In Silico and In Vivo Evaluation of Bacteriophage {phi}EF24C, a Candidate for Treatment of Enterococcus faecalis Infections

Along with the increasing threat of nosocomial infections byvancomycin-resistant Enterococcus faecalis, bacteriophage (phage)therapy has been expected as an alternative therapy againstinfectious disease. Although genome information and proof ofapplicability are prerequisites for a modern therapeutic phage,E. faecalis phage has not been analyzed in terms of these aspects.Previously, we reported a novel virulent phage, ${phi}$ EF24C, and itsbiology indicated its therapeutic potential against E. faecalisinfection. In this study, the ${phi}$ EF24C genome was analyzed andthe in vivo therapeutic applicability of ${phi}$ EF24C was also brieflyassessed. Its complete genome (142,072 bp) was predicted tohave 221 open reading frames (ORFs) and five tRNA genes. Inour functional analysis of the ORFs by use of a public database,no proteins undesirable in phage therapy, such as pathogenicand integration-related proteins, were predicted. The noncompetitivedirections of replication and transcription and the host-adaptedtranslation of the phage were deduced bioinformatically. Itsgenomic features indicated that ${phi}$ EF24C is a member of the SPO1-likephage genus and especially that it has a close relationshipto the Listeria phage P100, which is authorized for prophylacticuse. Thus, these bioinformatics analyses rationalized the therapeuticeligibility of ${phi}$ EF24C. Moreover, the in vivo therapeutic potentialof ${phi}$ EF24C, which was effective at a low concentration and wasnot affected by host sensitivity to the phage, was proven byuse of sepsis BALB/c mouse models. Furthermore, no change inmouse lethality was observed under either single or repeatedphage exposures. Although further study is required, ${phi}$ EF24C canbe a promising therapeutic phage against E. faecalis infections.

INTRODUCTION

Top

INTRODUCTION

Enterococcus species can be found in the environment and normalmicroflora of large mammals, and some of these bacterial speciesoccasionally cause a variety of diseases (30). The emergenceof vancomycin-resistant Enterococcus (VRE) disrupts effectiveconventional chemotherapy and causes fatal infections in nosocomialsettings. An increase in the number of VRE cases has been reportedworldwide (5, 9, 14). Among the enterococcal infections, E.faecalis is the most clinically isolated species (19, 30).

Bacteriophage (phage) therapy harnesses a live prokaryotic virusas a bioagent to target and destroy disease-causing bacteria(7, 15). Phage therapy has a long history of successful usein the former Eastern bloc countries, whereas it has almostno such history in the West (15, 27). The recent increase inthe number of multiple-drug-resistant bacteria including VREhas renewed the interest of the Western scientific communityin phage therapy (15, 27). However, because the past failuresin phage therapy resulted from a lack of scientific knowledgeof phage biology, this therapeutic approach needs to be scientificallyrationalized (27). Hence, each therapeutic phage needs to bewell characterized, like other approved drugs.

Genome analysis and proof of applicability are simple and effectivemethods for the primary evaluation of each phage. First, thephage genome reflects biological information such as morphology(e.g., drug formulation) and life cycle (e.g., propagation mechanismand drug efficacy), enabling the elucidation of phage drug features.In addition, genome analysis allows us to examine the safetyof the approach by determining the presence or absence of undesirablegenes such as pathogenic and integration-related genes (36,43). Moreover, the lytic activity and intrinsic effects of eachphage in vivo are usually unknown, so that in vivo therapeuticeffectiveness and phage toxicity must also be examined. Unfortunately,no therapeutic phage with such evaluation is currently availableagainst E. faecalis infections.

