Correspondence: Nikolay V. Volozhantsev, State Research Center for Applied Microbiology and Biotechnology, Obolensk, Moscow Reg. 142279, Russia. Tel.: +7 4967 36 01 47; fax: +7 4967 36 00 61; e-mail: email@example.com
Acinetobacter baumannii plays a significant role in infecting patients admitted to hospitals. Many A. baumannii infections, including ventilation-associated pneumonia, wound, and bloodstream infections, are common for intensive care and burn units. The ability of the microorganism to acquire resistance to many antibiotics, disinfectants, and dehydration assures its long-term survival in hospital settings. The application of bacteriophages is a potential tool to control A. baumannii infections. Bacteriophage AP22 lytic for A. baumannii was isolated from clinical materials and classified as a member of the Myoviridae family. The phage had an icosahedral head of 64 nm in diameter and a contractile tail of 85–90 nm in length. According to restriction analysis, AP22 had 46-kb double-stranded DNA genome. The phage AP22 exhibited rapid adsorption (> 99% adsorbed in 5 min), a large burst size (240 PFU per cell), and stability to the wide range of pH. The bacteriophage was shown to specifically infect and lyse 68% (89 of 130) genotype-varying multidrug-resistant clinical A. baumannii strains by forming clear zones. Thus, it could be used as a candidate for making up phage cocktails to control A. baumannii-associated nosocomial infections.
Nosocomial infections and multidrug resistance of pathogens causing these infections are the growing and recognized problems in the modern healthcare system.
Acinetobacter baumannii is a gram-negative, nonfermenting aerobic microorganism that plays a significant role in infecting patients admitted to hospitals. Acinetobacter baumannii is often a cause of hospital pneumonia, wound and catheter-related urinary tract infections, postsurgery complications, and bloodstream infections especially in critically ill or immunocompromised patients (Peleg et al., 2008; Towner, 2009). The ability of the microorganism to develop resistance to major groups of antibiotics, as well as to disinfectants, detergents, dehydration, and UV radiation, assures its long-term survival and nosocomial spread in hospital environments especially in intensive care and burn units (Wendt et al., 1997; Webster et al., 1998; de Oliveira & Damasceno, 2010). There is an important therapeutic problem to treat infections caused by this microorganism. In this context, novel antimicrobials that might be active against A. baumannii are urgently needed. The application of lytic bacteriophages is a potential approach allowing the solution to this problem.
The use of bacteriophages has been a success in treatments of some nosocomial bacterial infections, caused for example by Pseudomonas aeruginosa and Staphylococcus aureus (Merabishvili et al., 2009; Kutter et al., 2010). However, there are no bacteriophage preparations to control A. baumannii infections because of the absence of abundant phage collections to design therapeutics and narrow host range of available lytic phages.
Recently, several lytic bacteriophages infecting A. baumannii clinical strains have been characterized. The phage AB1 was isolated from a marine sediment sample and was lytic for one of five tested A. baumannii strains only. The phage was classified by authors as a member of the Siphoviridae family (Yang et al., 2010). In another work (Lin et al., 2010), phage φAB2 lytic for 25 of 125 multidrug-resistant (MDR) A. baumannii strains was isolated from hospital sewage water and characterized. The phage was attributed to the Podoviridae family. The lytic myophage Abp53 lysed 27% of the A. baumannii isolates tested was characterized in 2011 (Lee et al., 2011).
The purpose of our investigation was to isolate wide host range bacteriophages lytic for A. baumannii and study their biological properties. In the research, newly isolated Myoviridae lytic phage AP22 was characterized. The bacteriophage infected specifically and lysed 89 of 130 tested MDR clinically relevant A. baumannii strains obtained from hospitalized patients from several clinics of the Russian Federation.
