Global Phenotype Screening and Transcript Analysis Outlines the Inhibitory Mode(s) of Action of Two Amphibian-Derived, {alpha}-Helical, Cationic Peptides on Saccharomyces cerevisiae

Dermaseptin S3(1-16) [DsS3(1-16)] and magainin 2 (Mag 2) aretwo unrelated, amphibian-derived cationic peptides that adoptan ${alpha}$ -helical structure within microbial membranes and have beenproposed to kill target organisms via membrane disruption. Usinga combination of global deletion mutant library phenotypic screening,expression profiling, and physical techniques, we have carriedout a comprehensive in vitro analysis of the inhibitory actionof these two peptides on the model fungus Saccharomyces cerevisiae.Gene ontology profiling (of biological processes) was used toidentify both common and unique effects of each peptide. Resistanceto both peptides was conferred by genes involved in telomeremaintenance, chromosome organization, and double-strand breakrepair, implicating a common inhibitory action of DNA damage.Crucially, each peptide also required unique genes for maintainingresistance; for example, DsS3(1-16) required genes involvedin protein targeting to the vacuole, and Mag 2 required genesinvolved in DNA-dependent DNA replication and DNA repair. Thus,DsS3(1-16) and Mag 2 have both common and unique antifungalactions that are not simply due to membrane disruption. Physicaltechniques revealed that both peptides interacted with DNA invitro but in subtly different ways, and this observation wassupported by the functional genomics experiments that providedevidence that both peptides also interfered with DNA integritydifferently in vivo. This implies that both peptides are ableto pass through the cytoplasmic membrane of yeast cells anddamage DNA, an inhibitory action that has not been previouslyattributed to either of these peptides.

INTRODUCTION

Top

INTRODUCTION

Antimicrobial peptides have now been discovered from the microbial,animal, and plant kingdoms and are classified on the basis oftheir secondary structures (reviewed in reference 12). Thereis a perception that antimicrobial peptides kill microorganismsnonspecifically by disrupting the plasma membrane of targetcells with a detergent-like action. However, increasing evidenceindicates that the inhibitory actions of many peptides are complexand involve targets interior to the cytoplasmic membrane (51).Evidence from a range of studies has shown that an importanttarget for cationic peptides could be interference with thesynthesis or function of critical intracellular macromolecules,such as DNA and proteins.

Magainin 2 (Mag 2) from Xenopus laevis (53) and dermaseptinS3(1-16) [DsS3(1-16)] (31), a truncated derivative of DsS3 fromPhyllomedusa sauvagii (32) with full activity, are unrelated,cationic, ${alpha}$ -helical antimicrobial peptides that have been extensivelyinvestigated. Studies of Mag 2 have clearly shown that the peptidebinds to negatively charged bacterial membranes and inducesthe formation of a toroid pore that permeabilizes the membrane(29, 51). Similarly, studies on DsS3 have shown that this peptidealso binds to, and disrupts, membranes (15). Therefore, it isnot surprising that the principal mechanism of cell death inducedby these cationic peptides has been attributed to membrane disruption.However, membrane disruption alone may not be sufficient tokill microorganisms (reviewed in reference 51). Thus, in thiswork, we used functional genomic techniques to compare in detailthe inhibitory actions of Mag 2 and DsS3(1-16).

Genome-wide expression profiling has been employed extensivelyto determine the inhibitory effects of antimicrobials. The rationalefor this approach is that the mRNA profiles generated in responseto a particular antimicrobial represent the transcriptionalresponse of the cell to the inhibition of particular vital cellularfunctions, in effect, a signature of the type of stress induced.Research examining the response to antifungals includes studiesof the response of Candida albicans upon exposure to itraconazole(11), amphotericin B and nystatin (54), and, collectively, ketoconazole,amphotericin B, caspofungin, and flucytosine (25). More-sophisticatedapproaches have included the construction of a reference databaseor compendium of gene expression profiles occurring in Saccharomycescerevisiae as a consequence of either specific mutations ora range of chemical treatments (19). In addition, Bammert andFostel (6) used expression profiling to identify patterns ofgene expression convergent between drug-treated cells and cellswith genetic alterations in the same pathway targeted by thedrug and were able to assign a mode of action to a novel azoleantifungal.

Although this approach has been successful, problems with thisapproach include the assumption that a single interaction betweena drug and a target accounts for the inhibitory action of thedrug and that drug treatment also changes the expression ofmany genes that are not directly related to the drug targetand are thus functionally irrelevant. Thus, for more-precisemechanism-of-action studies, it would be beneficial to combineand complement expression profiling with functional screeningat the level of the entire genome. In this work, we combineda genome-wide phenotypic screen, using the S. cerevisiae deletionmutant library, with expression profiling to characterize theinhibitory action of two unrelated cationic, ${alpha}$ -helical peptideswhose principal inhibitory action is believed to be disruptionof microbial cell membranes. The data imply that the inhibitoryactions of these two peptides are very different and far morecomplex than simple disruption of the cellular membrane.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Antimicrobial peptides. DsS3(1-16)-NH₂ was synthesized according to the published sequenceALWKNMLKGIGKLAGK (31) and Mag 2 according to the sequence GIGKFLHSAKKFGKAFVGEIMNS(53) by Peptide Protein Research Ltd., Wickham, United Kingdom,to >95% purity and verified by high-performance liquid chromatographyand mass spectrometry. Experiments were carried out using astock solution of 50 mg ml^–1 in water.

Yeast strains and media. The yeast strain used in this study was S. cerevisiae BY4741(MATa his- ${Delta}$ 31 leu- ${Delta}$ 0 met- ${Delta}$ 150 ura- ${Delta}$ 30) from Research Genetics (Huntsville,AL). Gene deletion mutants were from the Research Genetics BY4741MATa haploid genome deletion mutant set. Malt extract broth(MEB; 0.6% malt extract, 0.6% dextrose, 0.18% maltose, 0.12%yeast extract), pH 7.0 (Difco), was used for all growth assays.