Previously, ${phi}$ EF24C was isolated and its biology was characterizedbriefly. ${phi}$ EF24C, which was classified in the family Myoviridaemorphotype A1, has a broad host specificity with strong virulenceagainst E. faecalis, including the VRE strains (47). The morphologyof ${phi}$ EF24C, together with the N-terminal sequences of its structuralproteins, implies a relationship to members of the SPO1-likephage genus (28, 47). Some of these members are considered tobe therapeutic and prophylactic phage candidates (e.g., Staphylococcusphages K and 812 and Listeria phage P100) (8, 20, 29, 40). Consequently, ${phi}$ EF24C was proposed as a putative therapeutic candidate. In thepresent study, the ${phi}$ EF24C genome was analyzed for the first time.Next, in vivo therapeutic effectiveness was evaluated usingsevere sepsis mouse models infected with an E. faecalis strainhaving either high or low phage sensitivity. In addition, phagetoxicity was briefly examined in vivo.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Culture media.
Both tryptic soy broth (TSB) and heart infusion broth were obtainedfrom Becton, Dickinson and Company (Sparks, MD). Constituentsof culture media used were purchased from Nacalai Tesque (Kyoto,Japan) unless otherwise stated. Heart infusion broth supplementedwith 20 mM MgCl₂ and 20 mM CaCl₂ (HIMC) was prepared. For thephage plaque formation assay, TSB-based solid media containing1.5% and 0.5% agar were used for the lower and upper layers,respectively.

Bacterial strains.
E. faecalis strain EF24 was employed as a host for the phagepropagation and phage plaque formation assay. E. faecalis strainsEF14 and VRE2 were employed in the animal experiments. Thesebacterial strains were described previously (47).

Phage purification.
Phage purification was carried out as described previously (47).Briefly, phage was propagated with host strain EF24 in 1 literof TSB medium. After the removal of bacterial debris by centrifugation(10,000 x g, 4°C, 10 min) and supplementation of the lysatewith polyethylene glycol 6000 (Sigma-Aldrich Co., MI) and NaCl(final concentrations of 10% and 0.5 M, respectively), phagewas precipitated by centrifugation (10,000 x g, 4°C, 30min). Phage precipitate was treated with DNase I (type II; Sigma-Aldrich)and RNase A (type IA; Sigma-Aldrich) (both 50 µg/ml).Finally, phage was sequentially purified by CsCl step gradientultracentrifugation (50,000 x g, 4°C, 2 h) twice. An S80AT3rotor and a GX series Himac CS 100GX microultracentrifuge (HitachiLtd., Tokyo, Japan) were used for ultracentrifugation. Afterthe phage band was collected, the purified phage was treateddifferently for each experimental purpose.

Genome sequencing.
The extraction of purified phage DNA was carried out as describedpreviously (47). Briefly, the purified phage suspension containingCsCl was diluted with AAS (0.1 M ammonium acetate, 10 mM NaCl,1 mM CaCl₂, 1 mM MgCl₂, pH 7.2) four times, and phage was pelletedby ultracentrifugation (100,000 x g, 4°C, 1 h). The phagepellet was incubated with proteinase K (Takara Bio, Kyoto, Japan),and phenol extraction and ethanol precipitation were conducted.Finally, the genome DNA was solubilized in water.

The phage DNA was digested by restriction endonuclease HindIII(50 ng DNA/U in 50 µl, 37°C, 4 h) (Takara Bio) andthen electrophoresed in 0.8% agarose. After visualization withethidium bromide (1 µg/ml), the DNA fragments were excisedand extracted from the gel. The DNA fragments were cloned intopUC19 vectors and transformed into competent Escherichia coliDH5 ${alpha}$ .

Sequencing of each cloned fragment was performed by PCR, usingthe BigDye Terminator v1.1 cycle sequencing kit (Applied Biosystems,CA). The PCR products were purified through a Sephadex G-50column (Sigma-Aldrich) and were then analyzed using an ABI Prism3100-Avant genetic analyzer (Applied Biosystems). Cloned fragmentswere sequenced by primer walking. The regions uncovered by thecloned fragments were amplified by PCR with primers based onthe cloned fragments, and then the PCR products were purifiedby LaboPass PCR (Hokkaido System Science, Sapporo, Japan) andsequenced by primer walking as described above. Both strandswere sequenced, and the sequence data were connected using theGenetyx-Mac ATSQ program, version 4.2.1 (Genetyx Co., Tokyo,Japan). The sequence coverage redundancy was at least double.