Materials and methods
Bacterial strains and their identification
MDR A. baumannii strains were isolated from clinical materials (wounds, tissue samples, sputum, bronchopulmonary lavage, pleural fluid, urine, bile, blood, and rinses of drainage and intravenous catheters) obtained from hospitalized patients of different clinics of the Russian Federation (Chelyabinsk, Moscow, Nizhni Novgorod, St. Petersburg) in 2005–2010. They were identified by amplified 16S rRNA gene restriction analysis using primers SP2-16S (5′-GATCATGGCTCAGATTGAACGC-3′) and ASP2-16S (5′-GCTACCTTGTTACGACTTCACCC-3′), and AluI restriction endonuclease. RFLP profiles were compared with those of A. baumannii 16S rRNA genes, whose nucleotide sequences were deposited in GenBank (accession numbers CP000863.1, CP000521.1, CP001172.1, CU459141.1, and CP001182.1).
Random amplification of polymorphic DNA (RAPD) analysis was subsequently used to discriminate the A. baumannii strains. Primers Wil2 (Williams et al., 1993) and 1247 (Akopyanz et al., 1992) previously used for typing other bacteria were applied.
Some other representatives of the genus of Acinetobacter such as A. lwoffii (six strains), A. anitratus (4), and A. calcoaceticus (3) and several other gram-negative microorganisms such as P. aeruginosa, Escherichia coli, Yersinia pseudotuberculosis, Yersinia enterocolitica, Klebsiella pneumoniae, Klebsiella oxytoca, Enterobacter cloacae, Pasteurella multocida, and Salmonella Enteritidis (three strains of each species) were used in the research.
All bacteria were grown in Luria–Bertani (LB) broth or nutrient agar (Himedia Laboratories Pvt. Limited, India) at 37 °C.
Phage isolation, propagation, and purification
Clinical materials and in-hospital environmental samples were used for phage isolation. Nonliquid samples were kept in 0.1 M Tris–HCl buffer, pH 7.0. The samples were cleared by low-speed centrifugation (7000 g for 30 min.) followed by filtration of the supernatants through 1.20- and 0.45-μm-pore-size membrane filters (Millipore) to remove bacterial debris. The purified filtrates were concentrated by ultracentrifugation at 85 000 g at 4 °C for 2 h (Beckman SW28 rotor). The spot test method as well as the plaque assay (Adams, 1959) was used to screen for the presence of lytic phage activity in the resultant concentrates using clinical A. baumannii strains of different RAPD groups. The plates were incubated overnight at 37 °C and examined for zones of lysis or plaques formation. Single plaque isolation was used to obtain pure phage stock. For that a single plaque formed on the A. baumannii lawn was picked up in SM buffer (10 mM Tris–HCl, pH 7.5, 10 mM MgSO4 × 7 H2O, and 100 mM NaCl) and replated three times.
Phage AP22 was propagated using liquid culture of identified A. baumannii clinical strain 1053 (OD600 nm of 0.3) at multiplicity of infection (MOI) of 0.1. The incubation was performed at 37 °C until complete lysis, and then chloroform was added. Bacterial debris was pelleted by centrifugation at 7000 g for 30 min. The phage lysate was concentrated by ultracentrifugation at 85 000 g at 4 °C for 2 h (Beckman SW28 rotor). The resultant pellet was carefully mixed with SM buffer and centrifuged at 13 000 g. Supernatant was treated with DNase (1 μg mL−1) and RNase (1 μg mL−1) at 37 °C. The nucleases were removed with chloroform. The phage preparation with the titer of 1012–1013 PFU mL−1 was purified by cesium chloride equilibrium gradient centrifugation at 100 000 g (Beckman SW41 rotor) for 24 h (Sambrook et al., 1989).
Phage host range determination
Host specificity of the phage was determined by double-layer method. Onto the surface of M9 medium (Sambrook et al., 1989) plates, 0.3 mL of liquid bacterial culture (108–109 PFU mL−1) and 4 mL of soft agar (LB broth supplemented with 0.6% agarose and calcium chloride, 400 μg mL−1) at 45 °C were poured and allowed to solidify. Then, the phage suspension and its several dilutions were spotted on the soft agar lawns and incubated at 37 °C for 18–24 h.
