The mechanisms of compartmentalization of intermediates and secretion of penicillins and cephalosporins in β-lactam antibiotic-producing fungi are of great interest. In Acremonium chrysogenum, there is a compartmentalization of the central steps of the CPC (cephalosporin C) biosynthetic pathway. In the present study, we found in the ‘early’ CPC cluster a new gene named cefP encoding a putative transmembrane protein containing 11 transmembrane spanner. Targeted inactivation of cefP by gene replacement showed that it is essential for CPC biosynthesis. The disrupted mutant is unable to synthesize cephalosporins and secretes a significant amount of IPN (isopenicillin N), indicating that the mutant is blocked in the conversion of IPN into PenN (penicillin N). The production of cephalosporin in the disrupted mutant was restored by transformation with both cefP and cefR (a regulatory gene located upstream of cefP), but not with cefP alone. Fluorescence microscopy studies with an EGFP (enhanced green fluorescent protein)–SKL (Ser-Lys-Leu) protein (a peroxisomal-targeted marker) as a control showed that the red-fluorescence-labelled CefP protein co-localized in the peroxisomes with the control peroxisomal protein. In summary, CefP is a peroxisomal membrane protein probably involved in the import of IPN into the peroxisomes where it is converted into PenN by the two-component CefD1/CefD2 protein system.
Very little is known about the subcellular localization of enzymes involved in the biosynthesis of secondary metabolites in fungi [1,2]. Acremonium chrysogenum is a filamentous fungus used for industrial production of β-lactam antibiotics that contains the cephem nucleus. CPC (cephalosporin C) biosynthesis begins with the condensation of the three precursor amino acids L-α-aminoadipic acid, L-cysteine and L-valine to form the tripeptide ACV (δ-L-α-aminoadipyl-L-cysteinyl-D-valine) [3,4]. This reaction is carried out by ACV synthetase , which is encoded by pcbAB . Later, oxidative ring closure of the ACV tripeptide by the IPN synthase encoded by pcbC  leads to formation of a bicyclic ring constituted by the four-membered β-lactam ring fused to the five-membered thiazolidine ring. The resulting compound, IPN (isopenicillin N), is the first compound in the biosynthetic pathway with antibiotic activity.
The conversion of IPN into PenN (penicillin N) in A. chrysogenum involves the activity of an IPN-CoA synthetase encoded by cefD1 and an IPN-CoA epimerase encoded by cefD2 . Both enzymes catalyse the isomerization of the L-α-AAA side chain of IPN to the D-enantiomer to give PenN.
After the epimerization step, a bifunctional enzyme with expandase and hydroxylase activities (encoded by cefEF)  converts PenN into deacetoxycephalosporin C and then into DAC (deacetylcephalosporin C). As a result, the five-membered thiazolidine ring of penicillin is replaced by a six-membered dihydrothiazine ring forming the cephem nucleus. Finally, DAC is converted into CPC by the DAC-acetyltransferase that uses acetyl-CoA as the donor of the acetyl group, encoded by the cefG gene .
Whereas in Penicillium chrysogenum there is a compartmentalization of the penicillin biosynthetic pathway (reviewed in ) between the cytosol and the peroxisomal lumen, in A. chrysogenum previous studies proposed that probably all CPC biosynthetic enzymes are cytosolic [12,13]. However, the amino acid sequences of the two-component IPN epimerase system (CefD1 and CefD2) contain PTS1 (peroxisomal targeting signal 1) sequences, suggesting that the epimerization step takes place in the peroxisomal matrix .
The distinct subcellular localization of the β-lactam biosynthetic enzymes in both filamentous fungal species implies transport of enzymes, precursors, intermediates and products through these compartments. However, although a good biochemical and genetic knowledge has accumulated on penicillin and CPC biosynthesis [1,2], there is not enough information about the systems involved in such transport.
Recently, our group has characterized the cefM gene of A. chrysogenum, which encodes a membrane protein of the MFS (major facilitator superfamily) that is located in small microbodies (probably peroxisomes). CefM protein seems to be involved in the translocation of the intermediate PenN from the peroxisomal lumen to the cytosol .
The question of how IPN is transported from the cytosol to the peroxisomal matrix remains unanswered. In order to search for proteins involved in IPN transport, we studied the DNA region located upstream of cefT . In the present paper, we report the characterization of a new gene encoding a membrane protein essential for CPC biosynthesis located in the cluster of early CPC genes. We provide evidence for its localization in peroxisomes and propose a role in IPN transport across the peroxisomal membrane.
