Published Ahead of Print 24 November 2010. 2011, 85(4):1507. DOI: 10.1128/JVI.01876-10. J. Virol. 

S. L. ParkerJae-Won Kim, Sangbom M. Lyi, Colin R. Parrish and John Membrane-Destabilizing Activity

1 HasµReovirus Outer Capsid Protein A Proapoptotic Peptide Derived from

http://jvi.asm.org/content/85/4/1507Updated information and services can be found at:

These include:

REFERENCEShttp://jvi.asm.org/content/85/4/1507#ref-list-1at:

This article cites 34 articles, 18 of which can be accessed free

CONTENT ALERTS more»articles cite this article),

Receive: RSS Feeds, eTOCs, free email alerts (when new

http://journals.asm.org/site/misc/reprints.xhtmlInformation about commercial reprint orders: http://journals.asm.org/site/subscriptions/To subscribe to to another ASM Journal go to:

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

on July 9, 2014 by C

ornell University Library

http://jvi.asm.org/

Dow

nloaded from

JOURNAL OF VIROLOGY, Feb. 2011, p. 1507–1516 Vol. 85, No. 40022-538X/11/$12.00 doi:10.1128/JVI.01876-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

A Proapoptotic Peptide Derived from Reovirus Outer Capsid Protein�1 Has Membrane-Destabilizing Activity�

Jae-Won Kim, Sangbom M. Lyi, Colin R. Parrish, and John S. L. Parker*Baker Institute for Animal Health, College of Veterinary Medicine, Cornell University, Ithaca, New York 14853

Received 3 September 2010/Accepted 10 November 2010

The reovirus outer capsid protein �1 is responsible for cell membrane penetration during virus entry andcontains determinants necessary for virus-induced apoptosis. Residues 582 to 611 of �1 are necessary andsufficient for reovirus-induced apoptosis, and residues 594 and 595 independently regulate the efficiency ofviral entry and reovirus-induced cell apoptosis, respectively. Two of three �-helices within this region, helix 1(residues 582 to 611) and helix 3 (residues 644 to 675), play a role in reovirus-induced apoptosis. Here, wechemically synthesized peptides representing helix 1 (H1), H1:K594D, H1:I595K, and helix 3 (H3) andexamined their biological properties. We found that H1, but not H3, was able to cause concentration- andsize-dependent leakage of molecules from small unilamellar liposomes. We further found that direct applica-tion of H1, but not H1:K594D, H1:I595K, or H3, to cells resulted in cytotoxicity. Application of the H1 peptideto L929 cells caused rapid elevations in intracellular calcium concentration that were independent of phos-pholipase C activation. Cytotoxicity of H1 was not restricted to eukaryotic cells, as the H1 peptide also hadbactericidal activity. Based on these findings, we propose that the proapoptotic function of the H1 region of �1is dependent on its capacity to destabilize cellular membranes and cause release of molecules from intracel-lular organelles that ultimately induces cell necrosis or apoptosis, depending on the dose.

Mammalian orthoreoviruses cause pathology in newbornmice by inducing apoptosis in infected cells (2). In addition,type 3 Dearing mammalian orthoreoviruses (Reolysin) are cur-rently in phase II/III trials for treatment of a variety of cancers(17, 28). Tumor killing by reovirus infection involves effectsthat are both direct and indirect (e.g., activation of a tumor-specific immune response) (25). However, infected tumor cellsundergo virus-induced apoptosis (26). Although much is nowknown about the specific apoptotic and signaling pathways thatare activated in reovirus-infected cells, the underlying mecha-nism of how the virus initiates apoptosis is not well understood.The �1 protein is the major determinant of reovirus-inducedapoptosis, and ectopic expression of �1 in cells induces apop-tosis (3, 5, 6). Apoptosis induced by ectopically expressed �1 isabrogated by coexpression of outer capsid protein �3, withwhich �1 assembles to form heterohexamers that form themajor portion of the outer capsid of the virus (3). Residues 582to 611 of �1 are necessary and sufficient for apoptosis induc-tion, and when fused to enhanced green fluorescent protein(EGFP), this fragment of �1 predominantly localizes to mito-chondrial membranes (3). In addition, recombinant T3D vi-ruses with substitutions in this region of �1 (K594D or I595K)are defective for induction of apoptosis and have attenuatedvirulence in newborn mice (5). How this particular region of�1 functions to initiate apoptosis is unknown. In the atomicresolution structure of the �1:�3 heterohexamer, this region of�1 forms a short loop followed by an amphipathic �-helix;residues 584 and 585 lie within the amphipathic �-helix (19).Amphipathic peptides are known to target and sometimes de-

stabilize membranes, and this region of �1 was previouslysuggested to be involved in membrane penetration during virusentry (20). Based on these previous findings, we hypothesizedthat one mechanism by which �1 induces apoptosis is by directdestabilization of intracellular membranes. Such destabiliza-tion could induce a cellular stress response or lead to directleakage of proapoptotic molecules from mitochondria or theendoplasmic reticulum (ER).

Here, we examined the capacity of peptides representingresidues 582 to 611 of �1 to perturb artificial and cellularmembranes.

MATERIALS AND METHODS

Cells. L929 cells were maintained in suspension in Joklik’s modified minimalessential medium supplemented with 4% fetal bovine serum, 2 mM glutamine,100 U ml�1 penicillin, and 100 �g ml�1 streptomycin. CV-1 cells were grown inHam’s F-12 medium (CellGro) supplemented with 10% fetal bovine serum(HyClone), 100 U ml�1 penicillin, 100 �g ml�1 streptomycin, 1 mM sodiumpyruvate, and nonessential amino acids (CellGro). CHO-K1 cells were grown inDulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetalbovine serum, 100 U ml�1 penicillin, and 100 �g ml�1 streptomycin.

Peptide synthesis and CD. Peptides were synthesized by FlexPeptide technol-ogy, purified by high-performance liquid chromatography (HPLC) and con-firmed by mass spectrometry at GenScript (Piscataway, NJ). Circular dichroism(CD) was used to analyze the secondary structure of the synthetic peptides.Initially, the peptides were analyzed at 50 �M in trifluorethanol or in thepresence of 10% phosphatidylcholine (PC) liposomes (prepared as describedbelow). Spectra were recorded with a model 400 AVIV CD spectrometer (Lake-wood, NJ) in a 1-cm-path-length cuvette at room temperature (RT). Structuralanalysis was performed using the Delta epsilon (DE) calculation method usingK2D CD spectra Deconvolution software (http://www.embl.de/�andrade/k2d.html).