Genome analysis.
The potential open reading frames (ORFs) that possibly encodethe gene products were first predicted by the following genepredication tools: GeneMark VIORIN (http://opal.biology.gatech.edu/GeneMark/)(3), FGENESB (http://www.softberry.com/berry.phtml), and MicrobialGenome Annotation Tools (http://www.ncbi.nlm.nih.gov/genomes/MICROBES/glimmer_3.cgi)(12, 46). ATG, TTG, and GTG were considered to be start codons,and TAA, TGA, and TAG were considered to be stop codons. TheORFs were then determined from the program-predicted ORFs basedon the criteria of a length of more than 72 nucleotides anda maximum length equivalent to that of the program-predictedORFs with stop codons at the same locations. To examine thetherapeutic eligibility of ${phi}$ EF24C strictly, such criteria werepurposely used to determine the ORFs in this study. To increasethe possibility of identifying protein-coding sequences, theribosomal binding site (RBS) sequence of each ORF was also subsequentlyinvestigated. Moreover, tRNA genes were predicted using thetRNAscan-SE program (http://lowelab.ucsc.edu/tRNAscan-SE/) (34).

The putative products of the ORFs were analyzed by BLASTP atthe National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi)and by InterProScan with BlastProDom, FPrintScan, HMMPIR, HMMPfam,HMMSmart, HMMTigr, ProfileScan, ScanRegExp, SuperFamily, HMMPanther,and Gene3D at the European Bioinformatics Institute (http://www.ebi.ac.uk/InterProScan/)(38). Transmembrane domains and signal peptides were also predictedby TMHMM (http://www.cbs.dtu.dk/services/TMHMM/) and SignalIP3.0 (http://www.cbs.dtu.dk/services/SignalP/), respectively(2, 31). Using an E value threshold of 0.1 for both BLASTP andInterProScan, the functions of the putative gene products werespecified.

The following genome features were also analyzed using in silicomolecular cloning genomics edition (In Silico Biology, Inc.,Yokohama, Japan): GC content, GC scanning, GC skew, cumulativeGC skew, and codon usage.

An unrooted phylogenic tree was constructed by the neighbor-joiningmethod, using DNASIS Pro (Hitachi Software Engineering Co.,Ltd., Tokyo, Japan). Bootstrap analysis was performed by resamplingthe data sets 10,000 times. Bootstrap values of greater than95% were considered to be statistically significant for thegrouping.

The gene order was compared with the most related phage genome;this comparison was manually performed by in-house BLAST search,using in silico molecular cloning genomics edition (In SilicoBiology, Inc.). An E value of less than 0.1 was considered asindicative of homology.

Comparative data for the following were retrieved from the GenBankdatabase: bacteriophage G1 (accession number AY954969), Staphylococcusphage K (accession number AY176327), Staphylococcus phage Twort(accession number AY954970), Listeria bacteriophage P100 (accessionnumber DQ004855), Lactobacillus plantarum bacteriophage LP65(accession number AY682195), and E. faecalis V583 (accessionnumber AE016830.1).

Phage preparation for animal experiments.
For the purpose of a mouse rescue experiment, the purified phagesample was continuously dialyzed against SMC (saline with 20mM MgCl₂ and 20 mM CaCl₂) (4°C, 30 min) and HIMC (4°C,30 min). On the other hand, the purified phage sample was dialyzedagainst SMC (4°C, 60 min) for the repeated administration.The titers (PFU/ml) of the phage were then determined by inoculationinto bacterial strain EF24. The phage was stored at 4°Cuntil use.