Screening of the yeast deletion mutant library. The S. cerevisiae genome gene deletion mutant set was screenedfor sensitivity and resistance to DsS3(1-16) and Mag 2. By useof a 48-pin replica plater, individual yeast mutant strainswere inoculated from the strain collection stored in 96-wellplates into sterile 96-well plates containing 200 µl MEB,pH 7.0, with G418 sulfate (1.5 µg ml^–1). The cultureswere incubated overnight at 30°C until they had grown toan optical density at 600 nm (OD₆₀₀) of 0.3, and these weretransferred into 96-well plates containing MEB, pH 7.0, with6 µg ml^–1 DsS3(1-16) or 15 µg ml^–1 Mag2 by using the replica plater as described above. The cultureswere agitated to resuspend the cells, and the OD₆₀₀ was measuredby using a Powerwave XS universal microplate spectrophotometer(BIO-TEK Instruments Inc., Winooski, VT). Deletion mutants showingan OD₆₀₀ significantly higher or lower than that of the isogenicparent were selected for detailed screening by monitoring growthturbidometrically with the Powerwave XS spectrophotometer inthe presence or absence of antimicrobial peptide. Growth wasmeasured in 48-well microtiter plates (Greiner Bio-One) withwells containing 300 µl MEB, pH 7.0, with or without 6µg ml^–1 DsS3(1-16) or 15 µg ml^–1 Mag2; S. cerevisiae culture was added to each sample well to givea final starting inoculum of 5 x 10³ cells ml^–1. Growthwas measured by determining the change in OD₆₀₀ every 15 minover 24 h at 30°C with continuous shaking. The data generatedwere imported into Excel software (Microsoft, Seattle, WA),and growth curves were generated.

RNA preparation. Yeast cultures were grown in 150 ml MEB, pH 7.0, in 250-ml flasksovernight at 30°C to an OD₆₀₀ of 0.25. The cells were treatedwith either 7 µg ml^–1 DsS3(1-16) or 15 µgml^–1 Mag 2 for 20 and 40 min. Cells were harvested andresuspended in 200 µl sterile water, and the suspensionwas dropped directly into liquid nitrogen to form cell pellets.The cell pellets were ground with a mortar and pestle underliquid nitrogen to form a fine powder which was taken up in2 ml TRIzol (Invitrogen). The cell paste was subjected to avortex for 1 min, incubated for 5 min at room temperature, andthen centrifuged at 12,000 x g for 10 min. The clear supernatantwas mixed with 0.4 volumes of chloroform, shaken for 15 s, andincubated for 10 min at room temperature. The mixture was centrifugedat 12,000 x g for 10 min, and the supernatant was combined with0.5 volumes of isopropanol to precipitate the RNA. This wasincubated for 10 min at room temperature and centrifuged at12,000 x g for 10 min to pellet the RNA. The isopropanol wastaken off, and the pellet was washed with 70% ethanol. The pelletwas air dried and resuspended in 500 µl diethyl pyrocarbonate-treatedwater.

Microarray hybridization and analysis. Microarray hybridization was performed using the AffymetrixGeneChip Yeast Genome S98 array as described previously (49).Microarray scanner output (in the form of .cel files) was normalizedby the robust multiarray average (RMA) method (22), using defaultparameters in R/Bioconductor software (www.bioconductor.org).Results were expressed in log₂ form. The results were trimmedto remove genes where the values in the paired control chipsdiffered from each other by more than 25%.

Ontology profiling of gene expression changes occurring in responseto peptide treatments was performed using the Gene Ontology(GO) Term Finder tool (http://db.yeastgenome.org/cgi-bin/GO/goTermFinder)in the Saccharomyces Genome Database (http://www.yeastgenome.org/).Only the top 10 most significant GO terms are shown (P <0.05).

Peptide-DNA interactions. The plasmid DNA pRS313 was extracted from an Escherichia coliculture by using a QIAprep Spin Miniprep kit from QIAGEN. DNAwas resuspended in 10 mM sodium phosphate buffer, pH 7.0, andconcentrations were determined by measuring the OD₂₆₀ in a Varianspectrophotometer. All gels were visualized using a UV Transilluminatorfrom UVP and photographed with the BioDoc-It imaging system.

Gel retardation experiments. The DNA binding assay was performed as described in the studyof Park et al. (40) except 100 ng of pRS313 (previously digestedovernight with BamHI) was used.

DNA strand break assay. The DNA strand break assay was performed by measurement of relaxationof supercoiled plasmid DNA (pRS313) to an open circular form(41). The reaction mixtures were prepared as described for theretardation experiments except that the plasmid DNA was notdigested with BamHI.

DNase I protection assay. The DNase I protection assay was performed according to themethods of Niidome et al. (37). A total of 500 ng of pRS313(previously digested overnight with BamHI) was mixed with increasingconcentrations of Mag 2 in 20 µl of binding buffer.

Measurement of DNA damage by quantitative PCR (qPCR) assay. The assay was exactly as described by Morton et al. (34). Wearbitrarily selected the S. cerevisiae gene IZH3. A 10-kb genomicsequence upstream and downstream of IZH3 was obtained from theSaccharomyces Genome Database, and forward and reverse primerswere designed using software made available by Integrated DNATechnologies and synthesized by MWG Biotech, Germany.