Cytotoxicity assay. We used a fluorescence-based cytotoxicity assay (CytoTox-Fluor; Promega, Madison, WI) to quantify cytotoxicity induced by incubationwith the synthetic peptides. L929 cells were seeded in 96-well plates and incu-bated with the indicated concentrations of each peptide for 1 h at 37°C in 100 �lof DMEM, and then the levels of cytotoxicity were measured following incuba-tion with the fluorescent substrate (100 �l) for an additional 30 min at 37°Caccording to the manufacturer’s instructions. Fluorescence was measured by

* Corresponding author. Mailing address: Baker Institute for Ani-mal Health, College of Veterinary Medicine, Cornell University, Hun-gerford Hill Road, Ithaca, NY 14853. Phone: (607) 256-5626. Fax:(607) 256-5608. E-mail: jsp7@cornell.edu.

� Published ahead of print on 24 November 2010.

1507

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

spectrofluorimetry (excitation wavelength [�ex] � 485 nm; emission wavelength[�em] � 520 nm), and cytotoxicity was calculated by the following equation:cytotoxicity � (mean for treated cells � mean for untreated cells)/standarddeviation (SD) for cell-free controls. Maximum cytotoxicity (100%) was set asthe fluorescence measured in cells treated with 30 �g ml�1 digitonin for 1 h at37°C.

Hemolysis assay. Hemolysis assays were performed as described previously(34). In brief, washed citrated chicken red blood cells (RBCs) in Alsevers solu-tion (Colorado Serum Co., Denver, CO) were suspended in cold phosphate-buffered saline (PBS) containing 2 mM MgCl2 (PBS-Mg) at a concentration of5% (vol/vol) with different concentrations of the indicated peptides. After incu-bation for 30 min at 37°C, the RBCs were pelleted by centrifugation (380 � g for5 min) and released hemoglobin in the supernatant was measured by absorbanceat 415 nm. Supernatant collected from RBCs incubated with 0.1% Triton X-100was used to assess complete release of hemoglobin, and RBCs incubated withPBS were used as a negative control. Data shown are the means SD of resultsfrom a minimum of three independent experiments.

Preparation of liposomes and liposome leakage assays. Phosphatidylcholine(PC) in chloroform (18:1; Avanti Lipids) was dried under nitrogen, rehydrated at1 mg ml�1 in PBS, pH 7.4, containing 12.5 mM aminonaphthalene trisulfonicacid (ANTS; Fluka) and 45 mM p-xylene-bis (N-pyridinium bromide) (DPX;Sigma), and then frozen and thawed (10�) in liquid nitrogen. Unilamellarliposomes (50-nm diameter) were prepared by extrusion (Lipex extruder; North-ern lipids, Inc.). Similar PC liposomes (200-nm diameter) were prepared in PBScontaining 300 mg ml�1 fluorescein isothiocyanate (FITC)-conjugated dextran(FD) with molecular weights of 40,000 (FD40) or 10,000 (FD10) (Sigma). Un-incorporated ANTS-DPX or FD was separated from liposomes by mixing theliposomes 1:1 (vol/vol) in 70% sucrose, overlaying them with 20% sucrose, andthen centrifuging them at 96,800 � g for 15 h at 4°C. Liposomes were harvestedfrom the top of the gradient and then reequilibrated in PBS by gel filtration ona Sephadex G-25 PD10 column (Amersham Biosciences). Dequenching ofANTS in response to liposome leakage was assayed by spectrofluorimetry (CaryEclipse spectrophotometer; Varian, Inc.) (�ex � 355 nm; �em � 520 nm). Thepercent leakage was calculated using the baseline fluorescence emission ofANTS-DPX-loaded liposomes in buffer only (F0), the sample fluorescence (FS),and the 100% leakage fluorescence (FT) (liposomes were lysed with 0.2% TritonX-100) as follows: percent leakage � [(FS � F0)/(FT � F0)] � 100. Leakage ofFD40 and FD10 from liposomes was also assayed by dequenching (�ex � 490 nm;�em � 500 to 540 nm). Maximum leakage of FD was assayed after addition of0.2% Triton X-100; in control experiments, we found that 0.2% Triton X-100 wasrequired to fully lyse liposomes.

We prepared liposomes with the lipid composition of Xenopus mitochondriausing egg phosphatidylcholine (PC; 47 mol%), egg phosphatidylethanolamine(PE; 28 mol%), soybean phosphatidylinositol (PI; 9 mol%), and brain phospha-tidylserine (PS; 9 mol%) mixed with 7 mol% of heart cardiolipin (CL) (AvantiPolar Lipids) as described in the text. For lower concentrations of cardiolipin, themol% of PC was increased. The lipid mixture was dried and resuspended in PBScontaining ANTS-DPX as described above. Liposome extrusion was as describedabove, with final extrusion through a 50-nm filter membrane.

Measurement of intracellular free Ca2� concentration. We assessed changesin intracellular free Ca2 levels in L929 cells. Spinner culture L929 cells (ap-proximately 3 � 107 cells) were harvested, washed in PBS buffer, incubated for1 h at 37°C in the dark in 6 ml of serum-free Joklik’s modified minimal essentialmedium containing 3 �M fura-2 acetoxymethyl ester (fura-2AM; Invitrogen) and8 �M pluronic F-127 (Invitrogen), then washed twice with Hanks balanced saltsolution (HBSS) buffer (130 mM NaCl, 5 mM KCl, 1.8 mM CaCl2, 1.0 mMMgCl2, 0.6 mM Na2HPO4, 0.4 mM KH2PO4, 5 mM glucose, 20 mM HEPES, pH7.4), resuspended in HBSS containing 2.5 mM probenecid (Sigma), and furtherincubated for 20 min at RT in the dark to allow esterification of fura-2AM. Follow-ing fura-2 loading, the cells were washed three times and resuspended in 3 ml ofHBSS containing 1 mM EGTA immediately prior to measurements. Cell suspen-sions (400 �l) were placed in a quartz cuvette with a magnetic stirrer. Experimentalreagents, including peptides, Triton X-100, carbachol, or phospholipase C (PLC)inhibitor (U-73122; Sigma), were added at the time points indicated in Results.Fluorescence was measured by spectrofluorimetry (�ex � 340 and 380 nm; �em �510 nm). The ratio of the fura-2AM–Ca2 complex (�ex � 340 nm; �em � 510 nm)to free fura-2AM (�ex � 380 nm; �em � 510 nm) was generated using the Grynk-iewicz equation to calculate internal calcium concentration (10).