Animal experiments.
All animal experiments were conducted with the approval of theAnimal Experiment Committee of Kochi Medical School. Female6- to 8-week-old BALB/c mice (weighing up to 18 g) were usedin the following experiments.

E. faecalis cells (either strain EF14 or strain VRE2) were grownin 300 ml of TSB medium at 37°C until the early stationaryphase (up to ca. 200 Klett units) and were then centrifuged(10,000 x g, 4°C, 10 min). The cell pellet was washed with300 ml of saline, centrifuged (10,000 x g, 4°C, 10 min),and finally resuspended in 3 ml of saline. After appropriatedilution using saline, the bacterial concentration (bacteria/ml)was determined by turbidity (in Klett units), measured witha Klett-Summerson photoelectric colorimeter (Klett Mfg. Co.,NY). EF14 was prepared to be at 1.0 x10¹⁰, 2.0 x 10¹⁰, 5.0 x10¹⁰, 1.0 x 10¹¹, 2.0 x 10¹¹, and 5.0 x10¹¹ bacteria/ml. VRE2was prepared to be at 5.0 x 10¹⁰, 1.0 x 10¹⁰, 2.1 x 10¹⁰, 5.0x 10¹⁰, 1.0 x 10¹¹, and 2.1 x 10¹¹ bacteria/ml. Saline (0.2ml; control) or bacterial suspension at different concentrationswas injected into the peritoneal cavities of 5 mice throughthe left side of the abdomen (in total, 70 mice were used).The survival rates for and activities of all tested animalswere observed for 7 days. The minimum bacterial concentrationshowing 100% lethality was determined as the minimum lethalbacterial dosage.

For the mouse rescue experiment by phage, the phage was dilutedin HIMC to the following concentrations (expressed in multiplicitiesof infection [MOI]): 100, 10, 1, 0.1, 0.01, 0.001, and 0.0001.Mice were inoculated on the left side of the abdomen with 0.2ml of the minimum lethal bacterial dose (in total, 70 mice).About 20 min after the inoculation of the lethal bacterial dose,0.2 ml of phage solution at different concentrations (in MOI),HIMC, or saline was administered to five mice on the right sideof the abdomen. The data on the survival rates of the mice wereanalyzed statistically with a two-tailed Fisher's exact test.Moreover, to measure the intrinsic effects of phage alone andbuffers, 0.5-ml aliquots of HIMC, saline, or phage suspension(total, 1.0 x 10¹² PFU) alone was administered into the abdominalcavities of 5 mice (a total of 15 mice). The survival ratesfor and activities of all tested animals were observed for 7days.

To examine the effects of repeated phage exposure, 0.5 ml ofan SMC phage suspension (in total, 3.5 x 10¹⁰ PFU) or SMC alonewas intraperitoneally administered seven times at 4-day intervalsinto the abdominal cavities of 10 mice (a total of 20 mice).The mouse survival rate was recorded for 2 months from the initialphage administration.

Nucleotide sequence accession number.
The genome data of phage ${phi}$ EF24C was deposited to GenBank (accessionnumber AP009390).

RESULTS AND DISCUSSION

Top

RESULTS AND DISCUSSION

General genome description.
The ${phi}$ EF24C genome was determined to be 142,072 bp. Also, thegenome sequence was circularly permuted because no definiteterminal ends were identified by genomic sequencing. The GCcontent was 35.7%. Two hundred twenty-one ORFs and five tRNAgenes were determined. When we checked for the presence of anRBS upstream from each ORF, most ORFs seemed to have a typicalRBS (see the supplemental material). In the following text,ORFs considered to have a functional putative product were definedas putative genes and are designated using an "orf" prefix.The putative gene (i.e., orf) product is likewise shown using"Orf."