CD spectroscopy. Circular dichroism (CD) spectroscopy was conducted on combinationsof the plasmid pRS313 (50 µg ml^–1) and antimicrobialpeptide DsS3(1-16) or Mag 2 (250 µg ml^–1). CD spectrawere measured in the far-UV (260- to 180-nm) range using a JascoJ-810 spectropolarimeter. pRS313 and the peptides were dissolvedin 10 mM sodium phosphate buffer, pH 7.0. For each experiment,a buffer-blank spectrum was subtracted from the spectrum ofthe peptide plus the buffer, and a plasmid-blank spectrum wassubtracted from the spectrum from the peptide-and-plasmid experiments.Data were also analyzed using the CDSSTR algorithm (http://public-1.cryst.bbk.ac.uk/cdweb/html/)(47) to show the secondary structural motifs for each peptide.

Microarray data accession number. In compliance with MIAME guidelines (http://www.mged.org/Workgroups/MIAME/miame.html),the data from this study have been submitted to the ArrayExpressrepository at the European Bioinformatics Institute website(http://www.ebi.ac.uk/arrayexpress/) under accession no. E-MEXP-1060.

RESULTS

Top

RESULTS

Phenotype screening identified many gene deletions conferring resistance and sensitivity to DsS3(1-16) and Mag 2. To gain a detailed understanding of the antifungal modes ofaction of DsS3(1-16) and Mag 2, we carried out a phenotypicscreening of mutants with the 4,847 nonessential gene deletionsin S. cerevisiae BY4741 MATa.

Deletion mutants displaying sensitivity. Of the 4,847 gene deletion mutants tested, 32 (0.7%) showedhypersensitivity to both DsS3(1-16) and Mag 2, 82 (1.7%) weresensitive to DsS3(1-16) alone, and 21 (0.4%) were sensitiveto Mag 2 alone (Table 1). We used the Saccharomyces Genome DatabaseGO Term Finder to identify significantly (P < 0.05) associatedgene groups in terms of biological process ontology. Hence,we conducted searches on the 32 gene deletion mutants showingcommon sensitivity to both peptides, thus representing commonmediators of resistance to ${alpha}$ -helical, cationic antifungal peptides.The most significant hits were represented by the terms "telomereorganization and maintenance" (9 out of 32 genes [29%] [P =1.38E–6]; VPS54, NUP60, RTF1, ARV1, RSC2, SRB5, CTK1,SNF7, and LDB7), "chromosome organization and biogenesis" (11out of 32 genes [35.4%] [P = 2.52E–6]; SGF29, VPS54, NUP60,RTF1, ARV1, RSC1, RSC2, SRB5, CTK1, SNF7, and LDB7), "lipidtransport" (3 out of 32 genes [9.6%] [P = 0.00037]; OSH3, CDC50,and ARV1), and "double-strand break repair" (3 out of 32 genes[9.6%] [P = 0.00125]; MMS22, RSC1, and RSC2). The observationthat a number of deletions in genes involved in maintenanceof chromosome integrity and DNA repair induce sensitivity toboth peptides implies that a common inhibitory action of thesemolecules could be due to the induction of DNA damage. Examplesincluded the deletions of MMS22, which acts in a double-strandbreak repair pathway that may be involved in resolving replicationintermediates or preventing the damage caused by blocked replicationforks (7); RSC1, RSC2, and LDB7, which are constituents of theRSC (remodel the structure of chromatin) complex involved inchromatin structure remodeling and double-strand break repairvia nonhomologous end joining (43, 48); and IMP2', a transcriptionalactivator believed to be involved in regulating ion homeostasisand DNA repair in response to DNA-damaging agents (28).

View this table:
[in this window]
[in a new window]

TABLE 1. Screening of the yeast disruptome revealed genes whose deletions confer sensitivity to the antimicrobial peptides DsS3(1-16) and Mag 2

Next, we studied the gene deletions that conferred sensitivityexclusively to either DsS3(1-16) or Mag 2 alone, thus representingindicators of unique, peptide-specific effects on the cells.Thus, ontology profiling (biological processes) of the genedeletions conferring sensitivity to DsS3(1-16) alone revealedsignificant hits for "vacuolar transport" (9 out of 82 genes[10.9%] [P = 1.14E–5]; VPS8, VPS61, VAC8, VAM3, VPS69,VMA6, VMA5, CUP5, and GGA1), "protein targeting to the vacuole"(5 out of 82 genes [6.0%] [P = 0.00099]; VPS61,VAC8, VPS69,CUP5, and GGA1), and "vacuolar acidification" (3 out 82 genes[3.6%] [P = 0.0026]; VMA6, VMA5, and CUP5). This suggests thatoptimal function of the vacuolar H⁺-ATPase is required specificallyfor resistance to the inhibitory effects of DsS3(1-16); perhapsto enable sequestration, and ultimately destruction of the inhibitorypeptide in the vacuole. However, other gene deletions that conferredspecific sensitivity to DsS3(1-16) included the deletions ofMRE11, encoding a protein vital for double-strand break repairand telomere maintenance (24); DPB3, encoding the third-largestsubunit of DNA polymerase II, capable of mismatch and nucleotide-excisionrepair, and required to maintain the fidelity of chromosomalreplication (21); and CTF4, encoding a chromatin-associatedprotein, required for sister chromatid cohesion (16).

In contrast to profiling of the gene deletions conferring sensitivityto DsS3(1-16), ontology profiling (biological processes) ofthe gene deletions conferring specific sensitivity to Mag 2revealed top hits for "second-messenger-mediated signaling"(3 out of 21 genes [14.2%] [P = 8.77E–05]; RCN1, SAC1,and SOK1), "endosome transport" (2 out of 21 genes [9.5%] [P= 0.00648]; VPS53 and ENT5), and "cell wall organization andbiogenesis" (3 out of 21 genes [14.2%] [P = 0.00994]; ROM1,SPR3, and CWH41).

