Specific Interaction of Novel Friunavirus Phages Encoding Tailspike Depolymerases with Corresponding Acinetobacter baumannii Capsular Types

Acinetobacter baumannii, a nonfermentative, Gram-negative, aerobic bacterium, is one of the most significant nosocomial pathogens. The pathogenicity of A. baumannii is based on the cooperative action of many factors, one of them being the production of capsular polysaccharides (CPSs) that surround bacterial cells with a thick protective layer.

IMPORTANCE Acinetobacter baumannii, a nonfermentative, Gram-negative, aerobic bacterium, is one of the most significant nosocomial pathogens. The pathogenicity of A. baumannii is based on the cooperative action of many factors, one of them being the production of capsular polysaccharides (CPSs) that surround bacterial cells with a thick protective layer. Polymorphism of the chromosomal capsule loci is responsible for the observed high structural diversity of the CPSs. In this study, we describe eight novel lytic phages which have different tailspike depolymerases (TSDs) determining the interaction of the viruses with corresponding A. baumannii capsular types (K types). Moreover, we elucidate the structures of oligosaccharide products obtained by cleavage of the CPSs by the recombinant depolymerases. We believe that as the TSDs determine phage specificity, the diversity of their

RESULTS
Phage isolation, morphology, and host specificity. Bacteriophages vB_AbaP_ APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 were initially isolated on the bacterial lawns of A. baumannii ACICU (K2), LUH5549 (K32), NIPH146 (K37), NIPH70 (K44), NIPH615 (K48), LUH5547 (K87), LUH5552 (K89), and MAR303 (K116) from sewage and environmental (river water) samples by using an enrichment procedure in 2018. The phages were named according to a rational scheme for the nomenclature of viruses of Bacteria and Archaea (14), where APK is a short laboratory name which means Acinetobacter phage and the number of the K type that is infected by the phage, for example, APK2 indicates the phage isolated on the bacterial lawn of A. baumannii ACICU belonging to K2.
As examined by transmission electron microscopy (TEM), all the phages had icosahedral heads of approximately 60 nm in diameter and short noncontractile tails of 10 nm in length (Fig. 2).
The host specificity of phages APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 was tested using a collection of A. baumannii strains belonging to 56 different K types (Table 1). K loci identified in the genomes of these strains were annotated earlier, and CPS structures for most of them were biochemically characterized. It was found that all of the phages except for APK2 were highly specific and able to infect only a strain of a certain K type. Phage APK2 could infect the representatives of two capsular types, namely, A. baumannii ACICU (K2) and A. baumannii B11911 (K93) (Fig.  1B). In return, phage APK93 isolated from another sample of wastewater on the bacterial lawn of A. baumannii B11911 (K93) was able to infect A. baumannii ACICU (K2). Based on this observation, we suggest that structural depolymerases of these phages recognize and degrade a similar linkage in the CPS structures of A. baumannii strains ACICU and B11911.
Phage genome organization and comparison. The phage linear genomes ranged from 41,105 to 42,402 bp in size, with direct terminal repeats (DTRs) of 381 to 417 bp at the genomes' ends, containing between 49 and 54 predicted genes located only on the forward strands ( Table 2). The G1C content of the genomes was 39.05% to 39.40%, similar to that of other A. baumannii viruses and close to the approximate average values for different A. baumannii strains (38.94% to 39.4% according to reference 16). No tRNA genes were identified. No genes encoding toxins and no products responsible for antibiotic resistance or related to lysogeny were determined in the phage genomes.
Almost all of the predicted proteins encoded by APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 had very close homologues in the genomes of phages belonging to the genus Friunavirus of the subfamily Beijerinckvirinae (recently designated by the International Committee on Taxonomy of Viruses [ICTV]), within the family Autographiviridae.
Based on the functions of homologous proteins, early, middle, and late gene regions were identified in APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 genomes, as well as in the genome of phage Fri1 and other Fri1-like viruses, such as vB_AbaP_AS11 and vB_AbaP_AS12 (7).
The functions of products encoded by genes of early regions (left parts of the genomes) were not determined, but it is likely that these proteins are involved in inhibiting or redirecting the activities of functionally important host system components to serve the needs of the viruses (17).