Microscopy and microinjection. Biotin-conjugated H1 and H1:I595K peptideswere synthesized, dissolved in PBS, and incubated for 30 min at RT with CV-1cells on 18-mm glass coverslips. The cells were washed extensively in PBS, thenfixed with 2% paraformaldehyde in PBS for 10 min, incubated with Alexa488Fluor-conjugated streptavidin (Invitrogen), and washed again in PBS, and then

the coverslips were mounted on glass slides with Prolong Gold plus DAPI(4�,6-diamidino-2-phenylindole; Invitrogen). Images were obtained with an in-verted microscope (Nikon TE2000) equipped with fluorescence optics through a60� 1.4-numerical-aperture (NA) oil objective with 1.5� optical zoom. Imageswere collected digitally with a Coolsnap HQ charge-coupled-device camera(Roper) and Openlab software (Improvision) and then prepared for publicationusing Photoshop and Illustrator software (Adobe).

To assess the distribution of biotinylated H1 in cells, CHO-K1 cells seeded on50-mm MatTek gridded dishes (MatTek, Ashland, MA) were preincubated with500 nM Mitotracker CMXRos (Invitrogen) in growth medium for 45 min at37°C. The labeling medium was replaced with medium lacking phenol red, andthe cells were microinjected with �6 nl of sample using an InjectMan NI2microinjector and a Femtoject micromanipulator (Eppendorf, Hauppauga, NY).Micropipette needles were prepared with a P-97 micropipette puller (SutterInstrument Co., Novato, CA). Samples injected were 50 �M biotin-conjugatedH1 that was preincubated with Alexa488-conjugated streptavidin or, as a control,Alexa488-conjugated streptavidin. Samples were prepared in PBS buffer andinjected into cells, and the cells were incubated at 37°C on a stage warmer(World Precision Instruments, Sarasota, FL). Fluorescence and phase-contrastimages were collected using a Nikon TE300 microscope and a Hamamatsu OrcaER charge-coupled-device camera. Images were collected using Simple PCIsoftware (Hamamatsu, Sewickley, PA) and contrast and brightness adjusted withPhotoshop (Adobe, San Jose, CA).

To assess the release of cytochrome c from cells microinjected with the H1,H1:I595K, or H3 peptide, CHO-K1 cells maintained at 37°C were microinjectedas described above with a 1:1 mixture of Alexa594 bovine serum albumin (BSA)(Invitrogen) and the indicated peptides (final concentration, 50 �M). Twentyminutes after microinjection, the cells were fixed in 4% paraformaldehyde inPBS, washed in PBS, permeabilized with 0.1% Triton X-100 in PBS plus 1%BSA, and then immunostained with anti-cytochrome c monoclonal antibody(MAb) (clone 6H2.B4; BD Biosciences) followed by Alexa594-conjugated goatanti-mouse IgG. Fluorescence and phase-contrast images were collected andprocessed as described above.

Peptide bactericidal activity. The bactericidal activity of peptides against Esch-erichia coli and Bacillus subtilis JH641 was assayed by colony inhibition. Bacterialcultures were grown to early log phase at 37°C in Luria-Bertani (LB) broth andthen diluted to 2 � 104 CFU/ml in 10 mM phosphate buffer (pH 7.4) containing0.03% LB broth. Equal volumes of peptide-containing solution and bacterialculture (25 �l of each) were mixed and incubated at 37°C for 3 h, after which thebacterium-peptide mixture was diluted 100-fold or 1,000-fold with cold phos-phate buffer and inoculated on LB agar plates, which were incubated at 37°Covernight. The number of bacterial colonies on each plate was counted thefollowing day and expressed as a percentage of the number of colonies obtainedwhen phosphate buffer was mixed with the bacteria instead of peptide.

Preparation of isolated mitochondria and cytochrome c release assay. Cyto-chrome c released from mitochondria purified from murine L929 cells wasassayed as described previously, with some modification (32). Cells in spinnerculture were harvested, washed with cold PBS, and then resuspended at 3 � 107

ml�1 in ice-cold mitochondrial buffer (20 mM HEPES, pH 7.5, 250 mM sucrose,1 mM EGTA, 1 mM EDTA, 100 mM KCl, 1 mM dithiothreitol [DTT]) contain-ing 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 2 �g ml�1 pepstatin, andcomplete protease inhibitor (Complete Mini; Roche). Cells were lysed byDounce homogenization (80 strokes), and nuclei and intact cells were removedby centrifugation (800 � g for 15 min at 4°C). The supernatant was then centri-fuged at 10,000 � g for 15 min at 4°C. The mitochondrial pellet was furtherwashed and resuspended in mitochondrial buffer before being used immediatelyfor further experiments. Both the mitochondrion-free supernatant and the mi-tochondrial pellet were saved for analysis. Isolated mitochondria were incubatedwith 100 �M the H1 or H3 peptide, treated with 1% Triton X-100, or incubatedwith an equal volume of PBS for 20 min at RT. After incubation, the sampleswere returned to ice and the mitochondrial and supernatant fractions wereseparated by centrifugation (16,000 � g for 15 min at 4°C). The pellet fractionwas resuspended in a volume of mitochondrial buffer equal to that of thesupernatant. Samples were then incubated in protein sample buffer with 200 mMDTT on ice for 15 min and then at 80°C for 15 min, cooled on ice, and clearedat 16,000 � g for 1 min, and then proteins were separated on 15% SDS-PAGE.Proteins were transferred to nitrocellulose membranes by Semi-Dry TransferCell (Bio-Rad). The membrane was blocked with 5% BSA in Tris-bufferedsaline–Tween 20 (TBST) buffer, then incubated with anti-cytochrome c MAb(1:1,000; 7H8.2C12; BD Biosciences) overnight at 4°C, washed, and then incu-bated with horseradish peroxidase (HRP)-conjugated donkey anti-mouse IgG(1:10,000; Jackson Laboratories). Proteins were detected on autoradiograph filmwith SuperSignal West Pico chemiluminescent substrate (Thermo Scientific).

1508 KIM ET AL. J. VIROL.

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

Statistical analysis. All experiments were performed a minimum of threetimes, and data are shown as mean standard deviation (SD). Preparation ofgraphs and statistical analysis was done using Prism GraphPad software.

RESULTS

A synthetic peptide representing residues 582 to 612 of �1 is�-helical in solution. We have previously shown that the C-terminal � fragment (residues 582 to 708) of the reovirus outercapsid �1 protein contains determinants necessary for virus-induced apoptosis. Residues 582 to 611 of �1 are necessaryand sufficient for apoptosis induction (3), and substitutionsK594D and I595K within this region in recombinant viruseshave reduced apoptosis-inducing capacity in vitro and in vivo(5). In the structure of the �1:�3 heterohexamer, residues 582to 611 consist of a short loop and an amphipathic �-helix (Fig.1A) (19).