Notably, mutants with deletions of genes known to be involvedin fungal drug resistance were not identified in large numbers,indicating that the inhibitory mode of action of these peptidesis different to well-known and characterized antifungal agents.Exceptions that induced sensitivity to DsS3(1-16) alone includeddeletions of RDS1, a transcriptional activator known to conferresistance to cycloheximide (1), and BFR1, a multicopy suppressorof brefeldin A sensitivity. The deletion of a gene already knownto be involved in drug resistance mechanisms, RCY1, conferredsensitivity to both peptides. RCY1 encodes an F-box proteininvolved in recycling plasma membrane proteins internalizedby endocytosis and confers resistance to the antifungals brefeldinA and monensin (36).

Intriguingly, the deletion of two genes, KTI12 and FYV12, knownto mediate resistance to yeast killer toxins, conferred sensitivityon DsS3(1-16) alone. KTI12 encodes a protein that associateswith a histone acetyltransferase complex that functions duringtranscription and was first identified in a screening for mutantsinsensitive to the Kluyveromyces lactis toxin zymocin, whichcauses cell cycle arrest in G₁ phase (5), and FYV12, a proteinof unknown function, is required for survival upon exposureto the K1 killer toxin (39).

Deletion mutants displaying resistance. Study of gene deletions that conferred resistance revealed thatno deletions conferred resistance to both DsS3(1-16) and Mag2, but 33 mutants (0.5%) were resistant to DsS3(1-16) aloneand only 4 (0.08%) were resistant to Mag 2 alone (Table 2).Notably, the deletion of a number of genes known to be involvedin the induction of programmed cell death in yeast (includingIZH2, IZH3, STM1, and AIF1) led to resistance to DsS3(1-16).Subsequently, we recently reported that DsS3(1-16) induced caspase-independentbut AIF-dependent apoptosis in yeast (34), and thus, we do notdiscuss the role of these genes further here.

View this table:
[in this window]
[in a new window]

TABLE 2. Screening of the yeast disruptome revealed genes whose deletions confer increased resistance to the antimicrobial peptides DsS3(1-16) and Mag 2^a

Notably, deletion of different genes involved in telomere maintenanceresulted in resistance to DsS3(1-16) and Mag 2. Resistance toDsS3(1-16) was induced by deletion of HST3, which encodes amember of the Sir2p family of NAD⁺-dependent protein deacetylasesthat are involved in telomeric silencing, radiation resistance,genomic stability, and short-chain fatty acid metabolism (8),and NNT1, a putative nicotinamide N-methyltransferase gene thatplays a role in ribosomal DNA silencing and in life span determination,deletion of which results in decreased telomeric silencing (3).Deletion of NMD2 induced resistance to Mag 2. NMD2 is also involvedin telomere maintenance, and its deletion confers increasedresistance to the anticancer drug cisplatin (4). Also, deletionof three genes involved in chromosome/chromatid segregationand cohesion, GLC8, SRC1, and SPO13, induced resistance to DsS3(1-16).

Related to a common theme of possible peptide interference withDNA, deletion of TOP1 induced specific resistance to Mag 2.TOP1 encodes the type IB topoisomerase that cleaves one DNAstrand and relaxes both positively and negatively supercoiledDNA. Top1p is the target of the antitumor drug camptothecinthat induces double-strand breaks during DNA replication (5).

Our findings from the genome deletion mutant screen analysissuggest, but do not explicitly define, the relative importanceof each deletion phenotype observed. Neither necessity nor sufficiencyof any candidate gene has yet been validated by complementation;thus, the potential cause-versus-effect relationships hypothesizedhave not yet been confirmed. This will form the basis of futuremore-detailed study of the precise role key genes play in mediatingresistance to the individual peptides.

In summary, the deletion mutant screen has indicated that DsS3(1-16)and Mag 2 have common inhibitory effects that may involve interferencewith DNA, chromosome integrity, and telomere maintenance. However,despite both peptides being cationic, with similar ${alpha}$ -helicalstructures, they also induce very different resistance/sensitivitypatterns in the global gene deletion mutant screen. Thus, itis also clear that each peptide has unique, independent antifungaleffects.

Analysis of up-regulated genes reveals both common and distinct gene expression profiles upon exposure to DsS3(1-16) or Mag 2. To obtain gene expression profiles occurring as a consequenceof sublethal exposure to DsS3(1-16) and Mag 2, concentrationsof each peptide, 7 and 15 µg ml^–1, respectively,that partially inhibited growth in shake-flask cultures andhad only a minor detrimental effect on cell viability were selected(data not shown).

Up-regulated genes common to both peptides. A total of 164 common genes were found to be up-regulated (>1.5-foldincrease in expression) upon treatment with DsS3(1-16) (7 µgml^–1) or Mag 2 (15 µg ml^–1) at both exposuretimes tested, 20 and 40 min postaddition (see Table SA in thesupplemental material). As described above, these gene expressionchanges represent a response to common effects induced by bothpeptides. GO profiling revealed that the most significant geneclusters identified belonged to the biological process terms"cellular catabolism," "ubiquitin-dependent protein catabolism,""modification-dependent protein catabolism," and "proteolysisduring cellular protein catabolism" (Table 3). Thus, exposureto both peptides induces pathways resulting in ubiquitin-dependentprotein breakdown and proteolysis. This response could be toremove excess inhibitory peptide directly or could be occurringas a consequence of the stress response induced by the peptidesaffecting endogenous cellular functions.