Examination of phage-resistant clones
Fifty phage-resistant clones were picked from the lysis zone formed by the phage on A. baumannii lawns from seven different plates. The clones were subjected to three cycles of purification, resuspended in a saline solution, treated with chloroform, and centrifuged. Supernatants were spotted on the phage-sensitive A. baumannii lawn. Also each resistant clone was grown in 30 mL LB broth in the presence of mitomycin C (0.3–1 μg mL−1). The samples were cleared by low-speed centrifugation (7000 g for 30 min.), and supernatants were concentrated 100–1000 times by ultracentrifugation at 4 °C for 2 h (85 000 g; Beckman SW28 rotor). The presence or absence of the phage was estimated by electron microscopy.
A putative prophage in the genomic DNA of the resistant clones was looked for using multiplex PCR with two pairs of primers specific to phage AP22 DNA, developed on the base of partial sequence of the phage genome. These were AP22A-f (5′-AGTTCGTTCTGCTGTTTGG-3′) and AP22A-r (5′-TCCTCAACATACCAAATCG-3′); AP22B-f (5′-GTGTTCATTTCGTTCTCTCA-3′) and AP22B-r (5′-CGACATTTCTCAACATCAGC-3′). As control of the PCR, primers for the gene 16S rRNA gene of A. baumannii were used.
Exponentially grown A. baumannii cells were mixed with the phage (MOI = 0.001) and incubated at room temperature. A volume of 100 μL of samples was taken in 1, 2, 3, 4, 5, 10, 15, and 20 min and mixed then with 850 μL of SM buffer supplemented with 50 μL of chloroform. After centrifugation, the supernatants were titrated for further determination of unadsorbed phages by the double-layer method at different time intervals. The adsorption constant was calculated according to the study by Adams (1959) for a period of 5 min.
A volume of 20 mL of host bacterial cells (OD600 nm of 0.3) was harvested by centrifugation (7000 g, 5 min, 4 °C) and resuspended in 0.5 mL LB broth. Bacterial cells were infected with the phage at MOI of 0.01. The bacteriophage was allowed to adsorb for 5 min at 37 °C. Then, the mixture was centrifuged at 13 000 g for 1 min to remove unadsorbed phage particles, and the pellet was resuspended in 10 mL of LB broth. Samples were taken at 5-min intervals during incubation at 37 °C within 2 h and immediately titrated. The procedure was repeated three times. Latent period was defined as the interval between adsorption of the phage to the host cell and release of phage progeny. The burst size of the phage (the number of progeny phage particles produced by a single host cell) was expressed as the ratio of the final count of released phage particles to the number of infected bacterial cells during latent period.
Phage infectivity at different pH values
The bacteriophage (108 PFU mL−1) was incubated in 1 mL of pH buffers at pH 2, 4, 7, 9, and 12 at room temperature. Samples were taken in 1, 3, 6, and 24 h and titrated using the double-overlay method.
Isolation and restriction analysis of phage genomic DNA
CsCl-purified bacteriophage particles were incubated in 0.5% SDS, 20 mM EDTA, and 50 μg mL−1 proteinase К at 56 °C for 3 h. The DNA was extracted with phenol/chloroform (1 : 1) and then precipitated with ethanol supplemented with sodium acetate. Restriction fragment length polymorphism (RFLP) analysis of the phage DNA was performed using endonucleases AluI, ApaI, BamHI, BglII, CfrI, ClaI, DraI, DraII, Eco52I, EcoR91I, EcoRI, EcoRV, HindIII, HinfI, MspI, NcoI, NheI, NotI, PstI, PvuII, RsaI, SalI, SmaI, SmiI, SspI, TaqI, VspI, and XmiI (Fermentas, Lithuania). The procedure was carried out according to instructions provided by the manufacturer. DNA fragments were separated by 0.8% and 1.5% agarose gel electrophoresis in TBE electrophoresis buffer. The approximate molecular sizes of separated DNA fragments were calculated using Quantity One software (Bio-Rad). One-kb DNA Ladder (Fermentas) and phage lambda DNA digested with HindIII were used as molecular markers.
Phage protein profiles
Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis was performed according to the Laemmli's standard protocol (Laemmli, 1970) using the CsCl-purified phage preparation.