Bioinformatic analysis revealed that 45.2% (100/221) of theputative ORF products were assumed to have either protein functionaldomains or similarity to other phage gene products or both.Overall, 20.8% (46/221) of the ORFs were deduced to encode functionalproteins. No genes coding for site-specific integrase, toxinand antibiotic resistance genes, or other pathogenic factorswere predicted. A BLASTP search on the hypothetical ORF-encodedproteins showed similarities with the gene products of the otherlarge virulent phages, including Staphylococcus phage K, G1,Twort, Lactobacillus phage LP65, and Listeria phage P100. Thegenomes of these phages are well characterized (8, 10, 32, 39).The genome map of ${phi}$ EF24C is shown in Fig. 1. The annotation ofthe genome is also shown in Table 1.

View larger version (30K):
[in this window]
[in a new window]

FIG. 1. Genome map of phage ${phi}$ EF24C. Arrows indicate putative ORFs and tRNA genes, along with their orientations. Functionally assigned genes are differently colored (blue, structural gene; red, lysis gene; green, DNA-associated gene; violet, tRNA gene). Speculated modules are enclosed by boxes (black, structural module; pink, DNA replication module). *, gene for structural protein.

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

TABLE 1. Features of phage ${phi}$ EF24C gene products and their functional assignments

The N-terminal amino acid sequences of the products of the orf16,orf23, orf24, orf40, orf69, and orf219 genes precisely correspondedto those of six virion proteins (42, 62, 16, 26, 21, and 12.6kDa, respectively) of ${phi}$ EF24C as reported in our previous paper(47). Only Orf16 and Orf23 were bioinformatically specifiedas major capsid protein (MCP) and tail sheath protein, respectively;the others were unknown. Except for that of Orf219, the N-terminalends were considered to be processed. During morphogenesis,the N-terminal 20 amino acid residues of the MCP (Orf16) wereconsidered to be removed, which was assumed to be mediated bythe putative prohead protease (Orf14). The other structuralproteins (Orf23, Orf24, Orf40, and Orf69) were assumed to besimilarly digested between methionine and alanine.

In general, functionally relevant genes are clustered as a modulein phage (6, 11). Three modules were speculated to range from1 bp to 71 kbp in ${phi}$ EF24C, according to genome annotation andstructural protein identification (Fig. 1). The large structuralmodule seemed to be associated with head and tail components(1 bp to 46 kbp). The replication-associated genes were clusteredas a DNA replication module (46 kbp to 68 kbp). The small structuralmodule included at least orf68 and orf69 (69 kbp to 71 kbp)due to an immunoglobulin (Ig)-like domain specified on Orf68(an Ig-like domain is typically found in a virion protein ofphage) and the identification of Orf69 as a structural proteinfrom the proteomic analysis (16, 17).

In the functionally uncategorized region, some putative genesassociated with de novo synthesis of nucleic acid precursorswere speculated. These included the genes for Orf103 (cytidinedeaminase), Orf203 (thymidylate synthase), and Orf208, Orf209,and Orf210 (ribonucleotide diphosphate reductase). Like T4 phage, ${phi}$ EF24C may produce its own modified base from host DNA breakdown(22, 37).

At the final stage of the latent period, progeny phages withina bacterial cell were released by the degradation of the cellwall (25, 27). This cell wall lysis is typically induced bytwo phage-encoded proteins called holin and endolysin (25, 27).At the late period of infection, holin forms a hole in the cellmembrane, and endolysin passes through the hole and destroysthe peptidoglycan structure (25, 27). In the phage ${phi}$ EF24C genome,the putative genes for endolysin (Orf9 and Orf10) and holin(Orf67) are thought to be distantly positioned, as occurs inphage T4 (37). orf9 and orf10 are thought to be located in onestructural module, and orf67 is thought to be located in veryclose proximity to another structural module. Thus, these genesare possibly expressed late in the period of infection (37).Consequently, Orf9, Orf10, and Orf67 may function as a holin-endolysinsystem.