View this table:
[in this window]
[in a new window]

TABLE 3. GO analysis of biological process genes >1.5-fold upregulated by treatment with DsS3(1-16) and Mag 2^a

Comparison of genes up-regulated by both peptides with the genedeletion mutant sensitivity data revealed only one gene, HAL5,whose deletion conferred sensitivity to both peptides and whoseexpression was up-regulated by both peptides at both time pointstested [for DsS3(1-16), 5.41-fold at 20 min and 6.41-fold at40 min; for Mag 2, 5.32-fold at 20 min and 4.8-fold at 40 min).HAL5 encodes a protein kinase that activates the Trk1-Trk2ppotassium transporters, increasing the influx of potassium anddecreasing the membrane potential. The resulting loss in electricaldriving force reduces the uptake of toxic cations and increasessalt tolerance. Indeed, overexpression of HAL5 enhances thetolerance of yeast cells to sodium and lithium, whereas genedisruption results in greater cation sensitivity (35). Notably,because HAL5 regulates the membrane potential, it is known tobe a pleiotropic determinant of sensitivity to other chemotherapeuticagents (45). Thus, it could be that the membrane potential isa significant mediator of the efficacies of both peptides. Thisconclusion is further supported by the observation that othergenes involved in cation homeostasis, for example, HAL9, CUP2,PTK2, ENA2, BSD2, ENA5, MNR2, and PPZ1, are up-regulated byboth peptides (see Table SA in the supplemental material).

A number of genes were notable for their strong up-regulationby both peptides (defined as an average increase in expressiongreater than sevenfold over both time points tested, 20 and40 min), for example, SPG4 (encoding a protein with unknownfunction but required for survival at high temperature duringstationary phase) (27), YER185W (encoding a transmembrane proteinwith unknown function), FMP48 (encoding a protein kinase withunknown function and localized to the mitochondria), PUT1 (encodingproline oxidase, a nuclear-encoded mitochondrial protein involvedin utilization of proline as the sole nitrogen source) (18),YDR034W-B (encoding a protein of unknown function whose greenfluorescent protein-fusion localizes to the cell periphery)(20).

Genes specifically up-regulated by DsS3(1-16). A total of 70 genes were found to be uniquely up-regulated (>1.5-foldincrease in expression) upon treatment with DsS3(1-16) (7 µgml^–1) alone at both exposure times tested, 20 and 40 minpostaddition (see Table SB in the supplemental material). Thesegene expression changes represent a response to specific effectsinduced by DsS3(1-16). GO profiling revealed that the most significantgene clusters identified belonged to the biological processterms "neutral amino acid transport," "response to stimulus,""response to reactive oxygen species," "actin cytoskeleton organizationand biogenesis," and "response to drug" (Table 4). The "responseto stimulus" cluster includes genes induced by general stressas well as genes induced by oxidative stress but also includesgenes such as PHR1, encoding a photolyase that repairs pyrimidinedimers in DNA in the presence of visible light; HEX3, involvedin genome stability and induced by DNA damage; and SIN3, involvedin chromosome integrity. The drug treatment cluster includesthree genes encoding ATP-binding cassette (ABC) transporterproteins involved in multidrug resistance; SNQ2, which alsoconfers resistance to singlet oxygen species (46); PDR5, whichactively exports various drugs across the membrane and is alsoinvolved in steroid transport, cation resistance, and cellulardetoxification during exponential growth (26); and finally,PDR10, encoding a protein similar to that encoded by PDR5, alsoinvolved in the pleiotropic drug resistance network.

View this table:
[in this window]
[in a new window]

TABLE 4. GO analysis of biological process genes upregulated >1.5-fold by DsS3(1-16) treatment^a

Comparison of genes up-regulated by exposure (at both time pointstested) to DsS3(1-16) alone with genes whose deletion has conferredspecific sensitivity to this peptide revealed three linked genes.MSC1 encodes a protein of unknown function, but deletion ofthe gene results in a defect in directing meiotic recombinationevents to homologous chromatids (44). The other two genes, YGR122Wand YPR115W, encode proteins of unknown function.

Notably, the genes showing the highest levels of induction specificallyby DsS3(1-16) at both time points, GPA2 (4.5- and 3.2-fold inductionat 20 min and 40 min, respectively), encoding the nucleotide-bindingalpha subunit of a heterotrimeric G protein that interacts withthe receptor Gpr1p and plays a role in signaling changes innutrient levels (13), and YAK1 (4.8- and 3.7-fold inductionat 20 min and 40 min, respectively), encoding a protein kinasethat is part of a glucose-sensing system involved in growthcontrol in response to glucose availability (33), are both involvedin signaling nutrient availability.

Genes specifically up-regulated by Mag 2. A total of 315 genes were found to be uniquely up-regulated(>1.5-fold increase in expression) upon treatment with Mag2 (15 µg ml^–1) alone at both exposure times tested,20 and 40 min postaddition (see Table SC in the supplementalmaterial). These gene expression changes represent a responseto specific effects induced by Mag 2. GO profiling revealedthat the most significant gene clusters belonged to the biologicalprocess terms "DNA replication," "DNA-dependent DNA replication,""DNA strand elongation," "DNA repair," and "maintenance of fidelityduring DNA-dependent DNA replication" (Table 5). Clearly, theserelated GO terms imply that Mag 2 has a specific effect on thesynthesis of new strands of DNA, using parental DNA as a templatefor the DNA-dependent DNA polymerases that synthesize the newstrands. The GO term "maintenance of fidelity during DNA-dependentDNA replication" indicates that Mag 2 may be directly interferingwith replication, causing DNA damage that has to be repairedby the genes present in the categories for the other significantGO terms "mismatch repair," "nucleotide-excision repair," and"double-strand break repair." Thus, these results imply thatMag 2 may have a more specific and potent DNA-damaging actionthan DsS3(1-16).

View this table:
[in this window]
[in a new window]

TABLE 5. GO analysis of biological process genes upregulated >1.5-fold by Mag 2 treatment^a

Analysis of genes up-regulated by exposure (at both time pointstested) to Mag 2 alone and gene deletions conferring specificsensitivity to this peptide revealed two genes: MRM2, encodinga mitochondrial 21S rRNA methyltransferase involved in telomeremaintenance and required for methylation of 21S rRNA (42), andROM1, encoding a GDP/GTP exchange protein that negatively regulatesRho1p, a protein that itself regulates protein kinase C andthe cell wall-synthesizing enzyme 1,3-beta-glucan synthase (38).

