APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 2011, p. 1413–1422 Vol. 77, No. 40099-2240/11/$12.00 doi:10.1128/AEM.01531-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Autophagy Deficiency Promotes �-Lactam Production inPenicillium chrysogenum�†

Magdalena Bartoszewska,1,2 Jan A. K. W. Kiel,1,2 Roel A. L. Bovenberg,3Marten Veenhuis,1,2 and Ida J. van der Klei1,2*

Molecular Cell Biology,1 and Kluyver Centre for Genomics of Industrial Fermentation,2 University of Groningen, P.O. Box 14,9750 AA Haren, Netherlands, and DSM Biotechnology Center, P.O. Box 425, 2600 AK, Delft, Netherlands3

Received 28 June 2010/Accepted 12 December 2010

We have investigated the significance of autophagy in the production of the �-lactam antibiotic penicillin(PEN) by the filamentous fungus Penicillium chrysogenum. In this fungus PEN production is compartmental-ized in the cytosol and in peroxisomes. We demonstrate that under PEN-producing conditions significantamounts of cytosolic and peroxisomal proteins are degraded via autophagy. Morphological analysis, based onelectron and fluorescence microscopy, revealed that this phenomenon might contribute to progressive deteri-oration of late subapical cells. We show that deletion of the P. chrysogenum ortholog of Saccharomyces cerevisiaeserine-threonine kinase atg1 results in impairment of autophagy. In P. chrysogenum atg1 cells, a distinct delayin cell degeneration is observed relative to wild-type cells. This phenomenon is associated with an increase inthe enzyme levels of the PEN biosynthetic pathway and enhanced production levels of this antibacterialcompound.

The discovery of penicillins in 1929 by Fleming is probablythe most important observation in the history of therapeuticmedicine development. Penicillins belong to the group of�-lactam antibiotics and are produced as secondary metabo-lites by several filamentous fungi (3). For industrial productionthe filamentous fungus Penicillium chrysogenum is used. Thefirst steps of the penicillin (PEN) biosynthetic pathway takeplace in the cytosol. The amino acid precursors L-�-aminoadi-pic acid (L-�-AAA), L-cysteine, and L-valine are condensedinto the tripeptide �-(L-�-aminoadipyl)- L-cysteinyl- D-valine(ACV) by the enzyme ACV synthetase (ACVS). Isopenicil-lin N synthase (IPNS) catalyzes the oxidative ring closure ofthe linear ACV tripeptide, which leads to the formation ofisopenicillin N (IPN), which has a bicyclic ring structure.The final step of penicillin biosynthesis, in which the hydro-philic L-�-AAA side chain of IPN is exchanged for a hydro-phobic acyl group, occurs inside peroxisomes via isopenicil-lin N acyltransferase (IAT) and phenylacetyl coenzyme Aligase (PCL) (2).

Peroxisomes are single-membrane-bound organelles presentin all eukaryotes. These cellular compartments are involved invarious metabolic pathways. The importance of peroxisomesfor efficient penicillin production in P. chrysogenum has beenwell documented (23, 24). Muller et al. (24) first suggested acorrelation between penicillin production and the volume frac-tion of peroxisomes. Later, it was shown that the high-produc-ing strain DS17690 has enhanced numbers of these organelles

relative to the original NRRL 1951 strain (41). Moreover,induction of peroxisome proliferation via overproduction ofPex11 in P. chrysogenum resulted in enhanced levels of peni-cillin production in two laboratory strains (14).

A remarkable feature of filamentous fungi is the differenti-ation of cells along the hyphae. These structures can be dividedinto actively growing regions (apical cells), metabolically activenongrowing regions (subapical cells), and the oldest part ofthe hyphae, which are comprised of degenerating, highlyvacuolated cells (27). Interestingly, it was suggested that�-lactam production is restricted only to some compart-ments of the hyphae in P. chrysogenum. Based on a struc-tured kinetic model describing growth, differentiation, andpenicillin production in submerged P. chrysogenum fermen-tations, it was suggested that antibiotic production is relatedto the amount of the metabolically active subapical regionsof the hyphae (27).

Autophagy is a highly conserved mechanism in which or-ganelles and proteins are degraded and recycled in the vacuolarlumen. This mechanism is crucial for maintenance of cellularhomeostasis, survival during nutrient starvation, and orches-tration of an efficient cellular response to stress (40). Althoughautophagy is generally considered a prosurvival mechanism,under specific conditions this process can also participate incell death (32, 34, 46). In filamentous fungi autophagy wasshown to be involved in nutrient recycling under starvationconditions and during developmental processes (30).

In this study we examined whether inhibition of autophagy-related processes is associated with a delay in the degenerationof late subapical cells. We demonstrate that autophagic deg-radation of cellular components occurs mainly in older sub-apical compartments of the hyphae under PEN productionconditions. This phenomenon may contribute to progressivedeterioration of these cells. Furthermore, the delayed deteri-oration observed in an autophagy-deficient P. chrysogenum

* Corresponding author. Mailing address: Molecular Cell Biology,Groningen Biomolecular Sciences and Biotechnology Institute, Uni-versity of Groningen, P.O. Box 14, 9750 AA Haren, Netherlands.Phone: 31 (0)50 363 3179. Fax: 31 (0)50 363 8280. E-mail: i.j.van.der.klei@rug.nl.

† Supplemental material for this article may be found at http://aem.asm.org/.

� Published ahead of print on 17 December 2010.

1413

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

mutant strain resulted in a significant increase in the amount ofPEN produced.

MATERIALS AND METHODS

Strains and cultivation conditions. P. chrysogenum strains used in this studyare listed in Table 1. For biochemical and ultrastructural analyses, P. chrysoge-num strains were cultivated on PEN induction medium (9) supplemented with0.05% phenoxyacetic acid. In order to induce nonselective autophagy, myceliawere grown for 24 h on PEN induction medium and then harvested by centrif-ugation and resuspended in PEN induction medium without a nitrogen source.For genetic manipulation purposes, P. chrysogenum strains were grown on YGGmedium (0.8% KCl, 1.6% glucose, 0.67% yeast nitrogen base [Difco], 0.15%citric acid, 0.6% K2HPO4, 0.2% yeast extract, pH 6.2, supplemented with peni-cillin and streptomycin [Gibco]). P. chrysogenum mycelia were cultivated at 25°Cat 200 rpm in batch cultivation mode.