Micro-organisms, culture media and antibiotic determination
A. chrysogenum C10 (ATCC 48272), a high-CPC-producing strain provided by PanLab, was used as the parental strain in the present study. All other strains were derived from A. chrysogenum C10; for sporulation they were grown in LPE (Le Page and Campbell) medium  for 7 days at 28 °C. Spores and mycelium fragments collected from six plates of LPE culture medium were inoculated into 100 ml of seed medium  in 500 ml shake-flasks and incubated at 25 °C for 48 h in an orbital incubator at 250 rev./min. A 10 ml aliquot of this seed culture was used to inoculate 100 ml of DP (defined production) medium  in 500 ml triple-baffled flasks (BellCo) and incubated at 25 °C in a rotary shaker at 250 rev./min. Samples were taken every 24 h. β-Lactam antibiotic production was assayed against Escherichia coli ESS2231 (a β-lactam-supersensitive test strain) and determined by HPLC as described previously by Ullán et al. .
DNA isolation and Southern blotting
Genomic DNA of A. chrysogenum was isolated as described previously . Genomic DNA samples (3 μg) from A. chrysogenum C10 and its transformants were digested with restriction enzymes and separated in 0.7% agarose gels. Digested total DNA was separated by agarose gel electrophoresis and blotted on to nylon membranes (Hybond NX; Amersham Pharmacia Biotech) . Southern blot hybridization was performed as described previously .
Total RNA was isolated from A. chrysogenum C10 mycelia with the RNeasy kit (Qiagen) as described previously .
DNA sequencing and intron analysis
DNA sequencing reactions were prepared by standard procedures  and automatic sequencing was performed using the AutoRead™ system (Pharmacia). To elucidate the presence of putative introns in the DNA sequence of cefP, the DNA region containing the expected intron splicing sites was amplified by RT (reverse transcription)–PCR (Promega) using RNA of A. chrysogenum 48 h cultures as template with the following primer pairs: INT-CP-DIR (5′-GCGAATGCGACCCCGAGGAGTA-3′) and INT-CP-REV (5′-TCGCAACAAAGAAGTAGGTGAAGA-3′); and INT-CP(B)-DIR (5′-CTACTACTTCGGGCAGCGGTT-3′) and INT-CP(B)-REV (5′-ATGTTATTTTCGTCAGTGTCC-3′).
The amplified regions were sequenced to confirm the presence of the three introns.
PCR and RT–PCR analysis
Genomic DNA amplification of the cefR gene was performed by PCR using the oligonucleotides R1R (5′-ATGCAGC-CCAGGTTATAA-3′) and RF5 (5′-ATTCATGTCACAAG-CCCC-3′) as specific primers. Expression of cefP and cefR was tested by RT–PCR using the primer pairs INT-CP-DIR/INT-CP-REV and RF6 (5′-ATAGCAGGGATGGCGACAG-3′)/R4R (5′-ACACCATCCGAAAGCACA-3′) respectively.
In vitro mutagenesis was performed with the QuikChange® site-directed mutagenesis kit (Stratagene) following the manufacturer's instructions. Oligonucleotides A1 (5′-GATTGGA-GAGAGATCTGATGATCTGGTGG-3′) and A2 (5′-CCACCA-GATCATCAGATCTCTCTCCAATC-3′) were used to introduce a BglII site into the cefP gene.
Plasmid constructions containing cefP and/or cefR
pCP comprises a BamHI fragment of 7.2 kb bearing the cefP gene under the control of its own promoter cloned into the BamHI site of plasmid pJL43 .
A BglII restriction site was introduced by in vitro mutagenesis into plasmid pCP. The cefP gene was inactivated by insertion of the hph (hygromycin B phosphotransferase) hygromycin-resistance cassette (subcloned from pAN7–1; ) into this BglII site. Plasmid pDP contains also the phleomycin-resistance gene ble under the control of the P. chrysogenum pcbC promoter as a second selective marker.
This plasmid carries a NotI fragment of 5.3 kb containing the cefR gene under the control of its own promoter cloned into the NotI site of the plasmid pBluescript II SK+ (Stratagene).
This plasmid contains a BamHI fragment of 8.5 kb bearing both the cefP and cefR genes under the control of their own promoters cloned into the BamHI site of pBluescript II SK+.
pcefP-DsRed (Discosoma sp. red fluorescent protein)
To obtain the fused cefP–DsRed gene, a BglII/StuI 2.7 kb DNA fragment obtained from plasmid pDP by PCR using the ST3F (5′-GGAAGATCTATGTTCGGATCTAGAGATGGC-3′) and T3DR-R (5′-AAAAGGCCTAATGTTATTTTCGTCAGTGTC-3′) oligonucleotides was inserted into the BglII/SmaI site of pEXpDsRed bearing the DsRed gene from Discosoma sp. (Clontech) flanked by the gdh gene promoter from Aspergillus nidulans and the terminator of the cyc1 gene from Saccharomyces cerevisiae.