Notable genes showing the highest level of induction by Mag2 alone included MSH2 (3.1- and 2.0-fold induction at 20 minand 40 min, respectively), a mitochondrial DNA-binding proteininvolved in mismatch repair of mitochondrial DNA (30); SPG1(2.1- and 3.89-fold induction at 20 min and 40 min, respectively),encoding a protein with unknown function that is required forsurvival at high temperatures during stationary phase (27);TPO3 (3.15- and 2.07-fold induction at 20 min and 40 min, respectively),encoding a plasma membrane polyamine transport protein specificfor spermine (2); and SSH4 (3.54- and 2.31-fold induction at20 min and 40 min, respectively), encoding a protein with unknownfunction whose overexpression confers resistance to the anti-inflammatorydrug leflunomide (14).

Analysis of down-regulated genes upon exposure to DsS3(1-16) and Mag 2. (i) Down-regulated genes common to both peptides. A total of 64 common genes were found to be down-regulated (>1.5-folddecrease in expression) upon treatment with DsS3(1-16) (7 µgml^–1) or Mag 2 (15 µg ml^–1) at both exposuretimes tested, 20 and 40 min postaddition (see Table SD in thesupplemental material). As described above, these expressionchanges represent a response to common effects induced by bothpeptides. GO profiling revealed that the most significant geneclusters identified belonged to the categories for the biologicalprocess terms "ribosome biogenesis and assembly" (P = 5.65E–20),"rRNA processing" (P = 2.25E–10), and "organelle organizationand biogenesis" (P = 2.85E–09) (see Table SE in the supplementalmaterial). This profiling indicates that both peptides havea general effect of shutting down protein synthesis in responseto stress via loss of ribosomes. Genes that encode proteinswith functions in ribosome structure, function, or biogenesishave been shown previously to be coordinately and transientlyrepressed by multiple environmental stress factors (9). Ribosomesynthesis is a major cellular biosynthetic activity, and therepression of these ribosomal components under stress conditionshas been proposed to free significant energy resources for othercellular processes.

LDB7 was the only gene that was down-regulated by both peptides(at 20 and 40 min exposure), and the corresponding deletionstrain also displayed sensitivity to both peptides. LDB7 isa component of the RSC complex involved in chromatin structureremodeling and double-strand break repair via nonhomologousend joining (48). Notably, in a screen for mutants conferringa low-dye-binding phenotype for the cationic dye Alcian Blue,LDB7 was isolated (10). While the molecular function of LDB7remains unknown, our demonstration that both peptides down-regulatethe expression of the gene could be explained by the abilityof LDB7 to enhance binding of basic, positively charged compoundssuch as cationic peptides, which would clearly be detrimental.However, this does not explain why deletion of LDB7 conferssensitivity to both peptides.

(ii) Genes specifically down-regulated by DsS3(1-16) or Mag 2. Space constraints do not allow detailed discussion or presentationof these results. Ontology profiling of genes specifically down-regulatedby DsS3(1-16) and by Mag 2 at both time points is shown in TablesSF and SG in the supplemental material, respectively. Confirmingthe observations made with up-regulated transcripts, the expressionprofiles for specifically down-regulated genes were very differentfor each peptide. The top biological process hits for DsS3(1-16)were "rRNA metabolism" (P = 1.53E–26), "ribosome biogenesis"(P = 4.51E–16), and "cytoplasm organization and biogenesis"(P = 5.59E–16). This profile indicated specific effectsof DsS3(1-16) similar to those that were identified as commonto both peptides, e.g., shut-down of ribosome biosynthesis.In contrast, the top process GO hits for Mag 2 were "regulationof metabolism" (P = 4.26E–05), "glycolysis" (P = 0.00011),and "external encapsulating structure organization and biogenesis"(P = 0.00013). These results indicated a general shut-down ofcellular metabolism and physiological processes in the presenceof Mag 2. Thus, these data support our previous observationsthat the individual peptides have very different effects oncells.

Mag 2 and DsS3(1-16) interact differently with DNA in vitro. The deletion mutant screen indicated that genes involved inmediating chromosome integrity and DNA damage repair were importantcommon mediators of sensitivity to both peptides. Furthermore,ontology profiling revealed that the genes most significantlyspecifically up-regulated by Mag 2 alone included genes involvedin DNA replication and repair. Thus, we explored the interactionof DsS3(1-16) and Mag 2 with DNA in more detail.

Recently, we reported that DsS3(1-16), but not Mag 2, inducesprogrammed cell death in yeast (34). In vitro, it was shownthat DsS3(1-16) retarded the electrophoretic mobility of DNAin an agarose gel and protected DNA from DNase I digestion butdid not induce direct DNA damage via double-stand breaks. Thus,we concluded that DsS3(1-16) was capable of interacting withDNA in vitro. To determine if exposure to DsS3(1-16) could induceDNA damage in vivo, we measured DNA damage by a qPCR assay (40).This assay revealed that DsS3(1-16) did induce some DNA damagein vivo, probably due to the onset of programmed cell deathand associated chromosome breakdown (measured by terminal deoxynucleotidyltransferase-mediateddUTP-biotin nick end labeling assay) (34). This may explainwhy, in this study, we identified many deletions of genes involvedin DNA repair and DNA integrity that were sensitive to the peptideand why the expression of some genes with similar functionswas induced upon exposure to DsS3(1-16).