P. chrysogenum niaD-deficient transformants were selected on plates contain-ing 1.25% KClO3 and supplemented with 0.185% adenine as sole nitrogensource. Phleomycin-resistant (Bler) P. chrysogenum strains were selected onPhleo-plates containing 50 �g/ml of phleomycin (Invitrogen) (17). To induceconidiospore formation, R-agar was used (0.52% [vol/vol] glycerol, 0.75% [vol/

vol] beet molasses, 0.5% yeast extract, 300 mM NaCl, 0.2 mM MgSO4 � 7H2O,0.370 mM KH2PO4, 3.3 �M NH4Fe(SO4)2 � 12H2O, 0.4 �M CuSO4 � 5H2O, and1.45 mM CaSO4 � 2H2O).

Escherichia coli DH5� and XL1-Blue cells, which were used for cloning pur-poses, were grown in super optimal broth with catabolite repression (SOC) orlysogeny broth (LB) (33) medium supplemented with the appropriate antibiotics.

Micrococcus luteus ATCC 9341, which was used for the PEN bioassay, wascultivated at 30°C on 2� TY medium (6).

Molecular techniques. Plasmids and oligonucleotides used in this study arelisted in Tables 2 and 3, respectively. Standard DNA manipulations were per-formed according to established procedures described by Sambrook et al. (33).Some of the cloning procedures were performed using the Multisite Gatewaytechnology (Invitrogen). Transformation of P. chrysogenum protoplasts was per-formed as described previously (4). Restriction enzymes (Fermentas and Roche),T4 DNA ligase (Fermentas), and other DNA-modifying enzymes were used asrecommended by the suppliers. PCRs were carried out with Phusion (Fermentas)or Expand (Roche) high-fidelity polymerase. After cloning, all DNA fragmentsobtained via PCR were sequenced (ServiceXS). Southern blotting was per-formed according to the DIG High Prime labeling and detection kit (Roche). Forin silico DNA sequence analysis, the Clone Manager 5 software (Scientific andEducation Software, Durham, NC) was used.

TABLE 1. P. chrysogenum strains used in this study

Strain Genotype or characteristics Reference or source

DS17690 High-PEN-producing strain, AmdS� 8DS54465 DS17690 with deletion of hdfA gene, AmdS� 37DS54465 GFP.SKL DS54465 with integrated PgpdA-GFP.SKL-TpenDE cassette at niaD locus, AmdS�,

chlorate resistantW. H. Meijer et al.,

submitted forpublication

DS17690 GFP DS17690 with integrated PpcbC-GFP-TpenDE cassette, AmdS� This studyatg1 DS54465 with deletion of atg1 gene, phleomycin resistant, AmdS� This studyatg1::GFP atg1 with PgpdA-GFP-TpenDE cassette integrated at niaD locus, phleomycin

resistant, AmdS�, chlorate resistantThis study

atg1::GFP.SKL atg1 with Pgpda-GFP.SKL-TpenDE cassette integrated at niaD locus, phleomycinresistant, AmdS�, chlorate resistant

This study

atg1::GFP � atg1 atg1::GFP with PATG1-ATG1-TATG1 cassette integrated at pyrG locus This study

TABLE 2. Plasmids used in this study

Plasmid Characteristicsa Reference or source

pDONRP4-P1R Multisite Gateway vector, Kanr Cmr InvitrogenpDONR221 Multisite Gateway vector, Kanr Cmr InvitrogenpDONRP2R-P3 Multisite Gateway vector, Kanr Cmr InvitrogenpDESTR4-R3 Multisite Gateway vector, Ampr Cmr InvitrogenpDESTR4-R3/AMDS pDESTR4-R3 with PgpdA-amdS cassette, Ampr 13pDELatg1 pDESTR4-R3/AMDS with atg1 deletion cassette, Ampr This studypENTR41-Patg1 pDONRP4-P1R with 5�-flanking region of atg1, Kanr This studypENTR23-Tatg1 pDONRP2R-P3 with 3�-flanking region of atg1, Kanr This studypENTR221phleo pDONR221 vector with PpcbC-phleomycin cassette, Kanr 16pMDB005 Plasmid containing PgpdA-eGFP-TpenDE cassette, Ampr This studypMDB006 pDONR221 with PgpdA-eGFP-TpenDE cassette, Kanr This studypENTR41-5�niaD pDONRP4-P1R with 5�-flanking region of niaD, Kanr Meijer et al., submittedpENTR23-3�niaD pDONRP2R-P3 with 3�-flanking region of niaD, Kanr Meijer et al., submittedpMDB007 pDESTR4-R3 with niaD::PgpdA-eGFP -TpenDE cassette, Ampr This studypGBRH2-eGFP pGBRH2 plasmid containing PpcbC-eGFP-TpenDE cassette, Ampr 13pNiGaNi Plasmid containing PgpdA-amdS expression cassette flanked by niaD

sequences, AmprLab collection

pBBK-007 Plasmid containing PgpdA-DsRed.SKL-TpenDE cassette, Ampr Meijer et al., submittedpEXP-5�niaD-PgpdA-

GFP.SKL-TpenDE-3�niaD

pDESTR4-R3 containing niaD-PgpdA-GFP.SKL-TpenDE cassette, Ampr Meijer et al., submitted

pENTR23-ATG1 pDONR P2R-P3 vector containing PATG1-ATG1-TATG1 cassette, Kanr This studypENTR221/AMDS pDONR 221 vector with PgpdA-amdS cassette, Kanr Gift from J. G. NijlandLMOPE41pyrG pDONR P4-P1R with pyrG gene, Kanr 26pCOMatg1 Plasmid containing PATG1-ATG1-TATG1 complementation cassette,

pyrG gene, and amdS marker, AmprThis study

a Kan, kanamycin; Cm, chloramphenicol; Amp, ampicillin.