Many amphipathic peptides interact with membranes and insome cases destabilize membrane barriers (18). We previouslyshowed that residues 582 to 611 fused to EGFP localize tomitochondria, lipid droplets, and the endoplasmic reticulum intransfected cells (3). Destabilization of mitochondrial and/orER membranes could in part explain the mechanism by which�1 induces apoptosis. We therefore examined whether a syn-thetic peptide representing residues 582 to 611 could destabi-lize cellular and artificial membranes. Peptides representingresidues 582 to 612 of �1 (peptide H1) and, as a negativecontrol, a peptide representing residues 639 to 675 of �1 (pep-tide H3), which does not induce apoptosis when expressed incells as a fusion with EGFP (3), were chemically synthesized.All of the peptides were soluble in aqueous solutions. We usedcircular dichroism to identify the secondary structures of thepeptides in trifluorethanol (H1 and H3), which is known tostabilize peptide structures. We found that the H1 and H3peptides were �-helical in solution (100%) (Fig. 1B). The H1and mutant peptides (K593D and I595K) had identical spectra,while H3 had deeper and higher peaks, most likely because ofthe greater length of H3 (Fig. 1B). We observed no significantchanges in the secondary structure conformers of the H1 pep-tide in the presence of liposomes (data not shown).

Peptide H1 (residues 582 to 611) directly permeabilizes li-posomes. To examine if the H1 peptide could destabilize mem-branes, we prepared small unilamellar liposomes (�50-nm di-ameter) consisting of the neutral lipid phosphatidylcholine thatcontained the fluorophore/quencher pair 8-aminonaphthalene1,3,6-trisulfonic acid (ANTS)–p-xylene-bis-pyridinium (DPX).We then assessed liposome leakage in response to increasingconcentrations of H1 by measuring dequenching of ANTS (1).We found that H1 induced a concentration-dependent leakageof ANTS from liposomes, with 50% leakage occurring at aconcentration of �12.6 �M. In contrast, the H3 peptide had noeffect on liposome membrane permeability (Fig. 2A). We con-clude that the H1 peptide has direct membrane-permeabilizingproperties.

Mutants of H1 lose the capacity to directly permeabilizeliposomes. We previously showed that recombinant viruses or�1 with the I595K or K594D substitution within the H1 helixwere defective in apoptosis induction (5). We therefore exam-ined the capacity of peptides bearing these substitutions topermeabilize PC liposomes. We found that unlike the wild-

type H1 peptide, H1:I595K did not induce liposome leakage.The H1:K594D peptide did induce leakage, but �2-fold-higherconcentrations than those of wild-type H1 were required toinduce 50% leakage (Fig. 2A). Thus, the capacity of �1, �, ortheir I595K and K594D substitution mutants to induce apop-tosis when expressed ectopically in cells correlates with thecapacities of peptides representing the region that is minimallyrequired for apoptosis induction to permeabilize artificialmembranes.

Synthetic H1 peptide causes size-selective leakage of FITC-dextrans from liposomes. The �1N fragment of �1 (Fig. 1A) isrequired for virus entry but is not involved in �1-mediated

FIG. 1. Linear representation and ribbon diagram of a monomer of�1 with sequence and CD spectra of the synthesized peptides. (A) Thedifferent fragments of �1 generated during virus entry are illustrated ina linear representation. The �-helical regions within the � fragment of�1 are colored red (H1), yellow, and blue (H3). The amino acidsequences of the H1, H1:K594D, H1:I595K, and H3 peptides areshown. Numbers indicate amino acid positions within �1. (B) Circulardichroism spectra of the synthetic peptides H1, H1:K594D, H1:I595K,and H3. All of the peptides showed minima at 208 and 222 nm,characteristic of alpha-helical secondary structure. Peptides (50 �M)were solubilized in trifluoroethanol to prevent aggregation and tostabilize peptide structure. Spectra were recorded at room tempera-ture with a model 400 AVIV CD spectrometer (Lakewood, NJ).

VOL. 85, 2011 REOVIRUS �1 AND PEPTIDE MEMBRANE PERMEABILIZATION 1509

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

apoptosis induction. Zhang et al. showed that a synthetic myr-istoylated �1N peptide likely forms size-selective pores in ar-tificial membranes (34). To address the possibility that the H1peptide also forms size-selective pores, we prepared liposomespreloaded with self-quenching concentrations of 10- or 40-kDafluorescein isothiocyanate (FITC)-dextran polymers. Leakageof the FITC-dextran molecules was assayed by fluorescencedequenching (Fig. 2B). We found that concentrations of H1that caused release of ANTS also caused release of 10 kDaFITC-dextran; however, the release of 40 kDa FITC-dextranwas much less efficient (Fig. 2B). These findings indicate thatsize-limited pores formed in liposomes treated with the H1peptide.

The H1 peptide permeabilizes cell membranes. (i) H1 iscytotoxic for L929 cells. Based on our finding that the H1

peptide directly permeabilized artificial membranes, we inves-tigated whether incubation of H1 with cells was cytotoxic. Us-ing an assay that quantifies release of protease activity fromdead cells (CytoTox-Fluor; Promega), we found that concen-trations of H1 of 50 �M or greater were cytotoxic to L929 cellsafter 1 h of incubation (Fig. 3A). In contrast, similar concen-trations of the H3 peptide had no effect on the viability of L929cells (Fig. 3A). The I595K and K594D mutant H1 peptideswere not cytotoxic, even at 200 �M (Fig. 3A).

(ii) H1 causes hemolysis of red blood cells. To further de-termine if the H1 peptide was directly permeabilizing theplasma membranes of cells, we assayed for leakage of hemo-globin from chicken red blood cells. Unlike for the �1N pep-tide, where concentrations of 10 �g ml�1 led to 70% hemolysis,we found that only 30% hemolysis occurred when chickenRBCs were incubated with 20-fold-higher concentrations ofH1 peptide (Fig. 3B). This difference in efficiency of hemoglo-bin release (�68 kDa) likely reflects a lower size exclusionlimit of pores formed by the H1 peptide than of those formedby �1N, which allow release of hemoglobin and 40-kDa FITC-dextran (34).