To explain why Mag 2 induced many more genes involved with DNArepair and DNA replication than DsS3(1-16) did, we carried outthe same assays described above with Mag 2 (Fig. 1). In vitro,we observed that Mag 2 retarded the migration of DNA in an agarosegel (Fig. 1a) in a fashion similar to that of DsS3(1-16) (34).Significantly, Mag 2 was able to completely retard the migrationof pRS313 into the gel at a lower DNA-to-peptide ratio (1:100)than DsS3(1-16) (34), indicating that the interaction of Mag2 with DNA was stronger than DsS3(1-16). The presence of Mag2 also resulted in plasmid DNA remaining in the loading wells,apparently unable to move into the gel (Fig. 1a). To study thisinteraction further, we also performed a DNase I protectionassay with the plasmid in the presence of Mag 2 (37). Thus,if the peptide binds to the plasmid DNA, it would be expectedthat the digestion of the DNA by DNase I would be inhibited.After adding Mag 2 to the plasmid DNA at various ratios, themixtures were then incubated with DNase I (Fig. 1b). Consistentwith the previous observation, DNA in the presence of Mag 2at a ratio of 1:50 or higher resulted in the complete protectionof pRS313 from digestion. Also, Mag 2 was able to protect pRS313from DNase at a lower DNA-to-peptide ratio (1:100) (Fig. 1b)than DsS3(1-16) (34), confirming that the interaction of Mag2 with DNA was stronger than DsS3(1-16).

View larger version (53K):
[in this window]
[in a new window]

FIG. 1. Mag 2 interacts with DNA in vitro and induces DNA damage. (a) Results from a gel retardation assay showing the effect of increasing concentrations of Mag 2 on the electrophoretic mobility of plasmid DNA (pRS313). Ratios indicate increasing peptide-to-plasmid concentrations. M, molecular weight markers (1-kb ladder). "pDNA" indicates migration of pRS313 without the presence of Mag 2. (b) The left panel shows results from a DNase protection assay showing the protection of pRS313 from digestion by 0.5 µg ml^–1 DNase I in the presence of increasing concentrations of Mag 2. Ratios indicate increasing peptide-to-plasmid concentrations. The right panel shows results from a control experiment, showing the digestion of pRS313 after exposure to 0.1 and 0.01 U µl^–1 DNase I. M, molecular weight markers. "pDNA" indicates the electrophoresis of pRS313 without the presence of Mag 2 or DNase I. (c) DNA strand break assay results showing the effect of exposure to increasing concentrations of Mag 2 on the relative abundance of supercoiled-form (SF) and relaxed-form (RF) bands of pRS313. In each case, representative results are shown.

To explore whether the interaction of Mag 2 with DNA could resultin direct damage, we performed DNA strand break assays (41).Strand breaks were assayed by measuring the relaxation of supercoiledplasmid to an open circular form. Exposure of plasmid DNA toincreasing concentrations of Mag 2 resulted in the gradual lossof supercoiled bands at DNA-to-peptide ratios of 1:12 and above(Fig. 1c). Therefore, in contrast to the interaction of DsS3(1-16)with DNA (34), interaction of Mag 2 with DNA in vitro is strongenough to induce direct DNA damage in the form of strand breaks.

To determine if exposure to Mag 2 could induce DNA damage invivo, we measured DNA damage by use of a qPCR assay (41). Thismethod detects and quantifies DNA damage because if damagedDNA is used as a template, this blocks the progression of thePCR DNA polymerase and results in the production of fewer PCRproducts than equivalent undamaged DNA (50). Thus, prior exposureof cells to Mag 2 resulted in no significant reduction in specificPCR products relative to an untreated control (data not shown).Exposure to 5 mM H₂O₂ resulted in the virtual abolition of PCRproducts. Therefore, despite Mag 2 inducing strand breaks invitro, we were unable to detect DNA damage in vivo. Thus, itseems likely that the huge up-regulation of genes involved inDNA repair and DNA replication that we detected upon exposureto Mag 2 occurs in response to the presence of the peptide targetingthe DNA and, in this case, allows the cells to adapt and preventdetectable Mag 2-induced DNA damage in vivo.

To gain further insight into the mechanism of interaction ofDsS3(1-16) and Mag 2 with DNA, we used CD. Far-UV CD spectroscopywas performed with DsS3(1-16) and Mag 2 either alone or in thepresence of pRS313 (Fig. 2a and b). Without DNA, applicationof the CDSSTR algorithm revealed that the two peptides havesimilar CD spectra that indicate they are largely in an unorderedstate with some ß-sheet (Fig. 2c). However, the additionof Mag 2 in the presence of DNA results in a modification ofthe CD spectrum that can be interpreted as a change in the secondarystructure content of the protein (Fig. 2b). The CDSSTR algorithmrevealed that in the presence of DNA, Mag 2 adopts an ${alpha}$ -helicalstructure (Fig. 2c). In contrast, in the case with DsS3(1-16),there was little evidence of change in secondary structure uponexposure to DNA; the peptide remained largely unordered. Detailedexamination of the interaction of the peptides with DNA wasnot the purpose of this study, but the CD analysis has shownthat the two peptides interact with DNA in solution in differentways.

View larger version (27K):
[in this window]
[in a new window]

FIG. 2. CD revealed that, unlike DsS3(1-16), Mag 2 adopts an ${alpha}$ -helical conformation upon binding to DNA. CD spectroscopy of peptides (250 µg ml^–1) binding to DNA (pRS313; 50 µg ml^–1) is shown. Far-UV spectra of DsS3(1-16) and Mag 2 with 10 mM sodium phosphate buffer, pH 7.0 (the spectrum shown constitutes a buffer-blank spectrum subtracted from the spectrum of the peptide plus the buffer) (a), and pRS313 DNA in 10 mM sodium phosphate buffer, pH 7.0 (the spectrum shown constitutes a DNA-blank spectrum subtracted from the spectrum of the peptide plus the DNA) (b). Results are expressed as mean residue ellipticity relative to the molar concentration of peptide. Representative results are shown. (c) Output from the CDSSTR algorithm (http://public-1.cryst.bbk.ac.uk/cdweb/html/) (47), showing the calculated proportion of secondary structural motifs for each peptide. Peptide names in bold type had 50 µg ml^–1 pRS313 added.