1414 BARTOSZEWSKA ET AL. APPL. ENVIRON. MICROBIOL.

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

Plasmid construction. (i) Construction of a P. chrysogenum atg1 deletionstrain. The atg1 deletion cassette was constructed using the Multisite Gatewaycloning system. The upstream flanking region (1.7 kb) and downstream flank-ing region (2.1 kb) of the atg1 gene were amplified by PCR with the primersPdelATG1_Fow1 and PdelATG1_Rev1 and with TdelATG1_Fow1 andTdelATG1_Rev1, respectively, using DS54465 genomic DNA as template. Theresulting PCR products were introduced into pDONRP2R-P3 (3�-flanking re-gion of atg1) and pDONRP4-P1R (5�-flanking region of atg1) by using theGateway BP clonase reaction, resulting in plasmids pENTR23-Tatg1 andpENTR41-Patg1, respectively. Subsequently, a Gateway LR reaction was executedwith pENTR41-Patg1, pENTRY221phleo, pENTR23-Tatg1, and the destinationvector pDEST R4-R3/AMDS, resulting in the formation of the atg1 deletionconstruct pDELatg1. This plasmid was linearized with NotI and transformed intoP. chrysogenum DS54465 protoplasts. Bler transformants were selected and an-alyzed by colony PCR using the primers 5� atgknout.for and Phleo.rev. Deletionof atg1 was confirmed by Southern Blotting using a 936-bp atg1 promoter-specificprobe, which was generated by PCR with primers FWRprobe and REVprobe.

(ii) Construction of P. chrysogenum atg1::GFP, atg1::GFP.SKL, and DS17690GFP strains. The pMDB005 plasmid, which contains PgpdA-eGFP-TpenDE, wasconstructed as follows: plasmid pGBRH2-eGFP was digested with BamHI andSmaI, and the enhanced green fluorescent protein (eGFP)-containing fragmentwas cloned between the BamHI and SmaI sites of plasmid pBBK007, therebyreplacing DsRed.SKL. Subsequently, pMDB005 was used as template to generatean attB1-PgpdA-eGFP-TpenDE-attB2 fragment by PCR using the primers MDB008and LMOp012. The resulting PCR product was introduced into pDONR221using Gateway technology, generating plasmid pMDB006. Next, the Gateway LRreaction was performed with pENTR41-5�niaD, pENTR23-3�niaD, pMDB006,and pDESTR4-R3 to obtain plasmid pMDB007. This plasmid was linearized

with KpnI and transformed into P. chrysogenum atg1 protoplasts. NiaD� trans-formants were selected based on their ability to grow on medium supplementedwith chlorate. Production of GFP was confirmed by fluorescence microscopy.

In order to generate the P. chrysogenum atg1::GFP.SKL strain, the pEXP-5�-niaD-PgpdA-GFP.SKL-TpenDE-3�niaD plasmid was linearized with KpnI andtransformed into P. chrysogenum atg1 protoplasts. NiaD� transformants wereselected based on their ability to grow on medium supplemented with chlorate.Production of GFP.SKL protein was confirmed by fluorescence microscopy.

In order to generate the P. chrysogenum DS17690 GFP strain, a 2.2-kb NotIfragment from pGBRH2-eGFP containing the PpcbC-eGFP-TpenDE cassette wascotransformed into P. chrysogenum DS17690 protoplasts together with a 6.2-kbNotI-SmaI DNA fragment containing the amdS selection marker from pNiGaNi.AmdS� cotransformants that showed GFP fluorescence were selected.

Complementation of the atg1 deletion strain. The atg1 complementation cas-sette was constructed using the Multisite Gateway cloning system. The atg1 gene,containing 1,500 bp upstream of the predicted translation start site and 500 bpdownstream of the stop codon, was amplified by PCR with the primersATG1comfor and ATG1comrev, with DS54465 genomic DNA as the template.The resulting PCR product was introduced into pDONR P2R-P3 by using theGateway BP clonase reaction, resulting in plasmid pENTR23-ATG1. Subse-quently, a Gateway LR reaction was performed with pENTR23-ATG1,pENTR221/AMDS, LMOPE41pyrG, and the destination vector pDEST R4-R3,resulting in the formation of the atg1 complementation construct pCOMatg1.This plasmid was transformed in circular form into atg1::GFP P. chrysogenumprotoplasts, thereby promoting single crossover at the pyrG locus. AmdS� trans-formants were selected. Correct integration was confirmed by colony PCR, usingthe primers pyrGupst and AMDS.rev.

FIG. 1. Disruption of the atg1 gene in P. chrysogenum. (A) Strategy for the integration of the phleomycin cassette in the atg1 locus (left).Southern blot analysis of HindIII-digested genomic DNA by using the indicated 936-bp probe identified the expected 1.23-kb band in wild-typeDNA and the 6.6-kb band in DNA from the atg1 mutants (right). (B) Morphology of colonies of the WT and the atg1 mutant strain after 7 daysof growth on R-agar plates at 25°C. The atg1 null mutant generates white colonies because of its reduced sporulation ability. The colony of theWT control normally forms green spores.