This plasmid contains the EGFP [(enhanced GFP (green fluorescent protein)]–SKL (Ser-Lys-Leu) gene which encodes a protein targeted to peroxisomes. To construct p43EFGP-SKL, a NotI 2.1 kb fragment from pGBRH2-EGFP-SKL  was inserted into the NotI site of pJL43.
Transformation of A. chrysogenum protoplast
A. chrysogenum protoplasts were obtained and transformed as described previously . Transformants were selected in TSA (tryptic soy agar; Difco) with sucrose (10.3%) supplemented with phleomycin (10 μg/ml) or hygromycin B (30 μg/ml).
cefP gene disruption
Plasmid pDP was transformed into A. chrysogenum C10, and transformants were selected by their hygromycin B resistance. Transformants showing a hygR phleS phenotype, indicating that double recombination had occurred, were selected.
Integration of the hybrid cefP–DsRed gene and subcellular localization of the hybrid protein
Protoplasts of A. chrysogenum C10 were transformed with the integrative plasmids pcefP-DsRed and p43EGFP-SKL (the latter is a derivative of pGBRH2-EGFP-SKL ). Plasmid pcefP-DsRed carries the hybrid cefP–DsRed gene under the control of the promoter of the A. nidulans gdh gene and the terminator of the cyc1 gene of S. cerevisiae. On the other hand, plasmid p43EGFP-SKL carries the EGFP gene coupled with a SKL sequence (for its targeting to peroxisomes) under the control of the pcbC gene promoter and the penDE gene terminator of P. chrysogenum. In addition, this plasmid carries the phleomycin-resistance cassette for transformant selection.
For the subcellular localization of the hybrid protein CefP–DsRed, spores of transformant TPDsRed-32 were inoculated in CCM (complex culture medium)  and incubated for 3 days at 25 °C and 175 rev./min for its adequate germination and growth. This seed culture was used to inoculate DP medium (10% inoculum) and incubated for 72 h at 25 °C and 250 rev./min.
The fluorescence emissions of hyphae were analysed by confocal laser-scanning microscopy using a Radiance 2000 laser confocal microscope (Bio-Rad Laboratories). Green (EGFP) and red (DsRed) fluorescent proteins were visualized with a number 13 filter (λex 470 nm, 20 nm bandwidth; λem 505–530 nm for GFP and λex 556 nm, 20 nm bandwidth; λem 586 nm for DsRed).
Identification of the cefP gene downstream of cefT
A SmaI fragment (4.6 kb) of the DNA region located downstream of cefT (Supplementary Figure S2 at http://www.BiochemJ.org/bj/432/bj4320227add.htm) was cloned into the pBluescript II KS+ plasmid in both orientations, giving rise to plasmids pP1a and pP1b. The inserts of pP1a and pP1b were sequenced completely on both strands. The nucleotide sequence was deposited in the EMBL database under the accession number AM231816 (cefP). Analysis of the nucleotide sequence of the 4.6 kb DNA insert revealed the presence of one open reading frame named cefP (because it encodes a peroxisomal protein; see below). The cefP gene is 2769 nt long and is interrupted by the presence of three introns. The presence of the introns was confirmed by RT–PCR as described in the Experimental section. cefP encodes a protein of 866 amino acids with a deduced molecular mass of 99.2 kDa. The amino acid sequence of CefP showed strong similarity throughout its entire length to uncharacterized integral membrane proteins (Supplementary Figure S1 at http://www.BiochemJ.org/bj/432/bj4320227add.htm) of Nectria haematococca (63% identical amino acids), Gibberella zeae (60% identical amino acids), Podospora anserina (52% identical amino acids), Chaetomium globosum (50% identical amino acids), Neurospora crassa (49% identical amino acids) and Coccidioides immitis (41% identical amino acids). Furthermore, we found in all proteins one DUF (domain of unknown function) 221 motif (Pfam accession number pfam02714) that occurs in a superfamily of hypothetical transmembrane proteins of which none has any known function (http://pfam.ccbb.re.kr/cgi-bin/getdesc?name=DUF221).
Interestingly, there were no proteins with high similarity in the genomes of the benzylpenicillin producers P. chrysogenum or A. nidulans (which do not produce cephalosporins), suggesting that this protein is likely to be specific for CPC biosynthesis.
To determine the number of TMSs (transmembrane spanners), the deduced CefP protein was analysed with the algorithms of five different programs: SOSUI  (http://bp.nuap.nagoya-u.ac.jp/sosui/sosui_submit.html), TopPred2  (http://bioweb.pasteur.fr/seqanal/interfaces/toppred.html), TMHMM  (http://www.cbs.dtu.dk/services/TMHMM/), TMPRED  http://www.ch.embnet.org/software/TMPRED_form.html) and HMMTOP  (http://www.enzim.hu/hmmtop/html/submit.html). Results showed nine putative TMSs with the SOSUI and TopPred2 bioinformatic tools, whereas the TMPRED algorithm predicted nine or ten TMSs. Nevertheless, TMHMM and HMMTOP programs increased the number of putative TMSs to 11. The evaluation of these methods by Möller et al.  revealed that the best algorithms are those of the HMMTOP and TMHMM programs. In summary, the CefP protein contains from nine to eleven TMSs, with 11 being the most probable number (Supplementary Figure S1 shows the location of every TMS). The number of amino acids in each TMS ranges from 17 to 23.
