Characterization of a novel Achromobacter xylosoxidans specific siphoviruse: phiAxp-1

Bacteriophages have recently been considered as an alternative biocontrol tool because of the widespread occurrence of antimicrobial-resistant Achromobacter xylosoxidans. Herein, we isolated a virulent bacteriophage (phiAxp-1) from a water sample of the Bohai sea of China that specifically infects A. xylosoxidans. Transmission electron microscopy revealed that phage phiAxp-1 belongs to the Siphoviridae. We sequenced the genome of phiAxp-1, which comprises 45,045 bp with 64 open reading frames. Most of the proteins encoded by phiAxp-1 have no similarity to sequences in the public databases. Twenty-one proteins with assigned functions share weak homology with those of other dsDNA bacteriophages infecting diverse hosts, such as Burkholderia phage KL1, Pseudomonas phage 73, Pseudomonas phage vB_Pae-Kakheti25, Pseudomonas phage vB_PaeS_SCH_Ab26, Acinetobacter phage IME_AB3 and Achromobacter phage JWX. The genome can be divided into different clusters for the head and tail structure, DNA replication and mazG. The sequence and genomic organization of bacteriophage phiAxp-1 are clearly distinct from other known Siphoviridae phages; therefore, we propose that it is a member of a novel genus of the Siphoviridae family. Furthermore, one-step growth curve and stability studies of the phage were performed, and the specific receptor of phiAxp-1 was identified as the lipopolysaccharide of A. xylosoxidans.

(including capsid and tail morphogenesis), DNA packaging, DNA replication, and MazG (a pyrophosphohydrolase 11 ). Putative functional assignments and significant similarities to other sequences are listed in Table 2.
The proteins with assigned functions share weak homology with proteins found in other dsDNA bacteriophages, infecting diverse hosts, such as Burkholderia phage KL1, Pseudomonas phage 73, Pseudomonas phage vB_Pae-Kakheti25, Pseudomonas phage vB_PaeS_SCH_Ab26, Acinetobacter phage IME_AB3 and Achromobacter phage JWX, with percent identities of 24-70%. Multiple genome alignments showed the weak homology of phiAxp-1 with these phages at the whole genome level (Fig. 3). In addition, 18 virion proteins were detected using liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS), of which 12 have been assigned putative functions (Table 3). Of the 21 phiAxp-1 proteins with a suggested function, half of them are structural and morphogenesis proteins. Based on BLASTP analysis, the capsid morphogenesis and DNA packaging module contains genes encoding the large terminase subunit, capsid and tail proteins (Fig. 2). As a phage with Siphoviral morphotypes, the genes for head and tail assembly are arranged together, with the head genes 5′ to the tail genes 12 . Terminase genes are involved in initiation of DNA synthesis and are responsible for the packaging of the concatameric DNA in phage capsids 13 . gp40 (terminase large subunit) shares similarity with the terminase large subunit protein from KL1 (70% identity), vB_PaeS_SCH_Ab26 (66% identity) and 73 (67% identity). Phylogenetic analysis of the available bacteriophage terminase proteins also demonstrated that phiAxp-1 could not be assigned to a branch (Fig. 4). Portal proteins are responsible for forming a ring that enables DNA to pass into the major capsid during assembly and out during infection, serving as a junction between the capsid and tail proteins 13,14 . gp41 encodes a protein similar to the portal proteins found in the genomes of vB_Pae-Kakheti25 (48% identity), vB_PaeS_SCH_ Ab26 (48% identity) and 73 (48% identity). gp42 (head morphogenesis protein) shares weaker identity with the head morphogenesis protein from KL1 (28% identity). gp43 (scaffold protein) is similar to vB_PaeS_SCH_Ab26 orf9 (46% identity), and gp44 (major capsid protein) is similar to the major capsid protein of IME_AB3 (54% identity). The phiAxp-1 genome encodes six putative tail proteins, including a minor tail protein (gp48), a major tail tube protein (gp49), two tail chaperonin proteins (gp50-51), a tail tape measure protein (gp53) and a tail component protein (gp59) ( Table 2). The tail tape measure protein (1044 amino acids [aa], encoded by gp53) is similar to the predicted KL1 tape measure protein orf21 (41% identity). The tape measure protein is important for the assembly of phage tails and is involved in tail length determination 15,16 . Importantly, the largest protein encoded by phiAxp-1 is the tail component protein (gp59, 1331 aa) rather than the tail tape measure protein, which is commonly the largest protein of siphovirus 12 . In addition, gp57-59 are similar to JWX orf25-27 (42-50% identity).