The phage was examined by negative contrast electron microscopy (EM; Brenner & Horne, 1959). The purified and concentrated virus preparation was fixed with 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) and then placed on transmission EM support grids followed by rinsing with distilled water several times. The phage samples were stained with 1% uranyl acetate aqua solution for further examination with a Hitachi H-300TM electron microscope (Japan). Electron microscope magnification was calibrated using T4 phage as size standard. At least 20 electronic phage images were used for phage morphology determination.
To obtain electron microphotographs of the phage–host cell interaction, the mixture of A. baumannii 1053 cells (107 CFU mL−1) with the phage AP22 (108 PFU mL−1) was incubated in 0.05 M phosphate buffer (pH 7.0) for 10 min at room temperature. The phage–cell mixture was fixed with 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) and analyzed as described earlier.
Results and discussion
The problem of the search for potentially therapeutic A. baumannii phages and their characterization has recently attracted considerable attention because of the increasing interest of the microorganism itself as a threatening causative agent of nosocomial infections worldwide.
Phage AP22 was isolated from clinical materials provided by the burn center of N.V. Sklifosovsky Scientific Research Institute of First Aid (Moscow). A phage was isolated from the burn center in a month and found to be identical in host range and RFLP assays with AP22. It can be assumed that the bacterial virus is stable in the hospital environments.
The phage AP22 formed clear, round plaques of 2–3 mm in diameter with haloes (zone of clearance around each plaque) on A. baumannii-sensitive strains. The plaques on agar plates appeared after 4 h of the incubation at 37 °C. The size of haloes ranged from some millimeters to some centimeters in diameter and tended to expand with time. The haloes surrounding the plaque centers can indicate the presence of phage depolymerase capable of degrading exopolysaccharide (EPS) secreted by A. baumannii cells (Marti et al., 2011). Numerous phages inducing enzymes capable of depolymerizing the gram-negative bacterial EPS are characterized by plaques with size-varying haloes (Sutherland et al., 2004).
In liquid culture, the titer of the phage lysate was close to 1010 PFU mL−1.
Morphological characterization of the phage AP22 using EM is shown in Fig. 1. The phage has an icosahedral head of 63–65 nm in diameter and a contractile tail of 85–90 nm in length. Thus, the bacterial virus was classified as a representative of the Myoviridae family. It has a 22- to 23-nm base plate with tail fibers, each 42–43 nm long. The fibers are clearly visible after tail contraction (Fig. 1c). The phage was morphologically similar to earlier isolated phage BS46 (Soothill, 1992; Ackermann et al., 1994).
The phage genome is presented by double-stranded DNA that is digested with restriction endonucleases HindIII, DraI, VspI, SspI, TaqI, AluI, RsaI, HinfI, MspI, CfrI, and EcoRI. It is partially digested with EcoRV, PstI, SalI, XmiI, SmiI, ClaI, BamHI, PvuII, BglII, EcoR91I, NcoI, and NheI. Numerous restriction fragments are formed by digestion of the phage DNA with endonucleases that recognize hexanucleotide palindromic sequences containing only A+T base pairs such as DraI, SspI, and VspI (Fig. 2). On the other hand, enzymes recognizing G+C–rich sequences (ApaI, SmaI, NotI, Eco52I) do not digest phage AP22 DNA (data not shown). It is quite possible that the phage genome has low G+C content. The phage genome size was estimated as 46 kb (Fig. 2).
Purified phage particles were subjected to SDS-PAGE for the detection of the number of structural phage proteins. Four major protein bands and three minor protein bands were detected, with molecular weights ranging from approximately 18–87 kDa (Fig. 3). The most predominant polypeptide band of approximately 32 kDa was presumably corresponding to a major capsid protein.
The phage infection process was investigated by the estimation of the AP22 adsorption efficiency to the host and one-step growth of the phage. As shown in Fig. 4a, the phage adsorption occurred rapidly; more than 99% of the phage adsorbed within 5 min. Adsorption constant of AP22 was 1.53 × 10−7 mL min−1.
One-step growth experiment was completed to determine the latent period and the phage burst size (Fig. 4b). The latent period of AP22 was 40 min, followed by the rise period (increasing in the concentration of phage particles) of 40 min, and plateau phase. The burst size was approximately 240 particles per one infected cell.