Analysis of Pex19 (peroxisome biogenesis factor 19) binding sequences (http://www.peroxisomedb.org/) in the CefP protein revealed one putative Pex19-binding site  between amino acids 460 and 469 (Supplementary Figure S1). These observations suggested that CefP may be a peroxisomal membrane protein  (see below).
Targeted inactivation of cefP results in IPN accumulation
To determine whether the CefP protein is involved in CPC biosynthesis, we inactivated the cefP gene by the double-marker technique [8,15,32,33]. Targeted inactivation of this gene was performed using the plasmid pDP (see the Experimental section). This plasmid carries an inactivated cefP gene (obtained by insertion of the hygromycin B resistance gene hph under the control of the gpd gene promoter, as a transformation marker). To confirm that targeted inactivation took place at the correct position, five transformants and the A. chrysogenum C10 parental strain (positive control) were analysed by Southern blotting. The DNA of all strains was digested with SmaI and hybridized with a 4.6 kb SmaI probe containing the cefP gene (Supplementary Figure S2A). Results showed that the control strain A. chrysogenum C10 (Supplementary Figure S2B, lane 6) hybridized with a genomic DNA band of 4.6 kb; however, in TDPs (transformants disrupted in cefP) [TDP-115 (Supplementary Figure S2B, lane 1), TDP-151 (Supplementary Figure S2B, lane 2) and TDP-206 (Supplementary Figure S2B, lane 3)] the 4.6 kb hybridization band was converted into two bands of 7.5 kb and 1.1 kb, as expected, by a canonical double recombination (Supplementary Figure S2A).
To confirm that the cefD1/cefD2 and cefEF/cefG genes were not disrupted during the A. chrysogenum transformation process, a Southern blot analysis was performed. The DNA of the three strains was digested with BamHI and hybridized with one probe of the cefD1/cefD2 bidirectional promoter region (1 kb HindIII-EcoRV; Supplementary Figure S2C) and another probe of the cefEF/cefG genes (7.3 kb BamHI; Supplementary Figure S2C). Results showed that the A. chrysogenum C10 (Supplementary Figure S2D, lane 4) and the transformants TDP-115 (Supplementary Figure S2D, lane 1) and TDP-206 (Supplementary Figure S2D, lane 3) hybridized with two genomic DNA bands of 14 kb (cefD1/cefD2) and 7.3 kb (cefEF/cefG) as expected (Supplementary Figure S2C). However, in the transformant TDP-151 the 14 kb hybridization band was converted into a band of 11 kb, suggesting a reorganization of the cefD1/cefD2 locus, probably by an integration of the pDP plasmid into the cefD1/cefD2 locus (Supplementary Figure S2D, lane 2). This transformant was therefore discarded.
The other two cefP disrupted transformants (TDP-115 and TDP-206) and A. chrysogenum C10 as a control strain were cultured in DP medium. In TDP transformants, a drastic reduction in production of the cephalosporins DAC and CPC was detected by bioassay (approx. 8% of the cephalosporins produced in A. chrysogenum C10), whereas under the same culture conditions, high cephalosporin production was observed in A. chrysogenum C10 (Figure 1A). Analysis by HPLC of the culture-broth supernatant (Figure 2) confirmed the results of the bioassay and showed that the CPC (Figure 1B) and DAC (Figure 1C) production in the disrupted strains were drastically reduced. However, there was an increase in penicillin (IPN and PenN) production (Figure 1D). Further HPLC analysis of the late culture broths (96 h, 120 h and 144 h) of the disrupted strains indicated that the penicillin in this mixture was IPN, whereas in the parental strain only PenN was detected (Figure 1E).
β-Lactam antibiotic production in liquid cultures of cefP-disrupted mutants and the parental strain
HPLC analysis of intracellular DAC (A) and extracellular CPC (B) production in the TDP-115 and TDP-206 disrupted mutants and A. chrysogenum C10
To study whether the protein encoded by cefP was involved in IPN transport to peroxisomes, the intracellular cephalosporin levels (DAC and CPC) were measured by HPLC  in cell extracts of the disrupted TDP transformants and the parental strain A. chrysogenum C10 grown in DP medium for 96, 120 and 144 h. Results showed, when compared with the parental strain, an inefficient biosynthesis of DAC (Figure 2A) and the lack of CPC formation (Figure 2B) in the disrupted mutants. Additionally, there was no PenN accumulation in cefP-disruptant strains (see below).