TABLE 3. Oligonucleotides used in this study

Oligonucleotide Sequencea (5�–3�)

PdelATG1_Fow1.......................5�-GGGG ACA ACT TTG TAT AGA AAA GTT GGG AGA TTC CGT CAC GAG ATG-3�PdelATG1_Rev1 .......................5�-GGGG ACT GCT TTT TTG TAC AAA CTT G GC CTG TTC CGT CTC TGG TAA-3�TdelATG1_Fow1 ......................5�-GGGG ACA GCT TTC TTG TAC AAA GTG GTC CGG AAG AAG GTA GCT GTT-3�TdelATG1_Rev1 .......................5�-GGGG ACA ACT TTG TAT AAT AAA GTT G CA TGC CAA TTC CAC GCT GAT-3�MDB008 ....................................5�-GGGG ACA AGT TTG TAC AAA AAA GCA GGC TGC TCT GTA CAG TGA CC GGT GAC TC-3�LMOp012 ..................................5�-GGGG AC CAC TTT GTA CAA GAA AGC TGG GTT CCCC TGA AAG AGT TGA TAT TGA AGG-3�FWRprobe.................................5�-GCTTCCGTCACCGTACAGTT-3�REVprobe .................................5�-CTACCTCAGCAGCAGCAATG-3�Phleo.rev ....................................5�-AACGGCACTGGTCAACTTGG-3�5� atgknout.for ..........................5�-GCTTCCGTCACCGTACAGTT-3�ATG1comfor .............................5� GGGG ACA GCT TTC TTG TAC AAA GTG GCC GAC TTT GCT CAA GG CGG TTC 3�ATG1comrev.............................5�-GGGG ACA ACT TTG TAT AAT AAA GTT GCC TGG TAT CTG CAA ACC CAA TCC-3�pyrGupst ....................................5�-GAACGCCTCGCAGACAATGCTC 3�AMDS.rev..................................5�-GCATGCCAGAAAGAGTCACC-3�

a The underlined sequences are the attB recombination sites for the Multisite Gateway system.

VOL. 77, 2011 AUTOPHAGY DEFICIENCY PROMOTES PENICILLIN PRODUCTION 1415

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

Bioassays. PEN bioassays using M. luteus ATCC 9341 as an indicator organismwere carried out as described previously (6) using the spent medium of culturesgrown on PEN induction medium supplemented with phenoxyacetic acid. Dryweights of the cultures were determined and used as a reference.

Biochemical analysis. Crude extracts of P. chrysogenum cells were prepared asdescribed previously (13). Protein concentrations were measured using the RCDC assay system (Bio-Rad) with bovine serum albumin as a standard. SDS-PAGE and Western blotting were carried out in accordance with establishedprocedures. Blots were decorated with polyclonal antibodies raised against P.chrysogenum IPNS, IAT, and transcription elongation factor EF1-� (7, 12, 42).

Microscopy. Wide-field fluorescence microscopy was carried out using an AxioObserver Z1 fluorescence microscope (Zeiss). Images were taken using an EC-Plan Neofluar 100�/1.3 numerical aperture objective and a Coolsnap HQ2camera (Roper Scientific Inc.). GFP fluorescence was visualized with a 470/40-nm band-pass excitation filter, a 495-nm dichromatic mirror, and a 525/50-nmband-pass emission filter. FM4-64 (Invitrogen) fluorescence was analyzed with a545/25-nm band-pass excitation filter, a 570-nm dichromatic mirror, and a 605/70-nm band-pass emission filter. Z-stack images were made with an interval of 1�m. High-resolution overall images were generated from individual tiles by usingthe MosaiX module. P. chrysogenum mycelia were stained with fluorescentprobes prior to microscopy investigation, in accordance with manuals providedby the suppliers (Molecular Probes and Invitrogen).

For electron microscopy cells were fixed and prepared as described previously(44).

NMR analysis of �-lactams. The quantitative 1H nuclear magnetic resonance(NMR) analysis of P. chrysogenum fermentation products was performed using aBruker Avance 600 spectrometer at 600 MHz. The filtrate was mixed with aknown amount of maleic acid as the internal standard and lyophilized. Thelyophilizate was dissolved in D2O and measured at 300 K. The 30-s delaybetween scans was more than five times the T1 of all compounds, and so theexact amounts of analyzed compounds were measured by comparing the ratiosbetween the integrals of the analyzed chemicals and the integral of the internalstandard.

RESULTS

Deletion of the atg1 gene results in inhibition of autophagyin P. chrysogenum. The Saccharomyces cerevisiae atg1 gene en-codes a serine-threonine kinase, which is involved in the earlysteps of autophagy induction (5). Inactivation of the Atg1protein in yeast or its ortholog in filamentous fungi results inimpairment of autophagy (11, 20, 31).

To determine the significance of the P. chrysogenum atg1gene in PEN production, multiple independent atg1 deletionstrains were constructed and analyzed. Correct integration ofthe deletion cassette was confirmed by Southern blotting (Fig.1A). The atg1 null mutants did not display any growth defectbut showed a strong reduction in conidiospore formation(Fig. 1B).

In order to analyze the significance of the Atg1 protein inturnover of cytoplasmic components, we constructed atg1strains that either produced GFP.SKL to mark peroxisomes(atg1::GFP.SKL) or GFP without any sorting sequence to vi-sualize the cytosol (atg1::GFP). Wild-type (WT) and atg1strains producing GFP.SKL or GFP were grown for 24 h onPEN induction medium and were subsequently subjected tonitrogen starvation conditions to induce autophagy. Underthese conditions, in WT cells a strong accumulation of GFP(Fig. 2B) and GFP.SKL (Fig. 2D) fluorescence was observed inthe vacuolar lumen, indicative of uptake and degradation ofthese components. This vacuolar fluorescence was below thelimit of detection in cells prior to the shift (Fig. 2A). In con-trast, in the atg1::GFP and atg1::GFP.SKL strains, vacuolarGFP fluorescence was never observed under nitrogen limita-tion conditions (Fig. 2C and E). As a control, an atg1 strain wasused that was complemented with the atg1 gene under the

FIG. 2. Autophagy is impaired in P. chrysogenum atg1 null mutants.Mycelia of P. chrysogenum strains expressing GFP or GFP.SKL werecultivated under PEN-inducing conditions for 24 h, harvested by cen-trifugation, and shifted to PEN-inducing medium in the absence of anitrogen source to induce autophagy. Vacuoles are marked red byFM4-64. (A) GFP fluorescence inside the vacuolar lumen of WT cellswas below the limit of detection at the onset of the experiment. (B andC) Degradation of cytosolic components was evident in WT cells after24 h of growth under nitrogen starvation conditions (B) but not in atg1cells (C). (D and E) In WT cells, also GFP.SKL was observed in thevacuoles due to pexophagy (D), but not in atg1 cells (E). Bars, 5 �m.

