI2,1 mmol/I EDTA, pH 5.5) followed by a wash with buffer A. Protoplasts were homogenised in buffer A supplemented with 1 mM PMSF. Differential centrifugation was used to separate the different organelles present in the homogenate (10 min 4,000 g, 10 min 6,000 g, 20 min 30,000 g). The final (30,000 g) pellet, containing mainly peroxisomes and mitochondria, was loaded on a discontinuous sucrose gradient (in 5 mmol/I MES, 1 mmol/I MgCI2,1 mmol/i EDTA, pH 5.5) and centrifuged for 2 h at 30,000 g. Peroxisomal peak fractions, as determined by catalase activity measurements, were pooled and purified on a second discontinuous sucrose gradient. Subsequently, purified peroxisomes were incubated in a buffer containing 1 mol/l NaCI and 1 mol/I Tris, and peroxisomal membranes were obtained 'Theilgaard et al., Biochem J. (1997) 327: 185-191

by centrifugation at 200,000 g. The proteins in the 200,000 g pellet fraction were solubilised in SDS-PAGE loading buffer, separated on 10 % SDS-PAA gels and blotted onto PVDF membranes which were stained with Coomassie Brillant Blue. Protein bands were cut from the membrane and used for microsequencing (Eurosequence, Groningen, The Netherlands). A dominant 23 kDa protein band gave an N-terminal sequence that was similar to the peroxisomal membrane protein Pex11 p from Saccharomyces cerevisiae (Erdmann and Blobel, 19953; Marshall et al., 19954) and Candida boidinii (Sakai et al., 19955). (Fig. 1).

Cloning of a cDNA encoding P. chrysogenum Pex1 1p Based on the N-terminal sequence of the 23 kDa P. chrysogenum peroxisomal membrane protein and its similarity to S. cerevisiae and C. boidinii Pex11 p, we designed primers for use in the polymerase chain reaction (PCR): PEX11-F1 (SEQ ID NO: 7), PEX11-F2 (SEQ ID NO: 8) and PEX11-R (SEQ ID NO: 9). As template in PCR reactions DNA was used from a P. chrysogenum cDNA library constructed in plasmid pCMV-SPORT4 (Custom made by Life Technologies, Inc., Frederick, Md, USA). cDNAs were cloned as 5'Sall-Notl 3'fragments between the Sall and Notl sites in the pCMV-SPORT4 vector.

To obtain DNA fragments carrying a portion of pex11, PCR was performed with the peul 1- specific primers and the M13/pUC universal and reverse sequencing primers. PCR with the PEX11-F1/universal primer combination resulted in the synthesis of a highly specific fragment of approximately 1 kb. This fragment was digested with Notl, partly filled-in and cloned between the Notl and Smal sites of pBluescript 11 SK+ (Stratagene, San Diego, CA, USA), and sequenced. Analysis of three clones revealed that the PCR fragment contained an open reading frame (ORF) lacking a startcodon that showed significant similarity to S. cerevisiae and C. boidinii Pex11 p (27 % and 26 % identity, respectively), as well as approximately 150-250 bp of nontranslated DNA in addition to a polyA tail. To clone the complete ORF encoding Pex11 p, we designed the PEX11-STOP primer (SEQ ID NO: 10).

This primer is located 3 bp 3'of the stop codon of the putative pex11 ORF. PCR was performed with the PEX11-STOP primer in combination with the M13/pUC reverse 2 Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning, a Laboratory manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 3 Ermann, R. and Blobel, G. (1995) J. Cell Biol. 128,509-523 4Marshall, P. A., Krimkevich, Y. I., Lark, R. H., Dyer, J. M., Veenhuis, M, and Goodman, J. M. (1995) J.

Cefl BioL 129,345-355 5 Sakai, Y., Marshall, P. A., Saiganji, A., Takabe, K., Saiki, H., Kato, N., and Goodman, J. M. (1995) J.

Bacteriol. 177,6773-6781

sequencing primer using DNA from the P. chrysogenum [A] cDNA library. A highly specific DNA fragment of approximately 0.8 kb was obtained, which was digested with Sall, partly filled-in and cloned between the Sall and Smal sites of the polylinker of pUC19 (Yanisch- Perron et a/., 19856). To be absolutely sure that the PCR had not introduced errors in the DNA sequence, two clones slightly differing in length were sequenced. Sequence analysis revealed that both clones contained an identical ORF representing the P. chrysogenum pex11 gene. Upstream of the ATG of the ORF either 4 (clone 12) or 43 bp (clone 18) were present in addition to the linker that was used to clone the cDNAs in the pCMV-SPORT4 vector (for the sequence of clone 12 see Fig. 2). An alignment between the deduced amino acid sequence of the putative Pex11 p and its homologues from S. cerevisiae and C. boidinii is shown in Fig. 3.