Taken together, these results indicate that the TDP strains are unable to convert IPN into PenN and cephalosporins in an efficient manner, probably due to a lack of IPN transport related to its conversion into PenN and DAC, and, as a result, an increased amount of extracellular IPN is accumulated.
Cell-free extracts allow in vitro CPC synthesis
All other genes of the CPC pathway appeared to be functional in the cefP-disrupted strains (Supplementary Figure 2 D, lanes 1 and 3). This was confirmed by in vitro conversion of IPN into CPC using cell-free extracts of the transformant TDP-115 (disrupted in cefP) and the parental strain grown in DP medium for 144 h as described previously by Teijeira et al. . Results showed that the IPN-CoA synthetase/IPN-CoA epimerase system (converting IPN into PenN), the expandase/hydroxylase (converting PenN into DAC) and DAC acetyltransferase (converting DAC into CPC) activities were present in the disrupted mutant as well as in the parental strain, because there was in vitro CPC synthesis (Table 1). The IPN is converted into CPC in the cell-free extract when the compartmentalized CefD1 and CefD2 enzymes are released.
|Strain||Genotype||CPC formed (ng of CPC/mg of cell dry weight)|
|A. chrysogenum C10||Parental high cephalosporin producer||8.12±0.8|
|A. chrysogenum TDP-115||ΔcefP−||9.42±0.7|
|Strain||Genotype||CPC formed (ng of CPC/mg of cell dry weight)|
|A. chrysogenum C10||Parental high cephalosporin producer||8.12±0.8|
|A. chrysogenum TDP-115||ΔcefP−||9.42±0.7|
Complementation of the cefP mutation requires additional genetic information
For complementation studies, plasmid pCP, bearing the intact cefP gene with its own promoter (Figure 3A), was transformed into the TDP-206 and TDP-115 strains and transformants were selected by their resistance to phleomycin. Total DNA was extracted from the TDP-206 and TDP-115 blocked mutants (from three transformants of each disrupted strain) and from the parental strain A. chrysogenum C10; the DNAs were digested with SmaI and hybridized with a 4.6 kb SmaI probe containing the cefP gene. Results showed that the probe hybridized with a band of 4.6 kb, as expected (Figure 3A, lanes 1–6), indicating that the cefP gene was integrated in a non-reorganized form. The 7.5 kb and 1.1 kb hybridization bands corresponded to the endogenous hybridizing DNA fragments in the cefP disrupted mutants (Figure 3A; lanes 7 and 8) used as host strains.
Southern blot and RT–PCR analysis of cefP complementation and cephalosporin production of the complemented strains
The effect of complementation of cefP on CPC production was studied with two different transformants named TCP (transformant complemented with cefP). Results showed that in transformants TCP-105 (obtained from TDP-206 complemented with the cefP gene) or TCP-12 (derived from TDP-115 complemented with the cefP gene), cephalosporin production (Figure 3B) was not restored. To confirm these results, 300 transformants (150 of each disrupted strain) that showed resistance to phleomycin were tested for cephalosporin production by the agar-plug method .The results showed again that in none of the cefP-complemented transformants was cephalosporin production restored (results not shown).
There is no expression of the cefP gene in the TCP strains
To determine whether cefP is really expressed in the complemented TCP strains we analysed its expression by RT–PCR using the primers indicated in the Experimental section. The RT–PCR results (Figure 3C) indicated that the A. chrysogenum cefP gene is expressed in the control A. chrysogenum C10 (Figure 3C, lane 1), but not in the disruptant strain TDP-115 (Figure 3C, lane 2) or in the complemented TCP-12 strain (Figure 3C, lane 3). Upstream of cefP there is a gene named cefR involved in the regulation of cef genes (F. Teijeira, R.V. Ullán and J.F. Martín, unpublished work). To clarify the lack of complementation, the expression of the cefR gene was analysed by RT–PCR in the parental strain, the TDP-115 disrupted transformant and the complemented strain TCP-12 (Figure 3D). Surprisingly, results showed that there is no expression of cefR in the disrupted TDP-115 (Figure 3D, lane 2) and in the complemented TCP-12 strains (Figure 3D, lane 3). To confirm that the cefR gene is not altered in the TDP-115 strain as a result of the transformation process, we amplified this gene by PCR using total genomic DNA of this strain as template. The primers used (see the Experimental section) were located on both sides of the cefR gene and amplified a DNA product of 3.3 kb. This DNA fragment was cloned and sequenced, confirming that there is no mutation in the cefR gene or in its promoter region.