The mazG gene provides a significant selective advantage to phages 17 . The putative MazG protein is encoded  Stability studies. Stability studies of phage phiAxp-1 were conducted with different pHs, disinfectants, temperatures and ions, using a temperature-controlled incubator or water baths. The results are summarized in Fig. 5. The phage was most stable at pH 7, there was a significant reduction in the phage titre either above or below pH 7; the phage titre was further decreased under extremely acidic (pH 4) or basic (pH 12) conditions (Fig. 5a). No significant loss of phage titre was observed from 4 to 37 °C. However, the phage titre dramatically decreased when the temperature is over 50 °C (Fig. 5b). The activities of phage phiAxp-1 were affected in the presence of ethanol (Fig. 5c), the phage was resistant to isopropanol ( Fig. 5d) at low concentrations (10%, v/v), whereas it became unstable with increasing concentrations: there was a significant reduction in phage titre at high concentration (95%, v/v). Many phages require divalent ions such as Ca 2+ or Mg 2+ for attachment or intracellular growth 18 . It may be necessary to treat phages with Ca 2+ or Mg 2+ to obtain an efficient phage infection. The effects of divalent ions on phage amplification were evaluated and the phage revealed divalent cation dependency for optimal infectivity. Divalent cations at no more than 20 mM were beneficial for plaque development (Fig. 5e).

Identification of the phage receptor.
In Gram-negative bacteria, outer membrane proteins or lipopolysaccharide (LPS) may function as specific phage receptors 19 . Therefore, it was necessary to test whether the degradation of cell surface proteins or LPS could inhibit phiAxp-1 binding 20 . A. xylosoxidans cells were treated with either proteinase K (to destroy surface proteins) or periodate (to destroy surface carbohydrates) before the phage adsorption assay to determine the possible nature of the phage receptor 20 . phiAxp-1 exhibited high infection efficiency when mixed with untreated and proteinase K-treated A. xylosoxidans cells (Fig. 6a): the majority of the phages were removed from the suspension after centrifugation by binding to A. xylosoxidans cells. This suggested that the functional receptor is not a protein. The broad substrate specificity of proteinase K meant that the possibility that the receptor is a protein resistant to proteinase K is unlikely 19 . When the phage was incubated with periodate-treated A. xylosoxidans, the majority of the phages remained in the supernatant (Fig. 6b). The significant increase of free phage particles suggested that the phages were unable to efficiently adsorb onto the periodate-treated bacteria. Therefore, the A. xylosoxidans receptor recognized and bound by phiAxp-1 is a carbohydrate structure, most likely LPS. Significant inactivation of phages was further confirmed using LPS purified from A. xylosoxidans, which demonstrated LPS is the adsorption target (receptor) of this phage (Fig. 7). The results revealed direct correlation between A. xylosoxidans LPS concentration and phage infectivity inhibition, and 12.5 μ g/ml of LPS was sufficient to inhibit the binding activity of 50% of 4.7 × 10 4 pfu phiAxp-1. LPS  of Escherichia coli 0111:B4 was used as a negative control. As shown in Fig. 7, E. coli LPS showed no phage inactivating ability compared with that from A. xylosoxidans LPS, thus indicating that LPS from A. xylosoxidans is specific for phage phiAxp-1. In this respect, it is consistent with the features of most phages with Gram-negative bacterial hosts.