The phage stability was investigated at different pH levels. The phage remained stable within 24 h between pH 4 and 9. But only 0.04% phages were infectious at pH 2, and 100% phage particles were inactivated completely at pH 12 in 1-h incubation (data not shown).
Acinetobacter baumannii clones resistant to phage AP22 were formed at the rate of 10−6 per a cell. A total of 50 phage-resistant clones of A. baumannii 1053 were analyzed to determine whether they are phage-resistant mutants or lysogens with inserted prophage. To reveal possible spontaneous induction, bacterial suspensions of each clone treated with chloroform were spotted on bacterial lawn of sensitive strain. Besides, the resistant clones were grown in the presence of different concentrations of mitomycin C to show possible presence of the phage in concentrated preparation by EM procedure. In both cases, there was no presence of the phage in the samples. A possibility of the prophage presence in genomic DNA of resistant clones was estimated by PCR with two pairs of primers specific to the phage DNA. It was shown the absence of prophage DNA in genomic DNA of resistant clones (Fig. 5).
Lytic activity and host specificity of the phage were tested against 130 identified A. baumannii genotype-varying MDR strains. These strains were isolated from patients of burn units, units of selective and emergency surgery, therapeutics units, intensive care units, and urology units in 2005–2010. Most of them were resistant to diverse groups of antibiotics, including aminoglycosides, fluoroquinolones, third- and forth-generation cephalosporins, and also cefoperazone sulbactam and carbapenems. All strains were divided into 10 groups by RAPD analysis. RAPD groups A1 and B1 predominated with 48% and 35% of the investigated strains, respectively, and were spread in clinics of a variety of Russian cities.
Unlike some other known A. baumannii phages, bacteriophage AP22 was found to have a broad range of lytic activity against A. baumannii multidrug-resistant clinically relevant strains. The phage was shown to specifically infect and lyse 68% (89 of 130) of A. baumannii strains by forming clear zones. Of particular interest is that the phage lysed 83% (88 of 106) of A. baumannii strains from those two RAPD groups that were dominating in some Russian hospitals between 2005 and 2010 (Table 1).
Table 1. Phage AP22 lytic activity toward Acinetobacter baumannii strains of different RAPD groups
The number of strains
Period of isolation (years)
Place of isolation
Phage-susceptible strains (%)
Wound, tissue sampling, sputum,
bronchopulmonary lavage, pleural
fluid, urine, bile, blood, and
hospital environmental rinses
Wound, tissue sampling, sputum,
bronchopulmonary lavage, pleural
fluid, urine, bile, blood, rinses of
drainage and intravenous catheters,
and hospital environmental rinses
Moscow, St. Petersburg
Moscow, St. Petersburg
Wound, sputum, and rinses of
The phage was also tested against some other representatives of the genus Acinetobacter (A. lwoffii, A. anitratus, and A. calcoaceticus), as well as several other gram-negative microorganisms such as P. aeruginosa, E. coli, Y. pseudotuberculosis, Y. enterocolitica, K. pneumoniae, K. oxytoca, E. cloacae, P. multocida, and S. Enteritidis. All of these bacteria were found to be insensitive to the phage.
In conclusion, the lytic bacteriophage AP22 belonging to the Myoviridae family specific for A. baumannii was isolated and characterized. Bacteriophage AP22 exhibited rapid adsorption (> 99% adsorbed in 5 min), a large burst size (240 PFU per cell), stability to the wide range of pH, and lytic activity toward a broad range of A. baumannii strains. Thus, phage AP22 should be considered as a candidate for inclusion in phage cocktails to control A. baumannii-associated nosocomial infections.
We are grateful to Drs Margarita Popova (City Clinical Hospital №6, Chelyabinsk), Tamara Spiridonova (N.V. Sklifosovsky Scientific Research Institute of First Aid, Moscow), Natalia Gordinskaya (Nizhny Novgorod Research Institute of Traumatology and Orthopedics of Public Health), Artemy Goncharov (The Saint Petersburg State Medical Academy), and Nadezhda Fursova for providing A. baumannii isolates and clinical samples for the research. This study was supported by the Federal Service for Supervision of Consumer Rights Protection and Human Welfare (scientific program number 01201172662).