1416 BARTOSZEWSKA ET AL. APPL. ENVIRON. MICROBIOL.

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

control of its own promoter. In this strain the N-limited-in-duced autophagy was fully restored (see Fig. S1A in the sup-plemental material). Also, this strain regained the ability toproduce green spores (see Fig. S1B).

These data suggest that autophagy is impaired in the P.chrysogenum atg1 null mutant.

Autophagic turnover of cytosolic components and peroxi-somes occurs in subapical regions. The kinetics of constitutiveautophagy in P. chrysogenum strains cultivated under PEN-inducing conditions was analyzed using fluorescence micros-copy. P. chrysogenum strains producing GFP were used tomonitor autophagy-related degradation of cytosolic compo-nents. In WT cells, accumulation of GFP inside the vacuolewas not observed at early time points of cultivation (0 to 24 hafter inoculation of spores), suggesting that recycling of cyto-solic components via autophagy was below the limit of detec-tion in young hyphal cells (Fig. 3A). However, after 48 h ofgrowth, GFP accumulation in vacuoles of subapical cells wasevident (Fig. 3C). This phenomenon was not observed in atg1deletion cells, indicating that accumulation of GFP inside thevacuole in P. chrysogenum WT cells represents an autophagicprocess (Fig. 3B and D). Comparable results were observed forperoxisomes, marked with GFP.SKL. However, accumulationof GFP.SKL in the vacuole was retarded relative to uptake ofits cytosolic counterpart, but it was massively detected predom-inantly in late subapical cells at approximately 70 h of cultiva-tion (Fig. 4A). During prolonged cultivation, these phenomenaoccurred consecutively in time, in successive cells of thehypha (Fig. 4B and 5A). In contrast, vacuolar GFP.SKLfluorescence was not observed in atg1 cells (Fig. 5B). Mor-phometric analysis revealed that atg1 mutant cells containedsignificantly enhanced numbers of peroxisomes in late subapi-cal cells relative to WT (Fig. 5C). These data imply that con-tinuous turnover of cytosol and peroxisomes occurs via autoph-agy-related processes in subapical compartments of hyphaeduring cultivation of P. chrysogenum under PEN-inducing con-ditions.

Autophagy contributes to cell deterioration in old subapicalcompartments. The substantial degradation of cytosolic com-

ponents and peroxisomes in subapical cells during cultivationof WT P. chrysogenum was particularly evident at later stagesof growth. In order to determine the effect of extensiveautophagy on the morphology of stationary-phase cells (af-ter 130 h of cultivation), the autophagy-deficient atg1 andWT P. chrysogenum strains were analyzed by electron mi-croscopy. These analyses revealed that old hyphal compart-ments of the WT strain were characterized by extensiveautophagy and cell deterioration, in contrast to the corre-sponding compartments in atg1 cells (Fig. 6). These datasuggest that inhibition of autophagy in P. chrysogenum maybe associated with a delay in progressive deterioration ofcells at late stages of growth.

Inhibition of autophagy-related processes in P. chrysogenumresults in enhanced levels of PEN biosynthesis enzymes andenhanced PEN production. To determine the effect of atg1deletion on the levels of the enzymes involved in �-lactamproduction, we cultured WT and atg1 mutant cells under PEN-inducing conditions and analyzed protein abundance levels ofthe cytosolic enzyme IPNS and the peroxisomal enzyme IATover time via Western blotting. This analysis revealed that bothIAT and IPNS protein levels were enhanced in crude extractsof atg1 cells relative to those in WT extracts at time points atwhich WT cells showed significant induction of autophagy insubapical compartments of the hyphae (Fig. 7A). In order toanalyze whether the enhanced amounts of PEN biosynthesisenzymes are accompanied by elevated PEN production levels,bioassays using Microccocus luteus as the indicator strain wereperformed. The data showed a significantly larger clearingsurface when culture supernatants of atg1 cells were used,relative to WT (Fig. 7B), implying that deletion of atg1 leads toenhanced production of antibacterial agents by P. chrysoge-num. Both increased antibacterial activity and enhanced levelsof PEN biosynthesis enzymes were observed for multiple in-dependent atg1 mutants (see Fig. S2 in the supplemental ma-terial).

In order to analyze whether the increased antibacterial ac-tivity of atg1 culture supernatants was caused by enhancedlevels of PEN produced, quantitative NMR determinations of

FIG. 3. Constitutive turnover of cytosolic components under PEN-inducing conditions. Fluorescence microscopy images show time-dependentdegradation of cytosolic components via autophagy in cells of the WT strain grown under PEN-inducing conditions. This process was impaired inthe atg1 strain. Vacuoles were stained by FM4-64. Degradation of cytosol was below the limit of detection after 24 h in both WT (A) and atg1 cells(B) but prominent after 48 h of cultivation in WT cells (C), but not in atg1 cells (D). Bars, 5 �m.

VOL. 77, 2011 AUTOPHAGY DEFICIENCY PROMOTES PENICILLIN PRODUCTION 1417

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

secreted penicillin V (PenV) were performed. The parentalhigh-PEN-producing strain (DS54465) and the isogenic atg1mutant strain were cultivated on PEN induction medium for 7days. Under these conditions the atg1 mutant produced 37%

more penicillin V (average, 1.3 g/liter) than the correspondingparental strain (average, 0.95 g/liter; P 0.033 [Student’s ttest]). At this time point the sugars were completely consumed,and cell densities (in terms of dry weight per liter) of both

FIG. 4. Turnover of peroxisomes under PEN-inducing conditions. Fluorescence microscopy images show degradation of peroxisomes over timein the WT strain grown under penicillin-inducing conditions. (A) Accumulation of GFP.SKL fluorescence inside the vacuolar lumen (arrows),indicating turnover of peroxisomes, was observed after approximately 70 h of cultivation, predominantly in late subapical cells. (B) Duringprolonged growth (96 h), degradation of peroxisomes inside the vacuole was observed in successively formed cells of the hyphae. The asteriskindicates an apical compartment. Bars, 20 �m.