Concluding Remarks
The clinical relevance of nosocomially-acquired infections caused by multi-resistant Achromobacter strains is increasing rapidly, becoming a critical problem 7 . Phages are re-emerging as promising potential therapies for the treatment of bacterial infections 21 . Here, we report a preliminary analysis of A. xylosoxidans bacteriophage phiAxp-1. This article presents the sequence analysis and a detailed genome annotation of phage phiAxp-1. The genomic data constitute an important resource to study and engineer phages to control specific bacterial species 22 . The analysis showed that phiAxp-1 does not easily fit into previously established groups of dsDNA bacterial viruses and may represent a distinct branch of the Siphoviridae family. Stability is the primary requirement for any possible commercial use of the phage, which can reduce the cost of storage significantly 23 . Therefore, in this study, the stability tests on phage phiAxp-1 under different pHs, temperatures, disinfectants and ions were performed for the potential practical application of phiAxp-1. Despite its importance, the molecular interactions between phiAxp-1 and the surface of A. xylosoxidans are still poorly understood. Phages bind to unique host-specific structures, allowing them to recognize a suitable host in a mixed bacterial population 24 . In this study, periodate treatment of A. xylosoxidans, but not proteinase K treatment, inhibited phage binding. Furthermore, purified LPS from the A. xylosoxidans showed phage-inactivating capacities thus confirmed that LPS of A. xylosoxidans is the receptor of phage phiAxp-1.
The emergence of phage-resistant mutants affecting phage receptors is a major concern regarding the use of phage therapy 25 . The LPS of Gram-negative bacteria commonly represents an important virulence factor and is of great significance in the pathophysiology of many disease processes 26 . Thus, the phage-resistant mutants resulting from the loss or alteration of the receptor will be avirulent or attenuated. Such mutants do not pose a problem during bacteriophage treatment 25 . Our future work will explore this possibility. These results suggest that phage phiAxp-1 is a promising candidate for controlling A. xylosoxidans and represents an advance in our current knowledge of A. xylosoxidans phages.

Methods
Bacterial strains and growth media. Luria-Bertani (LB) broth medium was used to grow the bacterial strains and to propagate the phage. A. xylosoxidans strain A22732 was used as the indicator strain for phage isolation.
Isolation of phage and host range determination. phiAxp-1 was isolated from a water sample of the Bohai sea of China using a double agar overlay plaque assay, as described previously for the isolation of lytic phages 27 . The water sample was centrifuged at 8,000 × g for 10 min to remove the solid impurities. The supernatants were filtered through a 0.22-μ m pore-size membrane filter to remove bacterial debris. The filtrates were then mixed with A. xylosoxidans culture to enrich the phage at 37 °C. The culture was centrifuged, and the supernatant was filtered through a 0.22-μ m pore-size membrane to remove the residual bacterial cells. Aliquots of the diluted filtrate were mixed with A. xylosoxidans culture. Then, 3 mL of molten top soft nutrient agar (0.7% agar) were overlaid on the solidified base nutrient agar (1.5% agar) 28 . Following incubation for 10 h at 37 °C, clear phage plaques were picked from the plate. The phage titre was determined using the double-layered method. The host range of the phage was tested against 57 clinical strains from our microorganism centre, as determined by standard spot tests 29 . Briefly, 10 μ l from a purified phage suspension containing approximately 10 8 pfu/mL were spotted in the middle of a lawn of bacteria and left to dry before incubation overnight. Bacterial sensitivity to a bacteriophage was established by bacterial lysis at the spot where the phage was deposited. Each strain was tested three times at 37 °C.

TEM.
To prepare phiAxp-1 for transmission electron microscopy studies, cell debris from 500 mL of A.
xylosoxidans strain A22732 infected with phiAxp-1 was pelleted by centrifugation. Phage particles were precipitated with 1 M NaCl and 10% polyethylene glycol (PEG) 8000 at 4 °C with stirring for 60 min. The precipitated phage particles were harvested. Phage particles were resuspended in Saline -magnesium (SM) diluent plus gelatin (SMG) (50 mM Tris-HCl [pH 7.5] containing 100 mM NaCl, 8.1 mM MgSO 4 and 0.01% (w/v) gelatin) and extracted with an equal volume of chloroform. After low-speed centrifugation, the aqueous phase was sedimented at about 25,000 × g for 60 min. Phage particles were negatively stained with 2% (wt/vol) phosphotungstic acid (pH 7), air dried, and examined under a Philips EM 300 electron microscope operated at 80 kV and 120 KEv.