(iii) H1 causes changes in cytosolic Ca2� levels. We hypoth-esized that the cytotoxicity that we observed in L929 cells wascaused by increases in cytosolic Ca2 concentrations. In sup-port of this hypothesis, we found that application of the H1peptide to fura-2-loaded L929 cells induced substantial andrapid increases in cytosolic [Ca2], with 50 �M H1 causing a 3-to 4-fold increase in cytosolic [Ca2] by 2 min after addition ofthe peptide. In contrast, application of 50 �M H3 had nonoticeable effect on cytosolic [Ca2] (Fig. 3C). As these exper-iments were carried out with Ca2-free buffer containingEGTA, this finding suggested that in addition to having directmembrane permeabilization effects, the H1 peptide might alsostimulate phospholipase C-mediated IP3 signaling pathwayswith release of Ca2 from intracellular stores. Addition ofpurified rotavirus NSP4 or a peptide representing NSP4114-135

induces elevation of cytosolic calcium in Sf9 insect cells, whichis blocked by the phospholipase C inhibitor (U-73122) (29).Treatment with U-73122 prevented the increase in Ca2 in-duced by 100 �M carbachol, a PLC receptor agonist (Fig. 3D).However, we found that pretreatment of cells with the PLCinhibitor (U-73122) had no effect on H1-induced increases inintracellular Ca2 (Fig. 3D). Based on these findings, we con-clude that the rapid changes in intracellular Ca2 cannot beattributed to direct permeabilization of the plasma membranealone. We speculate that the H1 peptide may access the intra-cellular membranes of the ER or mitochondria, causing therelease of calcium stores.

H1 peptide has antibacterial activity for Gram-negative andGram-positive bacteria. Based on the cytotoxicity of H1 forL929 cells, the membrane-permeabilizing activity of H1, andthe propensity of cationic amphipathic helical peptides to haveantibacterial activity (33), we tested whether the H1 peptidehad antibacterial activity. We found dose-dependent inhibitionof the growth of Gram-positive (B. subtilis) and Gram-negative(E. coli) bacteria by the H1 peptide (Fig. 4A), with completeinhibition of colony formation in this assay at concentrations ofH1 greater than 25 �M. In contrast, 50 �M the H3 peptide hadlittle effect (Fig. 4B), and 50 �M concentrations of the K594Dand I595K substitution mutants had decreased antibacterial

FIG. 2. Synthetic H1 but not mutant or H3 peptides cause releaseof small molecules from phosphatidylcholine liposomes. (A) ANTS-DPX liposome leakage assays of H1, H3, and mutant peptides. Pep-tides at the indicated concentrations were incubated with phospha-tidylcholine (PC)-containing liposomes loaded with ANTS and DPX,and leakage of ANTS was assayed by spectrofluorimetry (�ex � 355nm; �em � 520 nm). Means SD of results from three independentexperiments are shown. The percent leakage was calculated using thebaseline fluorescence emission of ANTS-DPX-loaded liposomes inbuffer only (F0), the sample fluorescence (FS), and the 100% leakagefluorescence (FT) (liposomes lysed with 0.2% Triton X-100) as follows:percent leakage � [(FS � F0)/(FT � F0)] � 100. (B) Phosphatidylcho-line (PC) liposomes containing self-quenching concentrations (300 mgml�1) of 10 or 40 kDa fluorescein isothiocyanate-conjugated dextran(FD10 or FD40) were incubated with the indicated concentrations ofH1 peptide. Dequenching of FD was assayed by spectrofluorimetry(�ex � 490 nm; �em � 520 nm). The percent leakage was calculatedusing the baseline fluorescence emission of FD-loaded liposomes inbuffer only (F0), the sample fluorescence (FS), and the 100% leakagefluorescence (FT) (liposomes lysed with 0.2% Triton X-100) as follows:percent leakage � [(FS � F0)/(FT � F0)] � 100. Means SD of resultsfrom three independent experiments are shown.

1510 KIM ET AL. J. VIROL.

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

activity compared to the level for WT H1 (Fig. 4B). We con-clude that the H1 peptide has antibacterial properties.

Effect of lipid composition on H1-induced ANTS leakage.Green fluorescent protein (GFP)-fused H1 localizes promi-nently to mitochondria in transfected cells, causes release of

cytochrome c, and induces apoptosis (3, 31). Mitochondrialmembranes have increased concentrations of the neutral lipidcardiolipin, and permeabilization of liposomes by the Bcl-2homology domain 3 (BH3)-only peptides derived from pro-apoptotic proteins Bax and Bak is promoted by increased con-

FIG. 3. The H1 peptide is cytotoxic and permeabilizes cellular membranes. (A) Different concentrations of H1, H1:K594D, H1:I595K, and H3were incubated with L929 cells for 1 h at 37°C, and then the percentage cytotoxicity (maximum cytotoxicity was set as the fluorescence measuredin cells treated with 30 �g ml�1 digitonin for 1 h at 37°C) was assayed with Live/Dead fixable cell stain (Invitrogen) using a fluorescence platereader. Means SD of results from 3 independent experiments are shown. (B) The indicated concentrations of peptides were incubated withchicken RBCs at 37°C for 1 h, and hemolysis was assayed as described in Materials and Methods. (C) The indicated concentrations of H1 or H3peptides were added (arrow) to L929 cells preloaded with 3 �M fura-2 acetoxymethyl ester (fura-2AM) in Ca2-free Hanks buffered salt solution,pH 7.4, in a stirred cuvette at 37°C in a spectrofluorimeter. Fluorescence measurements were collected before and after addition of peptides (�ex �340 and 380 nm; �em � 510 nm). Change in cytosolic [Ca2] (nM) was determined from emission data using the Grynkiewicz equation (10).(D) The cytosolic calcium elevation in L929 induced by H1 is not mediated by a phospholipase C pathway. Fura-2AM-loaded L929 cells wereincubated with H1 (50 �M) or the M5 muscarinic agonist carbachol (100 �M) (left arrow; solid traces), and changes in cytosolic calcium levelswere monitored by spectrofluorimetry. Alternatively, H1 or carbachol was added to cells, together with 10 �M U-73122 (right arrow; dotted traces).The traces shown are representative of three independent experiments.

VOL. 85, 2011 REOVIRUS �1 AND PEPTIDE MEMBRANE PERMEABILIZATION 1511

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

centrations of cardiolipin (16). We therefore prepared lipo-somes with lipid contents that approximated those of Xenopusmitochondria and examined the effect of the presence or ab-sence of 7 mol% cardiolipin on the efficiency of H1 membranepermeabilization. We found that the efficiency of H1 perme-abilization of liposomes was increased when liposomes con-tained cardiolipin (Fig. 5A). Based on these findings, we testedwhether H1 could directly permeabilize the outer mitochon-drial membranes of isolated mitochondria and cause release ofcytochrome c, which is normally retained in the mitochondrialintermembrane space. We isolated mitochondria from L929cells and then incubated them with different concentrations ofthe H1, H3, H1:K594D, and H1:I595K peptides. To assesspermeabilization of the outer mitochondrial membrane, weexamined for leakage of cytochrome c from the intermem-brane space. Although the H1 peptide localized to mitochon-

FIG. 4. The peptide H1 has a bactericidal activity on Gram-positiveand Gram-negative bacteria. (A) The capacity of different concentra-tions of the H1 peptide to inhibit bacterial growth was assayed asdescribed in Materials and Methods. (B) The indicated peptides weretested for their bactericidal activity at 50 �M. Results shown are themeans SD for three independent experiments.