Phylogenetic analyses of ${phi}$ EF24C within the SPO1-like phages.
The ${phi}$ EF24C genome contains approximately 142 kbp, the GC contentof which is 35.7%, as described above. The genome is circularlypermuted, and DNA polymerase A (Orf61) has been tentativelyidentified. These attributes, together with the morphologicaland biological features of ${phi}$ EF24C, suggest that ${phi}$ EF24C is a memberof the SPO1-like phage genus (28). Therefore, the phylogenicrelationships of ${phi}$ EF24C to the other SPO1-like phages were examined.

A phylogenic tree based on the MCP is frequently used in phagephylogenic analysis (1, 26). The MCPs of the following phageswere obtained from their genome sequences: Staphylococcus phageK, G1, Twort, Lactobacillus phage LP65, and Listeria phage P100.The MCP-based phylogenic analysis showed that ${phi}$ EF24C is mostclosely related to Listeria phage P100 among these phages (Fig.2A). Next, the gene organization of the ${phi}$ EF24C genome was alsocompared with that of Listeria phage P100. Genome synteny wasobserved, particularly on the predicted structural and DNA replicationmodules in ${phi}$ EF24C (ca. 70% of the genome) (Fig. 2B), whereasthe genes on the other region of ${phi}$ EF24C (the remaining ca. 30%of the genome) are not only functionally unknown but also dissimilarto any genes of Listeria phage P100. By considering the differencein bacterial species and phylogenic relation based on MCP, ${phi}$ EF24Cand Listeria phage P100 were considered to have evolved divergentlyfrom the same virus origin (44).

View larger version (17K):
[in this window]
[in a new window]

FIG. 2. Phylogenetic analyses of ${phi}$ EF24C. (A) Phylogenetic tree based on MCPs. (B) Comparison of the gene order between Enterococcus phage ${phi}$ EF24C and Listeria phage P100. The ORF of ${phi}$ EF24C is connected to that of P100 by a black line where the E value from the in-house BLAST search is less than 0.1.

Noncompetitive nature between replication and transcription directions.
The origin of replication (ori) and replication terminus (ter)can be deduced by GC skew and cumulative GC skew analyses (23,24, 45). The cumulative GC skew analysis for Fig. 3 shows thepredicted ori and ter, which are the lowest (ca. 1-bp) and highest(108-kbp) regions, respectively. Although the process of ${phi}$ EF24Cgenome replication remains unknown, it seems to replicate ina manner slightly different from that of another large phage,T4 of E. coli. The ${phi}$ EF24C genome seems to have only one ori,whereas phage T4 has multiple oris (37). Moreover, the speculatedreplication direction (ori to ter) matched with the directionof transcription (direction of genes). According to this mutualcorrespondence, bimolecular collisions during replication andtranscription can be avoided (13, 18). Hence, ${phi}$ EF24C is assumedto multiply without such mutual interference between replicationand transcription.

View larger version (16K):
[in this window]
[in a new window]

FIG. 3. GC-associated genome analysis. In the GC scanning, the low region in GC is indicated by the bar. In the cumulative GC skew, the origin of replication (black arrow) and the changing point of gene direction/termination of DNA replication (white arrow) are indicated.

Host-adapted translation.
Host and phage codon usage comparisons and phage tRNA analysisimplied efficient translation. The codon usages of ${phi}$ EF24C andE. faecalis V583 were mutually similar (Fig. 4A), which canindicate overall efficient translation in the host. Moreover,the codon usage of ${phi}$ EF24C exceeded that of the host on five predictedtRNAs whose genes were carried by the phage (Fig. 4B). Thus,we can infer that ${phi}$ EF24C supplies specific tRNAs on its own incase of tRNA deficiency.

View larger version (20K):
[in this window]
[in a new window]

FIG. 4. Optimization of ${phi}$ EF24C codon usage to its host codon usage and its possible tRNA function. (A) Comparison of codon usage between ${phi}$ EF24C and E. faecalis V583. (B) Location of tRNA genes (top), predicted secondary structures of tRNAs (middle), and phage and host codon usage comparisons on the tRNA anticodons (bottom). In the bar chart, codons for ${phi}$ EF24C tRNAs are indicated in boldface.