Complementation with both cefP and cefR genes is required for restoration of cephalosporin production
To elucidate whether both the cefP and cefR genes, or only cefR, were necessary to restore CPC synthesis in the cefP-disruptant strain, two series of strains transformed with either cefR alone or with both cefP and cefR were created. For this purpose, we constructed plasmids pB5.5R [bearing only the cefR gene (Supplementary Figure S3B at http://www.BiochemJ.org/bj/432/bj4320227add.htm)] and pCRP [bearing both the cefR and cefP genes (Supplementary Figure S3B)]. These plasmids were co-transformed with the help of pJL43  in the TDP-115 mutant and transformants were selected by resistance to phleomycin. We selected at random one transformant of each transformation, TCR-45 (transformant 45 complemented with cefR) and TCPR-27 (transformant complemented with cefR and cefP), which were analysed by Southern blot hybridization to confirm the correct gene integration (Supplementary Figures S3D and S3E). For this analysis, TCR-45, TDP-115 and C10 genomic DNAs were digested with SmaI and hybridized with a 759 bp EcoRI internal fragment of cefR as probe (Supplementary Figure S3B). Results showed that transformant TCR-45 (Supplementary Figure S3D, lane 3) showed hybridization, yielding a DNA band of 4.2 kb in addition to the endogenous 6.2 kb hybridization band [corresponding to the endogenous cefR in the host TDP-115 strain (Supplementary Figure S3D, lane 2)] indicating that the insert of the plasmid pCR was not reorganized.
Similarly, transformant TCPR-27 obtained with the pCRP plasmid was selected, and its DNA was digested with a mixture of BamHI and FspI and hybridized with the cefP probe (Supplementary Figure S2A). Results showed a correct integration of cefP and cefR genes (Supplementary Figure S3E, lane2), as expected (Supplementary Figure S3A), because there was a hybridization band of 8.5 kb [as in the C10 control (Supplementary Figure S3E, lane 3)] in addition to the endogenous 12.5 kb corresponding to the disrupted cefP gene that also appears in the TDP-115 strain (Supplementary Figure S3E, lane 1).
In summary, we obtained three types of complemented strains (Table 2), TCP-12 (complemented with cefP), TCR-45 (complemented with cefR) and TCPR-27 (complemented with both cefP and cefR). Analysis of the cephalosporin production of each of these strains by bioassay (Figures 4A and 4B) revealed that only in the cefP/cefR-complemented strain (TCPR-27) was there efficient cephalosporin biosynthesis. HPLC analysis of the culture broths confirmed the bioassay results and showed that in transformant TCPR-27, the CPC and DAC production was restored to levels similar to those of A. chrysogenum C10 (Figures 4C and 4D). In transformants TCR-45 and TCP-12 carrying only one complementing gene (either cefR or cefP) (Figures 4C and 4D), only a very small amount of cephalosporins was detected, as in the disrupted TDP-115 strain.
|Strain||Plasmid integrated||Complemented mutation|
|A. chrysogenum TDP-115||None||None|
|A. chrysogenum TCP-12||pCP||ΔcefP|
|A. chrysogenum TCR-45||pCR||ΔcefR|
|A. chrysogenum TCPR-27||pCRP||ΔcefP and ΔcefR|
|Strain||Plasmid integrated||Complemented mutation|
|A. chrysogenum TDP-115||None||None|
|A. chrysogenum TCP-12||pCP||ΔcefP|
|A. chrysogenum TCR-45||pCR||ΔcefR|
|A. chrysogenum TCPR-27||pCRP||ΔcefP and ΔcefR|
β-Lactam production in liquid cultures of complemented mutants (TCP-12, TCR-45 and TCPR-27), the cefP-disrupted strain (TDP-115) and the control A. chrysogenum C10
Analysis of extracellular penicillin levels by HPLC (Figure 4E) showed that the penicillin production decreased to normal parental levels in the TCPR-27 transformant with respect to the disrupted strain (47% reduction), whereas TCR-45 and TCP-12 strains showed less reduction (33%) than the cefP/cefR-complemented strain. Analysis of the penicillin isomers (IPN and PenN) in the late-culture-broth supernatants (96 h, 120 h and 144 h) revealed that transformant TCPR-27 secretes PenN at similar levels to the parental strain A. chrysogenum C10, whereas in the culture broth of the strains complemented with a single gene the only penicillin found was IPN (Figure 4F). Intracellular penicillin analysis of the strains (Figure 4G) showed that the intracellular penicillin detected in the cefP/cefR-complemented strain (TCPR-27) was PenN. In addition, RT–PCR analysis of the complemented TCPR-27 strain showed expression of cefP (Figure 3C, lane 4) and cefR (Figure 3D, lane 4).
In summary, integration of both cefR and cefP is necessary to restore cephalosporin production in the cefP-disruptant strain (see the Discussion).