Over-expression of pex11p in P. chrysogenum For over-expression of pex1 1p in P. chrysogenum the gene is for example placed under the control of the P. chrysogenum IPNS promoter. To this purpose the expression vector pGBRH2 was constructed (Fig. 4): Two PCR fragments containing the P. chrysogenum IPNS promoter and P. chrysogenum AT terminator regions were isolated by PCR using the primer combinations IPNS #1 (SEQ ID NO: 5) plus IPNS#2 (SEQ ID NO: 6) and AT #1 (SEQ ID NO: 3) plus AT#2 (SEQ ID NO: 4), respectively. The ATterminator fragment was digested with Smal and Notl and cloned between the Smal and Notl sites of pBluescript 11 KS+ (Stratagene). Subsequently, the IPNS promoter fragment, digested with Asp718 and Hindlll, was cloned between the Asp718 and Hindlll sites of the polylinker of the resulting plasmid. The P. chrysogenum pex11 gene is inserted as a Sa/l (blunted)-EcoRl fragment between the BamH1 (blunted) and EcoRl sites of the polylinker of pGBRH2, resulting in plasmid pGBRH2-PcPEX11. The expression plasmid is digested with Notl to release the PPNS-FcPEX11-TAT cassette. This fragment is co-transformed with a DNA fragment carrying the Aspergillus nidulans AMDS gene as a selection marker into various P. chrysogenum strains. Transformants were selected on acetamide agar and were screened for overproduction of Pex11 p using Western blotting with polyclonal antibodies. Figure 8 shows a clear example of the ultrastructural changes in the cell upon introduction of addition Pex11 p protein; a huge proliferation of the amount of peroxisomes is observed.

All these peroxisomes were shown to be functional; e. g. the contain acyltransferase (AT). The total amount of AT was equal to the parent strain (data not shown); therefore 6 Yanisch-Perron, C., Vieira, J., and Messing, J. (1985) Gene 33,103-119

we can conclude that the effect on penicillin production is solely caused by the number and volume of peroxisomes. Strains containing this enormous amount of peroxisomes showed a 2.1 fold increase in penicillin titre.

Cloning of genomic DNA containing pex11 Using the pex11 cDNA of P. chrysogenum as a probe allows to isolate the genomic region containing pex11. This allows the construction of mutants lacking pex11 (so-called p ex1 1strains) by disruption (single crossing-over) or deletion of (part on the gene (double crossing-over). pex11strains are analysed using ultrastructural (quantification of number/volume fraction of peroxisomes) and biochemical (production of penicillin) techniques.

Production of polyclonal antiserum against Pexl lp For the generation of antibodies against Pex11 p we construct a plasmid that allows synthesis in E coli of a fusion protein between the maltose binding protein and Pex11 p. To this effect the pex11 cDNA is isolated from clone #12 as a Sall-Smal fragment and cloned between the Sall and Hind I I l (blunted) sites of the pMal-C2 polylinker. A 640 amino acid fusion protein is produced by this plasmid. The fusion protein is isolated using the Protein Fusion and Purification System as described by New England Biolabs (Beverly, MA, USA) and subsequently used to immunise rabbits.

Legends to the figures Fig. 1. Sequence similarity between the N-terminal amino acid sequence of the dominant 23 kDa P. chrysogenum peroxisomal membrane protein and Pex11 p from Candida boidinii (Cb-Pex11 p) and Saccharomyces cerevisiae (Sc-Pex11 p). Identical residues are indicated by an asterisk, conservative replacements by a dot. Question marks indicate amino acids that could not be assigned with 100 % certainty during sequencing.

Fig. 2. Nucleotide sequence of clone 12 containing P. chrysogenum pex11 cDNA and the deduced amino acid sequence of Pex11 p. The linker used to clone the cDNAs in pCMV- SPORT4 as well as the PEX11-STOP primer used to isolate the cDNA are indicated.

Fig. 3. Multiple sequence alignment between the amino acid sequences of Pex11 p of P. chrysogenum (PCPEX11 P; 238 residues), Candida boidinii (CBPEX11 P; 256 residues) and Saccharomyces cerevisiae (SCPEX11 P; 236 residues). The N-terminal amino acid sequence of the 23 kDa peroxisomal membrane protein is also indicated. Identical residues are indicated by an asterisk, conservative replacements by a dot.

Fig. 4. The P. chrysogenum expression vector pGBRH2.

Fig. 5. Penicillin production rate is seriously hampered in microbody-deficient strains mut 1, 2 or 3, as compared to parent strain Wisconsin54-1255.

Fig. 6. Penicillin production rate versus peroxisomal cell volume in Wisconsin54-1255.

Fig. 7. Penicillin production rate versus peroxisomal fraction.

Fig. 8. Strong proliferation of microbodies in a Pex11p overproducing Penicillium chrysogenum strain as compared to a wild type strain (Wisconsin54-255) is shown.

Fig. 9. Electron microscopic observation of fusion of the vacuoles of P. chrysogenum fuse with peroxisomes during the latter stages of the fermentation, when the production rate is decreasing.