E. faecalis sepsis mouse model.
Sepsis mouse models are typically used for the preliminary assessmentof phage therapy against nosocomial bacteria (4, 35, 48-50).To examine therapeutic effectiveness and the effect of hostsensitivity difference on the phage in vivo, E. faecalis sepsismouse models using BALB/c mice were set up using two strains,EF14 and VRE2. In the previous study, EF14 had phage sensitivityabout 32 times greater than that of VRE2 (efficiency of plating[EOP] against EF14, 1; EOP against VRE2, 0.032) (47). Afterintraperitoneal bacterial inoculation at different concentrationsto mice, the minimum lethal bacterial dosages of EF14 and VRE2were determined to be 1.0 x 10¹⁰ and 4.2 x 10⁹ bacteria, respectively;these dosages resulted in 100% lethality within 2 days. Thesebacterial dosages were used for the following experiments toassess mouse rescue by phage.

In both cases, inoculation of the minimum lethal bacterial dosagesseemed to induce severe sepsis complications. As time elapsedafter bacterial inoculation, an increase in the number of bacteriain the blood was also observed (bacteremia). In addition, anincreasing frequency of unusual changes in the mice was observed,such as decrease in activity, low body temperature, shivering,blood clotting, and hyperventilation. These abnormal conditionswere considered to be typical features of severe sepsis, althoughhistological analyses were not performed (21, 41).

${phi}$ EF24C therapeutic effectiveness in vivo.
No abnormal mouse behavior or altered survival rate was observedfollowing the administration of saline, HIMC, or phage alone(1.0 x 10¹² PFU). Thus, the phage rescue experiments were consideredto be conducted without bias.

The in vivo therapeutic effectiveness of ${phi}$ EF24C was then examined.After the inoculation of the minimum lethal bacterial dosage,HIMC or ${phi}$ EF24C at different MOI of 10, 1, 0.1, 0.01, 0.001, and0.0001 was administered. Figure 5 shows the results of mouserescue experiments using ${phi}$ EF24C. The dose-dependent effectivenesswas observed to be less than MOI of 0.01 and 0.1 in the EF14and VRE2 mouse sepsis models, respectively. Compared with thecontrol (HIMC or saline treatment), the ${phi}$ EF24C treatment wassignificantly effective for both EF14-infected mice at MOI of10, 1, 0.1, and 0.01 (P < 0.01) and VRE2-infected mice atMOI of 10, 1, and 0.1 (P < 0.01) or 0.01 (P < 0.05). Accordingto these results, ${phi}$ EF24C can efficiently rescue mice infectedwith both EF14 and VRE2 at an MOI of 0.01. Under these experimentalconditions, the therapeutic efficacy of ${phi}$ EF24C did not seem tobe affected by the sensitivity of the host to the phage. The ${phi}$ EF24C burst size in vitro was reported as ca. 100 in a previouswork (47). Under in vivo conditions, the phage may not infectthe bacteria or propagate as efficiently as it does under invitro conditions. Therefore, these results suggest that theefficient propagation of ${phi}$ EF24C led to its significant therapeuticeffectiveness in vivo.

View larger version (14K):
[in this window]
[in a new window]

FIG. 5. Experiment examining mouse rescue by phage ${phi}$ EF24C administration. At approximately 20 min after the intraperitoneal inoculation of the minimum lethal bacterial dosage (EF14, 1.0 x 10¹⁰ bacteria; VRE2, 4.2 x 10⁹ bacteria) to BALB/c mice, different concentrations of phage (MOI of 10, 1, 0.1, 0.01, 0.001, and 0.0001) or HIMC medium (control) were administered to the opposite side of the abdominal cavities of five mice. The survival rate was recorded after 7 days. Values significantly different from the control values (P < 0.05 and P < 0.01) are indicated by asterisks and double asterisks, respectively.