CefP protein is localized in peroxisomes
In order to study the subcellular CefP localization, A. chrysogenum C10 was transformed with the plasmids containing the EGFP–SKL peroxisomal-targeted fluorescent protein and the CefP–DsRed fusion protein as described in the Experimental section. The phleomycin-resistant transformants were analysed by Southern blotting. Their DNAs were digested with a mixture of ApaI and EcoDR2 and hybridized with a probe internal to the DsRed gene. Results of the Southern blot analysis (Supplementary Figure S4B at http://www.BiochemJ.org/bj/432/bj4320227add.htm) revealed that the transformant TPDsRed-32 (Supplementary Figure S4B, lane 9) shows a hybridization band of 4.5 kb corresponding to the Pgdh/cefP/DsRed/Tcyc1 cassette, which is absent from other transformants (Supplementary Figure S4B, lanes 2–9) or in the parental strain (Supplementary Figure S4B, lane 1). Additionally, to confirm the correct integration of the PpcbC/EGFP/SKL/TpenDE cassette into the genome of the transformant TPDsRed-32, a new hybridization analysis was performed, digesting the DNA of TPDsRed-32 and of the parental A. chrysogenum C10 strain with NotI followed by hybridization with a probe corresponding to the EGFP–SKL gene. Results showed a single 2.1 kb hybridizing band corresponding to the correct integration of the EGFP-SKL expression cassette in the TPDsRed-32 transformant (Supplementary Figure S4C, lane 1).
Confocal microscopy analysis of the fluorescence in A. chrysogenum cells grown as indicated in the Experimental section revealed that the green fluorescent EGFP-SKL protein was located in the peroxisomes (Figure 5A) as expected  and the hybrid protein CefP–DsRed (red fluorescence) was found in the same location (Figure 5B). Superposition of both fluorescences resulted in a yellow colour (Figure 5C), confirming that CefP is a peroxisomal membrane protein.
Subcellular localization of the CefP protein
In A. chrysogenum, the central step of the biosynthetic pathway of CPC is the conversion of IPN into its D-isomer PenN. In CPC-producing fungi, this epimerization reaction is catalysed by two enzymes, isopenicillinyl N-CoA synthetase and IPN-CoA epimerase encoded by cefD1 and cefD2 respectively [8,34].
Bioinformatic analysis of CefD1 and CefD2 revealed that both proteins contain putative PTSs characteristic of peroxisomal matrix proteins , indicating a possible peroxisomal matrix localization of these enzymes . CefD2 protein contains putative PST1 [consensus sequence (R/L)-(L/V/I)-X5-(H/Q)(L/A)] and PST2 [consensus sequence (S/C/A)-(K/R/H)-L] signals, whereas CefD1 contains only a putative PTS1. Moreover, the optimum pH for the in vitro IPN epimerization  coincides with the estimate for the peroxisomal lumen . Supporting this possible peroxisomal location, putative CefD1 and CefD2 homologue proteins of P. chrysogenum have been found in the peroxisomal matrix 
Therefore the IPN precursor and PenN product of the epimerization step, which are hydrophilic antibiotics, require specific transport across the peroxisomal membrane
In the same ‘early’ CPC cluster, we previously found an open reading frame named cefM, located downstream from cefD1 in the same orientation, encoding an efflux pump protein with 12 TMSs bearing the characteristic motifs of drug/H+ antiporters (Family 3; drug efflux protein) . Targeted inactivation of cefM affected antibiotic biosynthesis significantly. The cefM-disrupted mutant showed a drastic reduction in extracellular PenN and cephalosporin production and accumulated intracellular PenN. An in vivo confocal microscopy study of the CefM–GFP fusion demonstrated that the CefM protein is located in the membrane of small microbodies where it appears to be implicated in PenN translocation from the microbody lumen to the cytosol . In conclusion, the epimerization process takes place inside the microbodies (peroxisomes), but it was still unknown which protein was responsible for IPN import from the cytosol to the peroxisomal matrix.
On the basis of gene-clustering patterns for secondary metabolite biosynthesis [37,38] we found in the present study a new open reading frame, named cefP, located upstream of the cefT gene . Bioinformatic analysis of the CefP protein revealed that this protein has 11 putative TMSs and one DUF221 (Pfam accession number pfam02714) motif that is found in a family of hypothetical transmembrane proteins. Computer analysis also revealed the presence of a Pex19p-binding domain (located between amino acids 460 and 469) characteristic of proteins that are recruited by the peroxin protein Pex19 to be incorporated into the peroxisomal membrane . The Pex19 protein acts both as an import receptor and as a cytosolic chaperone for peroxisomal membrane proteins .