1418 BARTOSZEWSKA ET AL. APPL. ENVIRON. MICROBIOL.

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

cultures were comparable (Fig. 7C). Thus, impairment ofautophagy-related processes in a high-PEN-producing strainresults in significantly increased production of PenV.

DISCUSSION

We have studied the role of autophagy in P. chrysogenumand have provided evidence that blocking autophagy promotesPEN production by this fungus. Autophagy has been impli-cated in many cellular processes, including survival under nu-trient starvation conditions, cell remodeling (25), removal ofdamaged organelles and proteins, and protection against oxi-dative stress (45). The core machinery executing autophagy isconserved among eukaryotes (25), including filamentous fungi(21, 30). Inhibition of autophagy in filamentous fungi by dis-ruption of specific ATG genes has provided evidence that thisprocess is involved in starvation-induced differentiation, in-

cluding the formation of aerial hyphae (31), conidiophores(15), and sexual reproductive structures (19, 29).

In our studies the role of autophagy in the �-lactam an-tibiotic-producing fungus P. chrysogenum was analyzed. Weshowed that disruption of the atg1 gene in P. chrysogenumimpairs autophagy and causes sporulation defects, as in otherfungi (20, 22, 28, 31). Interestingly, we observed that in P.chrysogenum the degradation of cytosolic components pre-ceded the turnover of peroxisomes under nitrogen starvationconditions. These data, together with previous observations(18), suggest that bulk autophagy, although considered anonselective process, may degrade cellular components in aspecific order. A more nonspecific autophagic response un-der nitrogen starvation conditions could be disadvantageousfor the cell, as it may degrade portions of essential subcellularcomponents required for basic cell function at the initiation ofthe recycling process.

FIG. 5. atg1 cells contain enhanced numbers of peroxisomes. Mycelia were cultivated under PEN-inducing conditions for 130 h. Peroxisomeswere labeled with GFP.SKL. (A) Significant turnover of peroxisomes was present in almost all cells along the hyphae of WT P. chrysogenum. (B) Anaccumulation of peroxisomes and absence of GFP.SKL fluorescence inside vacuoles were observed in the atg1 null mutant. Bars, 20 �m.(C) Peroxisome numbers in late subapical cells. Fluorescent spots were counted from grouped Z-stacks. The numbers of peroxisomes per �m ofhyphae are presented. The atg1 null mutant strain is characterized by a significantly increased number of peroxisomes in late subapicalcompartments (Student’s t test, P 0.0004). At least 65 subapical hyphal cells were randomly counted in three independent experiments. Scalebars represent the standard errors of means.

VOL. 77, 2011 AUTOPHAGY DEFICIENCY PROMOTES PENICILLIN PRODUCTION 1419

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

The mycelium of ascomycetous fungi is composed of inter-connected hyphae that are divided by perforated septa intomultinucleated compartments, which allows bulk flow of cy-tosol and organelles throughout the hyphae. Recycling of cel-lular components of mature hyphal compartments via autoph-agy and autolysis has been suggested to support apical cellgrowth under nutrient limitation conditions (35, 36). It is prob-able that at a late stage of industrial production of PEN in

batch-fed cultures the organism experiences nutrient limita-tions that result in induction of autophagy in mature hyphalcompartments.

In P. chrysogenum, the biosynthetic pathway of �-lactamantibiotics is compartmentalized and is suggested to take placein mature, vacuolated subapical hyphal compartments (10, 27).Extensive autophagy-related degradation of cytosolic compo-nents and peroxisomes in vacuolated, late hyphal elements was

FIG. 6. Morphological analysis of P. chrysogenum WT and atg1 cells. Electron micrographs of KMnO4-fixed P. chrysogenum hyphae are shownin WT (A) and atg1 (B) cells. Subapical compartments of WT but not atg1 hyphae showed extensive autophagy characterized by large vacuolescontaining autophagic bodies and reduced numbers of organelles relative to atg1 cells. M, mitochondrion; P, peroxisome; V, vacuole; AV,autophagosome. Bars, 2 �m.

FIG. 7. Effects of autophagy impairment on antibiotic production. (A) Western blot analysis of crude extracts of P. chrysogenum WT and atg1strains. Mycelia were cultivated under PEN-inducing conditions. Equal amounts of protein were loaded per lane. Polyclonal antibodies againstIPNS, IAT proteins, and transcription elongation factor EF1-� (loading control) were used to decorate the blots. Enhanced levels of both IPNSand IAT were detected in atg1 extracts relative to those in WT cells at all time points studied. (B) Spent media of atg1 and WT strains cultivatedunder PEN-inducing conditions were diluted and loaded into wells of a bioassay plate that had been overlaid with the PEN-sensitive indicator strainM. luteus. Cell densities (in terms of the dry weight per liter) of both WT and atg1 cultures were comparable at all time points studied. After 24 hof incubation at 30°C, the sizes of the growth inhibition zones (halos) were determined. The data indicate that the atg1 strain produced increasedamounts of antibacterial agents relative to the WT strain at all time points studied. (C) NMR quantitative determination of secreted PenV levels.Spores of atg1 and WT strains were inoculated in PEN induction medium supplemented with phenoxyacetic acid. Samples were taken after 7 daysof cultivation. At this time point the sugars were completely consumed and cell densities (in terms of dry weight per liter) of both cultures werecomparable (data not shown). The amount of detected extracellular �-lactam antibiotic, PenV, was significantly increased in the atg1 null mutantrelative to the parental strain (Student’s t test, P 0.033). Scale bars represent the standard errors of means of three independent experiments.