One-step growth curve. One-step growth experiments were performed as described previously 30 . Host strain A. xylosoxidans strain A22732 cells were harvested at exponential growth and resuspended in LB. The phage phiAxp-1 was added at a multiplicity of infection (MOI) of 0.0005 and allowed to adsorb for 5 min at room temperature. The mixture was centrifuged and the pellets containing infected cells were suspended in 10 ml of LB,   followed by incubation at 37 °C. Samples were taken at 10 min intervals (up to 110 min) and immediately diluted, and then titres were determined by the double-layered-agar plate method.  . The control (LB and "A22732 + acetate"), untreated strain (A22732), and treatment ("A22732 + ProtK" for proteinase K treatment and "A22732 + IO 4− " for periodate treatment) groups were tested for adsorption, as indicated in the x-axes. Error bars denote statistical variations. Significance was determined by a Student's t test for comparison between the treated and the untreated groups. *P 0.05. Isolation of phage DNA, genome sequencing and assembly. phiAxp-1 DNA was extracted from purified phage particles with phenol-chloroform (24:1, vol/vol) and precipitated with 100% ethanol. The samples were visualized on 0.7-1.0% agarose gels, and the purified phage DNA was sequenced using an Illumina HiSeq2500 sequencer. The sequence reads were filtered to remove low quality sequences, trimmed to remove adaptor sequences and the filtered sequences were assembled. The final assembled sequence was searched against the current protein and nucleotide databases (http://www.ncbi.nlm.nih.gov/) using Basic Local Alignment Search Tool (BLAST) software 31 . BLASTP was used to determine the similarity to described proteins in the National Center for Biotechnology Information [NCBI] database (http://www.ncbi.nlm.nih.gov). The CLC Main Workbench, version 6.1.1 (CLC bio, Aarhus, Denmark) was used for genome annotation. Computer-based predictions were checked manually. Phylogenetic analysis with the published genome sequences of related phages was conducted using ClustalW (Slow/Accurate, IUB). Whole genome comparisons were carried out using Mauve 32 . LC/ESI/MS/MS spectra (Q-TOF Ultima API, Micromass UK Ltd.) were used to identify the phage proteins, as described previously 33 .
Identification of the phage receptor. Receptor properties of phiAxp-1 were determined as described previously 19 . Briefly, A. xylosoxidans A22732 cultures were treated with sodium acetate (50 mM, pH 5.2) containing 100 mM IO 4− at room temperature for 2 h (protected from light) or proteinase K (0.2 mg/ml; Promega) at 37 °C for 3 h to determine whether proteinase K or periodate can destroy the phage receptor. The phage adsorption assay was then performed as previously described 20 . LB was used as a non-adsorbing control in each assay, and the phage titre in the control supernatant was set to 100%. Each assay was performed in duplicate and repeated twice 19 .
Phage inactivation by LPS. LPS extraction from A. xylosoxidans was performed using an LPS extraction kit from Intron Biotechnology (17144; Boca Scientific, Boca Raton, FL, USA), according to the manufacturer's instructions. LPS from Escherichia coli O111:B4 purchased from Sigma-Aldrich, Inc. (L2630; Sigma, USA) was used as a negative control to ensure that the possible effect was specific to A. xylosoxidans LPS. Both LPS of Escherichia coli O111:B4 and A. xylosoxidans A22732 are smooth type. The phage inactivation by LPS was performed as previously described 34 .
Nucleotide sequence accession number. The annotated genome sequence for the phage phiAxp-1 was deposited in the NCBI nucleotide database under the accession number KP313532.