DISCUSSION

Top

DISCUSSION

Despite having no sequence identity, DsS3(1-16) and Mag 2 havebeen shown to adopt amphipathic ${alpha}$ -helices in hydrophobic media,and thus, it has been proposed that their principal inhibitoryaction is due to membrane disruption (31, 52). Less certainis whether the membrane disruption induced by this interactionis the mechanism that kills the target microorganism.

Here, we have presented evidence implying that telomere maintenance,chromosome organization, and double-strand break repair arevital functions for maintaining resistance to the inhibitoryactions of both peptides. Furthermore, ubiquitin-dependent proteincatabolism, perhaps to remove peptides from inside the cells,was also shown to be important. Thus, Mag 2 and DsS3(1-16) sharecommon inhibitory actions, one of which appears to involve interferencewith the integrity of DNA. Interaction of other basic, cationicpeptides with negatively charged DNA has been proposed previously;for example, buforin 2, from the Asian toad Bufo bufo gargarizans,rapidly enters cells without perturbing the integrity of thecytoplasmic membrane and subsequently binds to DNA and RNA (40).

Study of factors unique to each peptide gave more subtle insightinto the unique inhibitory actions of Mag 2 and DsS3(1-16).With DsS3(1-16), genes involved in vacuolar transport, acidification,and protein targeting to the vacuole were vital in conferringresistance to this peptide. The vacuole plays a key role inthe sequestration and degradation of proteins delivered fromthe plasma membrane via endocytosis (23), and this could representan effective means of clearing excessive toxic peptide frominside the cell. DsS3(1-16) also induced genes mediating a responseto stress, reactive oxygen species, and drug treatment, implyingtoxic intracellular effects of the peptide and induction ofABC transporters, perhaps to remove excess peptide from thecell. In contrast, Mag 2 specifically induced genes involvedin both Ca²⁺- and cyclic AMP-mediated cell signaling and regulationof cell wall organization and biogenesis. The cell wall is animportant protective barrier to the yeast cell that the peptidehas to cross to reach the plasma membrane. We have preliminarydata showing that differences in the cell wall structure mediatehuge differences in the susceptibility of yeast to cationicpeptides (unpublished data). However, the most notable uniqueeffect of Mag 2 compared to DsS3(1-16) was the up-regulationof many genes involved in DNA-dependent DNA replication andDNA repair, implying that this peptide has a more significantdetrimental effect on the integrity of DNA than DsS3(1-16) does.

The deletion screen and expression analysis indicated that acommon inhibitory mechanism of both peptides could be theirability to interact with DNA. This was supported by in vitrogel retardation, DNase protection, and strand break assays thatindicated that both peptides bound DNA, that Mag 2 appearedto bind to DNA with greater affinity than DsS3(1-16), and, finally,that Mag 2, but not DsS3(1-16), could induce strand breaks invitro. Further differences in the nature of the peptides' interactionswith DNA were highlighted by CD analysis. These experimentsshowed that DsS3(1-16) was largely disordered in solution, whichwas supported by previous observations (15), and this did notchange in the presence of DNA. Mag 2, however, was disorderedin solution but in the presence of DNA adopted an ${alpha}$ -helical conformationsimilar to that observed in a hydrophobic environment. Thus,despite interaction with DNA being a possible common mechanismof inhibition of both peptides, each peptide appears to interactwith DNA differently in vitro. Crucially, this could accountfor some of the differences observed with the deletion mutantscreen and gene expression profiles obtained from cells exposedto each of the inhibitory peptides. For example, the specificup-regulation of genes involved in DNA-dependent DNA replicationand DNA repair that occurred upon exposure to Mag 2 furthersupported our in vitro observations that this peptide had amore significant detrimental effect on the integrity of DNAthan DsS3(1-16) did.

Collectively, our data imply that both peptides must be ableto pass though the cytoplasmic membrane of yeast cells to mediateinhibitory effects inside the cell. Supporting this conclusion,DsS3 has been shown to affect the viability of intraerythrocyticparasites at concentrations that did not permeate the host cell,indicating that the peptide can cross lipid bilayers and entercells via some unknown mechanism (15). Similarly, Mag 2 hasbeen shown to enter the cytosol of microbial cells (17). Therefore,it is clear that the two peptides have inhibitory actions notnecessarily linked to membrane disruption. Indeed, we have shownrecently that a principal antifungal action of DsS3(1-16), butnot of Mag 2, is the induction of programmed cell death (34).

In summary, we have shown that DsS3(1-16) and Mag 2 have bothcommon and unique inhibitory actions on yeast cells that arenot simply due to membrane disruption. In fact, both DsS3(1-16)and Mag 2 interact with DNA in vitro in subtly different ways,but more importantly, functional genomics experiments have providedevidence that both peptides also interact and interfere withDNA integrity differently in vivo. This implies that both peptidesare able to pass through the cytoplasmic membrane and interferewith the function of DNA in vivo, an inhibitory action thathas not previously been observed for either of these peptides.

ACKNOWLEDGMENTS

We thank J. Potter for expert assistance with CD.

The microarray work was supported by the Consortium for theFunctional Genomics of Microbial Eukaryotes, which is fundedby the UK Biotechnology and Biological Sciences Research Council(34/IGF13036). P.C. and C.O.M. also thank the BBSRC for supportingthis research (grant no. P20473).

FOOTNOTES

* Corresponding author. Mailing address: Centre for Biomolecular Sciences, School of Biology, University of St. Andrews, The North Haugh, St. Andrews KY16 9ST, United Kingdom. Phone: (44) (0)1334 463406. Fax: (44) (0)1334 462595. E-mail: pjc5{at}st-andrews.ac.uk

Published ahead of print on 10 September 2007.

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

Present address: Centre for Plant Integrative Biology, Schoolof Biosciences, University of Nottingham, Sutton Bonington LE125RD, United Kingdom.

REFERENCES

Top