1420 BARTOSZEWSKA ET AL. APPL. ENVIRON. MICROBIOL.

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

observed in these studies, which implies that autophagy couldhave an impact on the efficiency of PEN production. Indeed,our studies demonstrated that impairment of autophagy in P.chrysogenum led to significantly increased PenV production,possibly related to the increased amounts of PEN biosynthesisenzymes in the cells.

The elevated amounts of �-lactam antibiotic biosynthesisenzymes might positively affect PEN production (2, 38, 39, 43),but these enzyme levels are not the only factor that may impactproductivity of this secondary metabolite. The number of per-oxisomes may also play an important role in efficient PENproduction (14, 24, 41). Indeed, the disruption of atg1 results inan increased number of peroxisomes, since the autophagy-dependent removal of these organelles is impaired (1). Hence,it is likely that the presence of elevated numbers of peroxi-somes in subapical hyphal cells is one of the reasons for in-creased PEN production in cells of the atg1 strain.

The high-PEN-producing strains used now in industrialpenicillin fermentations are the result of time-consuming,extensive strain improvement programs. Application of ge-netic engineering has opened the possibility for targeted,comprehensive design of strains. We show here that a prom-ising way of improving PEN productivity in P. chrysogenummay be modulation of autophagy-related processes.

ACKNOWLEDGMENTS

This project was carried out within the research program of theKluyver Centre for Genomics of Industrial Fermentation, which is partof the Netherlands Genomics Initiative/Netherlands Organization forScientific Research. J.A.K.W.K. was supported by a grant from DSM,Delft, Netherlands.

We thank Rinse de Boer, Susan Fekken, and Wiebe H. Meijer forexcellent technical assistance.

REFERENCES

1. Aksam, E. B., et al. 2007. A peroxisomal lon protease and peroxisomedegradation by autophagy play key roles in vitality of Hansenula polymorphacells. Autophagy 3:96–105.

2. Brakhage, A. A. 1998. Molecular regulation of beta-lactam biosynthesis infilamentous fungi. Microbiol. Mol. Biol. Rev. 62:547–585.

3. Brakhage, A. A., et al. 2004. Regulation of penicillin biosynthesis in filamen-tous fungi. Adv. Biochem. Eng. Biotechnol. 88:45–90.

4. Cantoral, J. M., B. D. Díez, J. L. Barredo, E. Alvarez, and J. F. Martín. 1987.High-frequency transformation of Penicillium chrysogenum. Biotechnology(NY) 5:494–497.

5. Cheong, H., and D. J. Klionsky. 2008. Dual role of Atg1 in regulation ofautophagy-specific PAS assembly in Saccharomyces cerevisiae. Autophagy4:724–726.

6. Gidijala, L., et al. 2008. Production of functionally active Penicillium chry-sogenum isopenicillin N synthase in the yeast Hansenula polymorpha. BMCBiotechnol. 8:29.

7. Gidijala, L., I. J. van der Klei, M. Veenhuis, and J. A. Kiel. 2007. Repro-gramming Hansenula polymorpha for penicillin production: expression ofthe Penicillium chrysogenum pcl gene. FEMS Yeast Res. 7:1160–1167.

8. Harris, D. M., et al. 2006. Enzymic analysis of NADPH metabolism inbeta-lactam-producing Penicillium chrysogenum: presence of a mitochon-drial NADPH dehydrogenase. Metab. Eng. 8:91–101.

9. Hillenga, D. J., H. J. Versantvoort, A. J. Driessen, and W. N. Konings. 1994.Structural and functional properties of plasma membranes from the filamen-tous fungus Penicillium chrysogenum. Eur. J. Biochem. 224:581–587.

10. Justen, P., G. C. Paul, A. W. Nienow, and C. R. Thomas. 1998. Dependenceof penicillium chrysogenum growth, morphology, vacuolation, and produc-tivity in fed-batch fermentations on impeller type and agitation intensity.Biotechnol. Bioeng. 59:762–775.

11. Kamada, Y., T. Sekito, and Y. Ohsumi. 2004. Autophagy in yeast: a TOR-mediated response to nutrient starvation. Curr. Top. Microbiol. Immunol.279:73–84.

12. Kiel, J. A., V. I. Titorenko, I. J. van der Klei, and M. Veenhuis. 2007.Overproduction of translation elongation factor 1-alpha (eEF1A) suppressesthe peroxisome biogenesis defect in a Hansenula polymorpha pex3 mutantvia translational read-through. FEMS Yeast Res. 7:1114–1125.

13. Kiel, J. A., et al. 2009. Matching the proteome to the genome: the microbodyof penicillin-producing Penicillium chrysogenum cells. Funct. Integr.Genomics 9:167–184.

14. Kiel, J. A., I. J. van der Klei, M. A. van den Berg, R. A. Bovenberg, and M.Veenhuis. 2005. Overproduction of a single protein, Pc-Pex11p, results in2-fold enhanced penicillin production by Penicillium chrysogenum. FungalGenet. Biol. 42:154–164.

15. Kikuma, T., M. Ohneda, M. Arioka, and K. Kitamoto. 2006. Functionalanalysis of the ATG8 homologue Aoatg8 and role of autophagy in differen-tiation and germination in Aspergillus oryzae. Eukaryot. Cell 5:1328–1336.

16. Koetsier, M. J., et al. 2010. The Penicillium chrysogenum aclA gene encodesa broad-substrate-specificity acyl-coenzyme A ligase involved in activation ofadipic acid, a side-chain precursor for cephem antibiotics. Fungal Genet.Biol. 47:33–42.

17. Kolar, M., P. J. Punt, C. A. van den Hondel, and H. Schwab. 1988. Trans-formation of Penicillium chrysogenum using dominant selection markers andexpression of an Escherichia coli lacZ fusion gene. Gene 62:127–134.

18. Kristensen, A. R., et al. 2008. Ordered organelle degradation during starva-tion-induced autophagy. Mol. Cell Proteomics 7:2419–2428.

19. Liu, X. H., J. P. Lu, and F. C. Lin. 2007. Autophagy during conidia-tion, conidial germination and turgor generation in Magnaporthe grisea.Autophagy 3:472–473.