The middle regions of the phage genomes comprised nucleotide metabolism and DNA replication and repair genes encoding DNA primase, DNA helicase, ATP-dependent ligase, DNA polymerase I, 59-39 exonuclease, tRNA nucleotidyltransferase, endonuclease VII, phosphoesterase, deoxynucleotide monophosphate kinase, and also the gene encoding single-subunit viral RNA polymerase. Interestingly, genes encoding tRNA nucleotidyltransferases were absent in the genomes jf phages APK32, APK37, and APK48 and thus, most likely, are not essential for phage nucleotide metabolism. The presence of open reading frames (ORFs) encoding putative HNH endonucleases was also revealed in the middle clusters of all studied phages, except for phages APK89 and APK116. Some homing endonuclease genes were found immediately upstream (APK87_g25) or downstream (APK2_g22, APK37_g23, APK44_g23, APK48_g21) of the DNA polymerase I gene. In some cases, HNH endonuclease genes (APK32_g24, APK44_g21) interrupted the DNA polymerase gene into the two domains, and in some cases, they were located between genes encoding 59-39 exonuclease and DNA endonuclease VII (APK32_g29, APK48_g26) instead of genes encoding tRNA nucleotidyltransferase.
Late genome regions were highly conserved among all Fri1-like viruses, containing genes encoding structural proteins (head-tail connector protein, scaffolding protein, major capsid protein, tail tubular proteins A and B, internal virion proteins A, B, and C, and tailspike), proteins associated with bacterial cell lysis (holin and endolysin), and with proteins associated with the packaging of DNA (DNA maturase A and B) (Fig. 3).
Tailspikes of phages APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 were formed by single proteins encoded by the genes located at the end of structural modules of phage genomes immediately after internal virion proteins A to C ( Fig. 3; Table 2).
At the amino acid level, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 TSDs were found to differ significantly from the proteins encoded by the other A. baumannii phages deposited in GenBank, except for their N-terminal domains (approximately the first 160 amino acids of the proteins). These domains were responsible for the attachment of variable CPS-recognizing/degrading parts of the tailspikes to the tail tubular structures of phage particles and, as expected, were very conservative within the phages of the same taxonomic group sharing similar morphology and structural components.
BLASTp analysis revealed that APK2_gp43 was identical to the proteins encoded by phages vB_AbaP_APK2-2 (APK2-2_gp43; accession no. AZU99292) and vB_AbaP_ APK93 (APK2_gp43; GenBank accession no. AZU99342) and was highly homologous to the proteins encoded by Acinetobacter phage IME200 (YP_009216489; 99% at the amino acid level) and Acinetobacter phage SH-Ab 15519 (YP_009598268; 97% at the amino acid level). This, most likely, indicates that tailspikes of phages APK2, IME200, and SH-Ab 15519 can interact specifically with CPS of the same structure.
Deletion mutants lacking the N-terminal domains of the tailspikes were cloned, expressed, and purified by immobilized metal ion affinity chromatography, followed by ion-exchange chromatography. The recombinant proteins were stable for at least 2 months at 4°C, retaining sufficient depolymerase activities. An example of serial 2-fold titration of one of the purified recombinant depolymerases on the bacterial lawn of a host strain, after 8 h of incubation, is presented in Fig. 1C.
The spectra of depolymerase activity of recombinant proteins were tested against a panel of A. baumannii strains with the confirmed CPS structures (Table 1) belonging to 56 different K types. All purified recombinant depolymerases, except for APK2_ gp43, were found to be highly specific and formed opaque haloes only on the bacterial lawns of A. baumannii strains of the corresponding K types. APK2_gp43 formed haloes on the bacterial lawns of K2 and K93 strains (data not shown), meaning that this depolymerase effectively degrades CPSs of both types. In order to investigate the role of TSDs in phage infection of corresponding A. baumannii host cells, we have conducted a series of competition experiments (Fig. 4). For this, bacterial cell cultures preincubated with purified TSD proteins were mixed with several phage dilutions and plated on agar dishes. After overnight incubation, phage titers were measured. A negative-control experiment series where phage host bacterial FIG 3 Comparison of phage genomes using the following color scheme: light gray, genes with unknown functions; green, genes encoding enzymes of nucleotide metabolism, proteins involved in DNA replication and repair, and RNA polymerase; red, structural protein genes; yellow, genes encoding proteins associated with lysis; violet, DNA-packaging protein genes. Maps were created with Easyfig. cells were pretreated with bovine serum albumin (BSA) showed no significant differences in phage titers, whereas coincubation with APK2_gp43, APK32_gp46, APK37_gp44, APK44_gp44, APK48_gp43, APK87_gp48, APK89_gp46, and APK116_gp43 resulted in A. baumannii host cells becoming nonsusceptible to infection by the corresponding phages. Thus, the addition of 20 mM TSD proteins to the cells completely inhibits plaque formation. This means that the TSDs effectively degraded capsular polysaccharide layers surrounding A. baumannii host cells and, after that, specific phages could not adsorb to the cells. The results obtained confirm that the CPSs are the primary receptors for the phages and that TSDs play a crucial role in the initial step of the phage-bacterial cell interaction.