FIG. 5. H1 preferentially permeabilizes liposomes containing the mitochondrial lipid cardiolipin (A) but does not cause release of cytochromec from isolated mitochondria (B) or mitochondria in CHO-K1 cells microinjected with H1 (C). (A) ANTS-DPX-loaded liposomes containing amixture of lipids characteristic of mitochondria from Xenopus oocytes were prepared with or without 7 mol% cardiolipin. Leakage of ANTS wasquantified as before upon incubation with the indicated H1 concentrations. Means SD of results from three independent experiments are shown.(B) Mitochondria isolated from L929 cells were incubated with 100 �M the H1 or H3 peptide or with equal volumes of PBS or 1% Triton X-100for 20 min at RT. After incubation, mitochondria were pelleted, and cytochrome c present in the pellet (P) and supernatant (S) fractions wasassayed by immunoblot analysis. A sample of purified cytochrome c (marked C) was loaded in the first lane. (C) CHO-K1 cells were microinjectedwith a 1:1 mixture of Alexa594 BSA, and the H1, H1:I595K, or H3 peptide (final concentration, 50 �M). Cells were incubated for a further 20 minat 37°C and then fixed and immunostained with mouse anti-cytochrome c MAb followed by Alexa488-conjugated goat anti-mouse IgG.

1512 KIM ET AL. J. VIROL.

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

drial membranes (Fig. 6B), we found that incubation of iso-lated mitochondria with 100 �M the H1 or H3 peptide did notcause cytochrome c release, suggesting that any pores formedby the H1 peptide in mitochondrial membranes do not allowegress of cytochrome c from the mitochondrial intermembranespace (Fig. 5B).

Microinjected H1 does not cause release of cytochrome cfrom mitochondria. We have previously shown that expressionof GFP-� causes release of cytochrome c from the mitochon-drial intermembrane space at 48 h posttransfection (31). How-ever, we found that purified mitochondria did not release cy-tochrome c when they were incubated with the H1 peptide(Fig. 5B). To test whether the H1 peptide could cause releaseof cytochrome c from mitochondria in cells, we microinjectedH1 (final concentration, 50 �M) together with BSA fluores-cently labeled with Alexa594 into CV-1 cells. After an incuba-tion period of 20 to 30 min, the cells were fixed and immuno-stained for cytochrome c. We found that, similar to ourfindings with purified mitochondria, the H1 peptide did notcause release of cytochrome c from mitochondria in cells aftera 20-min incubation period (Fig. 5C). As expected, microin-jection of the H1:I595K or H3 peptide also did not causerelease of cytochrome c. We conclude that although the H1peptide can permeabilize the plasma membrane and artificialmembranes, it does not permeabilize mitochondrial mem-branes to an extent that will allow release of cytochrome c.

Biotinylated H1 localizes to mitochondria when microin-jected into cells. A GFP-H1 fusion protein localizes to mi-

tochondria in transfected cells (3). We therefore assessedthe subcellular distribution of a biotinylated form of H1complexed with streptavidin-Alexa488 after it was micro-injected into CHO-K1 cells. Control microinjections ofstreptavidin-Alexa488 alone were diffusely distributedthroughout the cell (Fig. 6A). However, the biotinylated H1colocalized with the mitochondrial marker MitotrackerCMXRos (Fig. 6B). In addition, 1 min following microin-jection, the mitochondria in several H1-injected cells be-came punctate and aggregated and differed from the normaltubulovesicular distribution of mitochondria in control in-jected cells (Fig. 6B, insets). These findings indicate that H1localizes to mitochondrial membranes and disrupts normalmitochondrial architecture.

Biotinylated H1 but not H1:I595K peptide forms a filamen-tous mosaic on the plasma membrane when incubated withcells. To further investigate the mechanism of H1-inducedplasma membrane permeability, we assessed the cell surfacedistribution of biotinylated H1 and H1:I595K. CV-1 monkeykidney cells were incubated with 50 �M H1 for 30 min at RTand then washed and fixed, and the distribution of biotinylatedH1 was detected by the addition of streptavidin-Alexa488. Thebiotinylated H1 peptide formed a filamentous mosaic-like pat-tern on the plasma membrane of the CV-1 cells. In addition,larger aggregates of peptide could be seen adherent to theplasma membrane. In contrast, the mutant peptide H1:I595Kdistributed as punctate spots on the plasma membrane and didnot form larger aggregates (Fig. 7). We conclude from these

FIG. 6. Microinjected H1 localizes to mitochondrial membranes and disrupts normal mitochondrial architecture. CHO-K1 cells labeled withMitotracker CMXRos were microinjected with Alexa488-conjugated streptavidin (A) or 50 �M a biotinylated H1 peptide bound with Alexa488-conjugated streptavidin (B). Phase-contrast and fluorescent images were collected before and 1 min after microinjection. Injected cells are markedwith red stars. Insets show an enlarged view of the region outlined in dotted lines. Bar, 10 �m.

VOL. 85, 2011 REOVIRUS �1 AND PEPTIDE MEMBRANE PERMEABILIZATION 1513

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

findings that the H1 peptide can form filamentous arrays uponinteraction with cellular membranes perhaps similar to thosedescribed for some cell-penetrating peptides upon interactionwith artificial membranes (22, 23).

DISCUSSION

The reovirus outer capsid protein �1 is the primary deter-minant of reovirus-induced apoptosis (3, 5–7). �1 also has an

additional critical role in membrane penetration during virusentry. In this regard, a myristoylated N-terminal fragment of�1 (�1N; residues 2 to 42) that is produced by autocatalyticcleavage of �1 during virus entry has direct membrane-perme-abilizing activity (21, 34). The C-terminal � fragment of �1,although not functionally required for entry, increases its effi-ciency and was suggested to have chaperone-like activity (13).In previous work, we showed that ectopic expression of GFP-

FIG. 7. The H1 peptide forms filaments on the plasma membrane. (A) CV-1 cells were incubated with 50 �M biotin-conjugated H1,biotin-conjugated H1:I595K, or PBS. Biotinylated peptides were detected with Alexa488-conjugated streptavidin. Nuclei are stained with DAPI.Bar, 10 �m. (B) Higher-magnification images of cells incubated with biotinylated H1 showing aggregates and filaments and biotinylated H1:I595Kshowing punctate spots. Bar, 5 �M.