20. Liu, X. H., et al. 2007. Involvement of a Magnaporthe grisea serine/threoninekinase gene, MgATG1, in appressorium turgor and pathogenesis. Eukaryot.Cell 6:997–1005.

21. Meijer, W. H., I. J. van der Klei, M. Veenhuis, and J. A. Kiel. 2007. ATGgenes involved in non-selective autophagy are conserved from yeast to man,but the selective Cvt and pexophagy pathways also require organism-specificgenes. Autophagy 3:106–116.

22. Mukaiyama, H., et al. 2009. Autophagy-deficient Schizosaccharomycespombe mutants undergo partial sporulation during nitrogen starvation. Mi-crobiology 155:3816–3826.

23. Muller, W. H., et al. 1992. Involvement of microbodies in penicillin biosyn-thesis. Biochim. Biophys. Acta 1116:210–213.

24. Muller, W. H., et al. 1991. Localization of the pathway of the penicillinbiosynthesis in Penicillium chrysogenum. EMBO J. 10:489–495.

25. Nakatogawa, H., K. Suzuki, Y. Kamada, and Y. Ohsumi. 2009. Dynamics anddiversity in autophagy mechanisms: lessons from yeast. Nat. Rev. Mol. CellBiol. 10:458–467.

26. Opalinski, L., J. A. Kiel, T. G. Homan, M. Veenhuis, and I. J. van der Klei.Penicillium chrysogenum Pex14/17p: a novel component of the peroxisomalmembrane that is important for penicillin production. FEBS J. 277:3203–3218.

27. Paul, G. C., and C. R. Thomas. 1996. A structured model for hyphal differ-entiation and penicillin production using Penicillium chrysogenum. Biotech-nol. Bioeng. 51:558–572.

28. Pinan-Lucarre, B., A. Balguerie, and C. Clave. 2005. Accelerated cell deathin Podospora autophagy mutants. Eukaryot. Cell 4:1765–1774.

29. Pinan-Lucarre, B., M. Paoletti, K. Dementhon, B. Coulary-Salin, and C.Clave. 2003. Autophagy is induced during cell death by incompatibility andis essential for differentiation in the filamentous fungus Podospora anserina.Mol. Microbiol. 47:321–333.

30. Pollack, J. K., S. D. Harris, and M. R. Marten. 2009. Autophagy in filamen-tous fungi. Fungal Genet. Biol. 46:1–8.

31. Richie, D. L., et al. 2007. Unexpected link between metal ion deficiency andautophagy in Aspergillus fumigatus. Eukaryot. Cell 6:2437–2447.

32. Samara, C., P. Syntichaki, and N. Tavernarakis. 2008. Autophagy is re-quired for necrotic cell death in Caenorhabditis elegans. Cell Death Differ.15:105–112.

33. Sambrook, J. F., and T. Maniatis. 1989. Molecular cloning, a laboratorymanual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

34. Scott, R. C., G. Juhasz, and T. P. Neufeld. 2007. Direct induction ofautophagy by Atg1 inhibits cell growth and induces apoptotic cell death.Curr. Biol. 17:1–11.

35. Shoji, J. Y., M. Arioka, and K. Kitamoto. 2006. Possible involvement ofpleiomorphic vacuolar networks in nutrient recycling in filamentous fungi.Autophagy 2:226–227.

36. Shoji, J. Y., M. Arioka, and K. Kitamoto. 2006. Vacuolar membrane dynam-ics in the filamentous fungus Aspergillus oryzae. Eukaryot. Cell 5:411–421.

37. Snoek, I. S., et al. 2009. Construction of an hdfA Penicillium chrysogenumstrain impaired in non-homologous end-joining and analysis of its potentialfor functional analysis studies. Fungal Genet. Biol. 46:418–426.

38. Theilgaard, H., M. van Den Berg, C. Mulder, R. Bovenberg, and J. Nielsen.2001. Quantitative analysis of Penicillium chrysogenum Wis54-1255 trans-formants overexpressing the penicillin biosynthetic genes. Biotechnol. Bio-eng. 72:379–388.

39. Thykaer, J., and J. Nielsen. 2003. Metabolic engineering of beta-lactamproduction. Metab. Eng. 5:56–69.

40. Todde, V., M. Veenhuis, and I. J. van der Klei. 2009. Autophagy: principlesand significance in health and disease. Biochim. Biophys. Acta 1792:3–13.

41. van den Berg, M. A., et al. 2008. Genome sequencing and analysis of thefilamentous fungus Penicillium chrysogenum. Nat. Biotechnol. 26:1161–1168.

VOL. 77, 2011 AUTOPHAGY DEFICIENCY PROMOTES PENICILLIN PRODUCTION 1421

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from

42. van der Lende, T. R., et al. 2002. �-(L-�-Aminoadipyl)-L-cysteinyl-D-valinesynthetase, that mediates the first committed step in penicillin biosynthesis,is a cytosolic enzyme. Fungal Genet. Biol. 37:49–55.

43. Veenstra, A. E., P. van Solingen, R. A. Bovenberg, and L. H. van der Voort.1991. Strain improvement of Penicillium chrysogenum by recombinant DNAtechniques. J. Biotechnol. 17:81–90.

44. Waterham, H. R., et al. 1994. The Hansenula polymorpha PER1 gene is

essential for peroxisome biogenesis and encodes a peroxisomal matrix pro-tein with both carboxy- and amino-terminal targeting signals. J. Cell Biol.127:737–749.

45. Yorimitsu, T., and D. J. Klionsky. 2005. Autophagy: molecular machinery forself-eating. Cell Death Differ. 12(Suppl. 2):1542–1552.

46. Yu, L., et al. 2006. Autophagic programmed cell death by selective catalasedegradation. Proc. Natl. Acad. Sci. U. S. A. 103:4952–4957.

1422 BARTOSZEWSKA ET AL. APPL. ENVIRON. MICROBIOL.

on August 20, 2020 by guest

http://aem.asm

.org/D

ownloaded from