Structures of the CPSs of A. baumannii phage host strains. To elucidate mechanisms of the action of phage depolymerases, structures of the phage host CPSs, which are the primary receptors for depolymerase-carrying bacteriophages, were characterized.
Structures of the K87 and K89 CPSs of A. baumannii LUH5547 and LUH5552, respectively, have not been reported earlier but, according to our unpublished data, are as shown in Fig. 6. The CPS of strain LUH5552 has a pentasaccharide K unit, which in addition to common monosaccharides contains 3-acetamido-3,6-dideoxy-D-galactose (Fuc3NAc). The CPS of LUH5547 is distinguished by the presence of D-GlcA, 6-deoxy-Ltalose (L-6dTal), and three residues of L-rhamnose (L-Rha) in a heptasaccharide K unit. Details of the structure elucidation of these CPSs will be reported elsewhere.
Structures of the oligosaccharides obtained were established by one-and twodimensional 1 H and 13 C NMR spectroscopy (25) and were confirmed by high-resolution electrospray ionization mass spectrometry (HR ESI-MS) ( Table 4). All oligosaccharides had the same monosaccharide composition as the CPSs they were derived from.
The 1 H and 13 C NMR spectra of the K unit monomers were fully assigned by twodimensional shift-correlated experiments ( 1 H-1 H correlation spectroscopy [COSY], 1 H-1 H total correlation spectroscopy [TOCSY], and 1 H-13 C heteronuclear single quantum coherence [HSQC] spectroscopy) and compared with the data of the corresponding CPSs (Tables 5 to 11). Linkage and sequence analyses by two-dimensional 1 H-1 H rotating-frame nuclear Overhauser effect (ROESY) and 1 H-13 C heteronuclear multiple-bond The 13 C NMR chemical shifts of all but two monosaccharide residues in the K unit monomers were essentially the same in the oligosaccharides and the corresponding CPSs, whereas those of the residues at the reducing and nonreducing ends of the oligosaccharides were different. On this basis, the glycosidic linkages that were cleaved in the CPSs by the TSDs could be identified (Table 3). As expected, the monosaccharides that occupy the reducing end of the oligosaccharides (D-GlcNAc in 5 and 8, D-GalNAc in 2, 6, 10, and 12) were present in two anomeric forms (a and b), which showed in the 13 C NMR spectrum the C-1 signals characteristic for nonlinked monosaccharides at d 92.3-92.6 and 95.7-96.5, respectively (Tables 5 to 11). In oligosaccharide 1, the reducing end was occupied by a legionaminic acid derivative (Legp5Ac7R), as followed from a displacement of the C-6 signal of this monosaccharide from d 73.1 in the CPS to d 70.9 in 1 ( Table 5). This difference reflected conversion, upon cleavage of the glyosidic linkage of the a-linked Legp5Ac7R residue with the axial carboxyl group, into a more stable b-anomer with the equatorial carboxyl group (26).
Furthermore, the 13 C NMR chemical shifts of unit C (b-Gal) in 1 were typical of the corresponding nonsubstituted residues (Table 5) (27). Hence, this monosaccharide, which was substituted in the corresponding initial CPS, was located at the nonreducing end in 1 (Fig. 5). Unit B (b-Gal in 10, a-GalNAc in 8), unit C (b-GalNAc in 6), and unit D (a-Rha in 5, b-GalNAc in 2) were disubstituted in the CPSs but became monosubstituted in the corresponding oligosaccharides ( Fig. 5-7). This conclusion followed from significant upfield displacements (by 7.3 to 10.2 ppm) of the signals for C-4 of unit D in 2 and C-3 of units B, C, or D in 5, 6, 8, and 10, compared with their positions in the corresponding CPSs (Tables 5 to 10). Therefore, these carbons that were linked in the CPSs became nonlinked in the oligosaccharides. These data defined the structures of the oligosaccharides obtained by depolymerization of the CPSs (Fig. 5-7) and, as a result, identified the linkages that were cleaved by phage depolymerases ( Table 3).