Eligibility of ${phi}$ EF24C as a potential therapeutic phage.
${phi}$ EF24C is considered to be eligible as a therapeutic phage forthe following reasons. First, as demonstrated in a previousstudy, ${phi}$ EF24C has broad host specificity and strong virulenceagainst E. faecalis strains. Second, undesirable genes for phagetherapy such as integration-related and pathogenic (e.g., toxinand antibiotic resistance, etc.) genes have not yet been identified.Third, we can infer from the following features that the biologicalnature of ${phi}$ EF24C is appropriate for a therapeutic phage: itsde novo nucleic acid synthesis from host DNA breakdown, itsholin-endolysin system, its host-adapted translation, and thenoncompetitive nature of its transcription and replication.Fourth, its genomic features, together with its morphology,allow ${phi}$ EF24C to be categorized in the SPO1-like phage genus,some members of which, including phages K and P100, are usedor are under consideration for therapy or prophylaxis. Phylogeneticand genome synteny analyses revealed a close relationship between ${phi}$ EF24C and the Listeria phage P100, which has been approved forprophylactic use and has been commercialized (EBI Food Safety[http://www.ebifoodsafety.com/]) (8, 42). Thus, ${phi}$ EF24C can beencompassed in the general therapeutic phage group. Fifth, accidentalhomologous recombination is also not likely in the use of ${phi}$ EF24C,because similarities between the ${phi}$ EF24C genome and the host E.faecalis V583 genome were restricted to tRNA genes and a fewgenes for hypothetical proteins, as for the other therapeuticphages K and P100 (data not shown). This animal experiment showedthat a single low-dosage administration of ${phi}$ EF24C can effectivelytreat sepsis in mice without the effect of host sensitivityto the phage observed in vitro (i.e., EOP difference betweenstrains EF14 and VRE2). The phage bacteriolytic action is believedto be the primary mechanism of mouse rescue effects. However, ${phi}$ EF24C efficacy may be altered in different mouse strains andanimals. For example, some possible phage-inactivating factors,such as phage-neutralizing antibody and liver or bile acids,may alter phage efficacy, and the phage may not efficientlyreach the focus of infection. Therefore, further study is required(i.e., of the pharmacokinetics and pharmacodynamics associatedwith the compromised route of administration).

Safety issues are of great concern in phage therapy. Surprisingly,no significant side effects of phage therapy have been reportedto date in the East (36, 43). In this experiment, phage administrationdid not cause any lethality or mouse behavior change both inthe mouse rescue experiment and in the administration of a high-concentrationdosage of phage alone. In addition, repeated exposure to phage(administration seven times at 4-day intervals) did not causeany change in mouse behavior. However, different phages havedifferent molecular features, and different mouse strains havedifferent levels of immunity, so the safety of each phage mustbe examined in the future (25, 33, 36). In this study, ${phi}$ EF24Cwas investigated primarily as a therapeutic phage. Althoughfurther development of this phage and other methods is stillnecessary to address some remaining problems, ${phi}$ EF24C is a promisingtherapeutic phage against E. faecalis infections.

ACKNOWLEDGMENTS

We thank Hiromi Kataoka (Clinical Laboratory Centers, KochiMedical School) for access to DNASIS Pro and Toshimitsu Uchiyama(Toho University, Tokyo, Japan) for helpful scientific advice.

This study was supported by The Special Research Project ofGreen Science, Kochi University.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology and Infection, Kochi Medical School, Kohasu Oko-cho, Nankoku City, Kochi 783-8505, Japan. Phone: 81-88-880-2323. Fax: 81-88-880-2324. E-mail: matuzaki{at}kochi-u.ac.jp

Published ahead of print on 2 May 2008.

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

REFERENCES

Top