To study the CefP protein role we inactivated the cefP gene by the double-marker procedure [8,14,15,33]. Cephalosporin production decreased drastically in the disrupted TDP strains, indicating that this protein is essential for CPC biosynthesis in vivo. The disrupted mutants accumulated IPN in the culture broths due to an overflow of this intermediate when its intracellular level increases. A similar accumulation of IPN takes place in the cefD1/cefD2-disrupted mutants , indicating that the IPN epimerization into PenN is indeed blocked in TDP mutants. However, unlike what we found in the cefD1/cefD2-disrupted mutants, the TDP strains constructed in the present study show in vitro IPN epimerase activity for the efficient conversion of IPN into PenN. Similar in vitro conversion of PenN into CPC using cell-free extracts of the cefM-disrupted strain showed that the late CPC enzymatic activities were present, although cefM-null mutants were unable to produce this β-lactam antibiotic .
These results indicate that IPN transport across the peroxisomal membrane is blocked as a consequence of the lack of the CefP transporter in the cefP-disrupted strains, explaining the increased secretion of the intermediary IPN into the culture broth. It is known that the CefT protein is involved in the secretion of hydrophilic β-lactams, including IPN [40,41]. CefT seems to be responsible for the secretion of the intracellular IPN excess from the cell into the extracellular medium in the epimerase-null or cefP-disrupted strains. In the wild-type and improved CPC-production strains, IPN is secreted in low amounts because in those strains it is efficiently converted into cephalosporins [2,42].
Surprisingly, complementation in trans with the single cefP gene under the control of its own promoter was not sufficient to restore a wild-type phenotype in the TDP mutant, but required the simultaneous introduction of the cefR gene located upstream of cefP in the same orientation. RT–PCR analysis showed that cefR is not expressed in the TDP mutants. Apparently, a functional CefP protein is required for induction of cefR expression, perhaps through the transport of an inducer molecule. Complementation with the entire cefR/cefP fragment (8.5 kb) restored the wild-type phenotype, indicating that both linked genes form a cluster that is essential and necessary for an efficient β-lactam biosynthesis (F. Teijeira, R. V. Ullán and J. F. Martín, unpublished work).
The present study confirms the compartmentalization of the biosynthesis of CPC in A. chrysogenum [2,14], particularly in the transport of penicillin (IPN and PenN) across the peroxisomal membrane. The confocal microscopy experiments reported in the present paper clearly indicate that the microbodies reported in our previous paper  are authentic peroxisomes since the p43EGFP-SKL peroxisome-targeted protein co-localizes with the DsRed fluorescent CefP hybrid protein. The role of CefP is probably to transport IPN from the cytosol to the peroxisomal matrix where it is epimerized to form PenN. Finally, the CefM protein  secretes the epimerization product (PenN) from the peroxisomal lumen into the cytosol where it is converted into CPC by the expandase/hydroxylase  and DAC acetyltransferase  activities. We may conclude that peroxisomes play an indispensable role in the biosynthetic pathway of CPC in A. chrysogenum since they seem to contain the enzymes that catalyse the epimerization step as demonstrated by characterization of the CefM  and CefP carriers. The usefulness of this compartmentalization of the CPC biosynthetic process in A. chrysogenum, as in P. chrysogenum for penicillin biosynthesis, is to allow division of precursors and enzymes and to aid in the regulation of the reactions involved. These advances in the compartmentalization of β-lactam biosynthetic enzymes in fungi may serve as a model for the localization of other secondary-metabolite biosynthetic enzymes in microbodies of plants and other fungi.
We thank Dr J.A.K.W. Kiel and Dr M. Veenhuis (Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands) for providing the pGBRH2-EGFP-SKL plasmid. We acknowledge the excellent technical assistance of A. Sánchez, B. Martín, J. Merino, A. Casenave and A. Mulero.
Discosoma sp. red fluorescent protein
domain of unknown function
green fluorescent protein
hygromycin B phosphotransferase gene
Le Page and Campbell
major facilitator superfamily
peroxisome biogenesis factor 19
peroxisomal targeting signal
transformant complemented with cefP
transformant complemented with cefR
transformant complemented with cefR and cefP
transformant disrupted in cefP
Ricardo Ullán performed most of the experimental work with the help of Fernando Teijeira. Susana Guerra participated in the genetic analysis of the mutant strains. Immaculada Vaca performed studies on secretion of the intermediates. Juan Martín directed the work and prepared the manuscript.
This work was supported by the European Union (Eurofungbase) [grant number LSSG-CT-2005–018964], and the Torres Quevedo Program [grant number PTQ06–2-0114 (to F.T.)].
Present address: Department of Chemistry, Faculty of Sciences, University of Chile, Las Palmeras 3425, Ñuñoa, Santiago de Chile, Chile.
The nucleotide sequence data reported for cefP will appear in the GenBank®, EMBL, DDBJ and GSDB Nucleotide Sequence Databases under the accession number AM231816.