1514 KIM ET AL. J. VIROL.

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

fused to residues 582 to 611 of �1 (within �) localized tomitochondria and induced apoptosis (3). More recently, wehave found that ectopic expression of �1 or reovirus infectioncauses release of cytochrome c and apoptosis in cells lackingthe proapoptotic BCL-2 family members Bax and Bak (31).We now show that residues 582 to 611 of �1 also have directmembrane-permeabilizing activity. These findings raise thepossibility that one mechanism by which �1 initiates apoptosisin reovirus-infected cells is by directly permeabilizing cellularmembranes. The �1 protein localizes to mitochondria and theER in infected cells (3), and we speculate that direct perme-abilization of these membranes with release of proapoptoticmolecules into the cytosol initiates the apoptotic process ininfected cells. Our data show that the H1 peptide directlypermeabilizes artificial membranes, causing release of smallmolecules up to 10 kDa in size, but cannot itself cause releaseof cytochrome c from the intermembrane space of purifiedmitochondria. In contrast, ectopic expression of GFP-� causesrelease of cytochrome c by 48 h posttransfection (31). It ispossible that the release of cytochrome c and smac/DIABLO,which occurs during reovirus infection (14, 15), is a conse-quence of direct permeabilization of the outer mitochondrialmembrane by �1, as the full-length �1 protein may be able tooligomerize and form larger pores in intracellular membranes.However, cleaved tBid is required for reovirus-induced apop-tosis (8). Therefore, it is possible that �1 or a fragment of �1acts together with cleaved tBid to permeabilize mitochondrialmembranes in the absence of Bax and Bak. The activation ofcalpain during reovirus infection (9) might be explained byleakage of calcium from the ER as a consequence of �1-induced membrane damage.

Our data also indicate that the H1 peptide can directlypermeabilize the plasma membranes of cells, leading to cyto-toxicity that is accompanied by increases in intracytosolic cal-cium. Although we initially expected that the increased calciumresulted from influx from the extracellular environment, wefound that cytosolic calcium levels significantly increased evenwhen extracellular calcium was absent. We further showed thatthis increase in cytosolic calcium was not a consequence ofactivation of phospholipase C signaling. Our findings that di-rect application of biotinylated H1 to cells leads to localizationof the peptide to mitochondria suggest that in following plasmamembrane permeabilization, H1 can access the cytosol andtarget mitochondrial and perhaps other intracellular mem-branes. A recombinant mutant virus bearing the I595K muta-tion in �1 has significantly decreased capacity to activateNF- B and induce apoptosis (5). The efficiency of this mutantin inducing hemolysis of red blood cells and in entering cellswas comparable to that of wild-type virus. However, the H1peptide bearing this mutant was not able to permeabilize cel-lular membranes. Our findings suggest a strong correlationbetween �1-induced membrane permeabilization and apopto-sis induction but clearly discriminate the capacity of � to per-meabilize membranes from its proposed chaperone role duringvirus entry.

The H1 peptide had moderate antibacterial effects againstGram-positive and Gram-negative bacteria. Peptides derivedfrom HIV-1 gp41 have also been shown to kill Gram-positiveand Gram-negative bacteria (27). The mechanism of the anti-bacterial action of H1 is likely related to its membrane-desta-

bilizing properties. The action of cationic amphipathic antibac-terial peptides such as magainin on Gram-negative bacteria isthought to be due to the capacity of these peptides to displaceMg2 and Ca2 cations from lipopolysaccharide (LPS), caus-ing localized destabilization of the bacterial outer membrane.The peptides can then disrupt the cytoplasmic membrane in aprocess called self-promoted uptake (11, 24).

The formation of filamentous structures on the plasmamembrane by H1 indicates that it can self-associate in thepresence of membranes to form higher-order oligomeric struc-tures. In contrast, the mutant H1:I595K, which does not de-stabilize membranes, although able to associate with theplasma membrane, did not form such structures, suggestingthat the capacity of H1 to self-associate or aggregate may berequired for membrane destabilization. The micellar aggregatemodel of peptide-mediated membrane destabilization pro-poses that peptide aggregation within the membrane leads tothe formation of pores (12). Furthermore, collapse of suchaggregates is predicted to allow translocation of peptides intothe cytosol. In support of this mode of action for H1, we havefound that some cells incubated with biotinylated H1 havepeptide labeling on organellar membranes (data not shown).

We have proposed that when large numbers of viral particlesare added to cells, the � fragment of �1 may gain access to thecytosol of cells and induce apoptosis (3, 5). This hypothesiswould in part explain the capacity of UV-inactivated virions toinduced apoptosis (30) and the requirement for disassembly ofvirions in order to induce apoptosis (4). Our current findingsnow indicate that the minimal region of � required for apop-tosis induction has direct membrane-destabilizing activity andcan target mitochondrial membranes. However, we found thatneither incubation of the H1 peptide with purified mitochon-dria nor microinjection of H1 into cells led to release of cyto-chrome c from the mitochondrial intermembrane space. Takentogether, these findings support a model whereby C-terminalfragments of �1 produced during virus entry and delivered intothe cytosol during membrane penetration or produced in thecytosol as a consequence of the action of cytosolic proteasescan target and destabilize mitochondrial and other intracellu-lar membranes. Such destabilization would be predicted tolead to direct release of calcium and to activation of cellularstress response pathways that would in aggregate promoteapoptosis. The recent finding that tBid is required for reovirus-induced apoptosis supports this model (8). We have previouslynoted differences in the localization patterns of �1 and GFP-�,with GFP-� being much more strongly associated with mito-chondrial membranes and �1 more often seen in associationwith the ER and lipid droplets (3). We propose that targetingof �1 to the ER leads to local release of ER calcium. Calciumrelease from the ER could be responsible for activating cal-pain, which has been shown to be required for reovirus-in-duced apoptosis (9). Our current findings suggest that reovi-rus-induced apoptosis may initiate as a consequence of�1-induced intracellular membrane permeabilization.

ACKNOWLEDGMENTS

We thank Brian Ingel and Wendy Weichert for excellent technicalassistance. We thank Susan Daniel for technical advice.

VOL. 85, 2011 REOVIRUS �1 AND PEPTIDE MEMBRANE PERMEABILIZATION 1515

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from

This research was supported by Public Health Service awards R01AI063036 (J.S.L.P.) and a Burroughs Wellcome Fund Investigators ofPathogenesis of Infectious Disease award (J.S.L.P.).

REFERENCES

1. Billen, L. P., C. L. Kokoski, J. F. Lovell, B. Leber, and D. W. Andrews. 2008.Bcl-XL inhibits membrane permeabilization by competing with Bax. PLoSBiol. 6:e147.

2. Clarke, P., and K. L. Tyler. 2009. Apoptosis in animal models of virus-induced disease. Nat. Rev. Microbiol. 7:144–155.

3. Coffey, C. M., et al. 2006. Reovirus outer capsid protein �1 induces apoptosisand associates with lipid droplets, endoplasmic reticulum, and mitochondria.J. Virol. 80:8422–8438.

4. Connolly, J. L., and T. S. Dermody. 2002. Virion disassembly is required forapoptosis induced by reovirus. J. Virol. 76:1632–1641.

5. Danthi, P., C. M. Coffey, J. S. Parker, T. W. Abel, and T. S. Dermody. 2008.Independent regulation of reovirus membrane penetration and apoptosis bythe �1 � domain. PLoS Pathog. 4:e1000248.

6. Danthi, P., M. W. Hansberger, J. A. Campbell, J. C. Forrest, and T. S.Dermody. 2006. JAM-A-independent, antibody-mediated uptake of reovirusinto cells leads to apoptosis. J. Virol. 80:1261–1270.

7. Danthi, P., et al. 2008. Reovirus apoptosis and virulence are regulated byhost cell membrane penetration efficiency. J. Virol. 82:161–172.

8. Danthi, P., et al. 2010. Bid regulates the pathogenesis of neurotropic reovi-rus. PLoS Pathog. 6:e1000980.

9. Debiasi, R. L., et al. 1999. Reovirus-induced apoptosis is preceded by in-creased cellular calpain activity and is blocked by calpain inhibitors. J. Virol.73:695–701.

10. Grynkiewicz, G., M. Poenie, and R. Y. Tsien. 1985. A new generation of Ca2

indicators with greatly improved fluorescence properties. J. Biol. Chem.260:3440–3450.

11. Hancock, R. E. 1997. Peptide antibiotics. Lancet 349:418–422.12. Hancock, R. E., and D. S. Chapple. 1999. Peptide antibiotics. Antimicrob.

Agents Chemother. 43:1317–1323.13. Ivanovic, T., et al. 2008. Peptides released from reovirus outer capsid form

membrane pores that recruit virus particles. EMBO J. 27:1289–1298.14. Kominsky, D. J., R. J. Bickel, and K. L. Tyler. 2002. Reovirus-induced

apoptosis requires both death receptor- and mitochondrial-mediatedcaspase-dependent pathways of cell death. Cell Death Differ. 9:926–933.

15. Kominsky, D. J., R. J. Bickel, and K. L. Tyler. 2002. Reovirus-inducedapoptosis requires mitochondrial release of Smac/DIABLO and involvesreduction of cellular inhibitor of apoptosis protein levels. J. Virol. 76:11414–11424.

16. Kuwana, T., et al. 2002. Bid, Bax, and lipids cooperate to form supramo-lecular openings in the outer mitochondrial membrane. Cell 111:331–342.

17. Lal, R., D. Harris, S. Postel-Vinay, and J. de Bono. 2009. Reovirus: rationaleand clinical trial update. Curr. Opin. Mol. Ther. 11:532–539.

18. Leuschner, C., and W. Hansel. 2004. Membrane disrupting lytic peptides forcancer treatments. Curr. Pharm. Des. 10:2299–2310.

19. Liemann, S., K. Chandran, T. S. Baker, M. L. Nibert, and S. C. Harrison.2002. Structure of the reovirus membrane-penetration protein, �1, in acomplex with is protector protein, �3. Cell 108:283–295.

20. Nibert, M. L., and B. N. Fields. 1992. A carboxy-terminal fragment of protein�1/�1C is present in infectious subvirion particles of mammalian reovirusesand is proposed to have a role in penetration. J. Virol. 66:6408–6418.

21. Odegard, A. L., et al. 2004. Putative autocleavage of outer capsid protein �1,allowing release of myristoylated peptide �1N during particle uncoating, iscritical for cell entry by reovirus. J. Virol. 78:8732–8745.

22. Plenat, T., S. Boichot, P. Dosset, P. E. Milhiet, and C. Le Grimellec. 2005.Coexistence of a two-states organization for a cell-penetrating peptide inlipid bilayer. Biophys. J. 89:4300–4309.

23. Plenat, T., et al. 2004. Interaction of primary amphipathic cell-penetratingpeptides with phospholipid-supported monolayers. Langmuir 20:9255–9261.

24. Powers, J. P., and R. E. Hancock. 2003. The relationship between peptidestructure and antibacterial activity. Peptides 24:1681–1691.

25. Prestwich, R. J., et al. 2009. Immune-mediated antitumor activity of reovirusis required for therapy and is independent of direct viral oncolysis andreplication. Clin. Cancer Res. 15:4374–4381.

26. Strong, J. E., M. C. Coffey, D. Tang, P. Sabinin, and P. W. Lee. 1998. Themolecular basis of viral oncolysis: usurpation of the Ras signaling pathway byreovirus. EMBO J. 17:3351–3362.

27. Tencza, S. B., J. P. Douglass, D. J. Creighton, Jr., R. C. Montelaro, and T. A.Mietzner. 1997. Novel antimicrobial peptides derived from human immuno-deficiency virus type 1 and other lentivirus transmembrane proteins. Anti-microb. Agents Chemother. 41:2394–2398.

28. Thirukkumaran, C., and D. G. Morris. 2009. Oncolytic viral therapy usingreovirus. Methods Mol. Biol. 542:607–634.

29. Tian, P., et al. 1995. The rotavirus nonstructural glycoprotein NSP4 mobi-lizes Ca2 from the endoplasmic reticulum. J. Virol. 69:5763–5772.

30. Tyler, K. L., et al. 1995. Differences in the capacity of reovirus strains toinduce apoptosis are determined by the viral attachment protein �1. J. Virol.69:6972–6979.

31. Wisniewski, M. L., et al. 27 October 2010. Reovirus infection or ectopicexpression of outer-capsid protein �1 induces apoptosis independently of thecellular proapoptotic proteins Bax and Bak. J. Virol. [Epub ahead of print.]doi:10.1128/JVI.01982-10.

32. Yang, J., et al. 1997. Prevention of apoptosis by Bcl-2: release of cytochromeC from mitochondria blocked. Science 275:1129–1132.

33. Zelezetsky, I., and A. Tossi. 2006. Alpha-helical antimicrobial peptides—using a sequence template to guide structure-activity relationship studies.Biochim. Biophys. Acta 1758:1436–1449.

34. Zhang, L., et al. 2009. Requirements for the formation of membrane poresby the reovirus myristoylated �1N peptide. J. Virol. 83:7004–7014.

1516 KIM ET AL. J. VIROL.

on July 9, 2014 by Cornell U

niversity Libraryhttp://jvi.asm

.org/D

ownloaded from