The CPSs of strains ACICU and B11911, having the same main chains consisting of disaccharide repeats but different side chains (Fig. 7), were cleaved with the same depolymerase APK2_gp43, in the same manner, by the b-D-GalpNAc-(1!3)-b-D-Galp (AfiB) linkage (Fig. 7). The same linkage was also cleaved by a TSD of bacteriophage w AB6, in the CPS of A. baumannii 54149 (6), having the same structure as the CPS of strain ACICU.
The 1 H and 13 C NMR spectra of the larger oligosaccharides 3, 7, 9, and 11 showed two series of signals, one being the same as in the corresponding K unit monomers 2, 6, 8, and 10 and the other being the same as in the corresponding CPS K units. Based on these and HR ESI-MS data (Table 4), it was concluded that these oligosaccharides represented K unit dimers having the structures shown in Fig. 5 to 7. A comparison of the NMR spectra of oligosaccharide 4 and those of the CPS of strain NIPH615 (22), combined with the HR ESI-MS data ( Table 4), showed that 4 was a K unit dimer with a b-D-GlcpNAc residue at the reducing end (Fig. 5).

DISCUSSION
In the scope of phage application for the development of preparations to control multidrug-resistant A. baumannii strains, a comprehensive characterization of prospective candidates to be included in these preparations is required. One of the important characteristics for the applied usage of virulent phages is a host range-the ability of a phage to demonstrate a lytic activity against diverse, clinically important bacterial strains that circulate in the area of estimated application. Taking into account that the CPSs are the primary receptors for depolymerase-carrying A. baumannii phages (5-9), we believe that the range of their lytic spectra depends on the prevalence of A. baumannii strains with appropriate CPS structures.
High diversity and variability of the CPS structures imply the existence of the same variety of phage depolymerases that can specifically recognize and cleave different CPSs. Therefore, preparations ("cocktails") comprising different phages with the corresponding CPS-recognizing proteins are required to eliminate bacteria of different capsular types within the same species. Certainly, in rare cases, several variants of K loci can be responsible for the synthesis of CPSs of the same structure (for example, KL3/ 22), or some phages like APK2, described in this work, are able to infect bacteria  Table 5  APK37 gp44 Table 11 a For full CPS structures, see Fig. 5 to 7.  (8,9,28), K19 (phages Fri1 and AS11) (7), K27 (phage AS12) (7), K45 (phage vB_AbaM_B9, which also performed lysis from without in a K30 strain) (29), and K91 (phage AP22) (13) capsular types of A. baumannii have been described. For some other depolymerase-carrying phages, there is no information as to which capsular type their A. baumannii host strains belong (30)(31)(32).
In this work, eight novel A. baumannii bacterial viruses and depolymerases, encoded in their genomes, were characterized. All phages share a similar genome organization with previously reported viruses of the Friunavirus genus, with the most variable region falling into gene regions encoding CPS-recognizing/degrading domains of tailspikes or structural depolymerases. Seven of the phages, namely, APK32, APK37, APK44, APKK48, APKK87, APKK89, and APK116, are the first reported viruses specific to A. baumannii strains of the K32, K37, K44, K48, K87, K89, and K116 capsular types, respectively. Phage APK2 was shown to infect A. baumannii K2 like the earlier described phages w AB6 (6) and vB_AbaP_B3 (8) and also A. baumannii K93 like phage APK93. Interestingly, phages APK2 and APK93 were initially isolated from different samples in 2018 and were given their names because of their isolation on the K2 and K93 bacterial lawns, respectively. Genome sequencing of these phages revealed that they are close homologs and that their receptor-recognizing/binding proteins or TSDs are completely identical. Therefore, these phages can be recognized as closely related variants.  Analysis of oligosaccharide products obtained by degradation of the A. baumannii CPSs by recombinant depolymerases APK2_gp43, APK32_gp46, APK37_gp44, APK44_ gp44, APK48_gp43, APK87_gp48, APK89_gp46, and APK116_gp43 showed that all TSDs studied were specific glycosidases that cleaved the CPSs by the hydrolytic mechanism, to give a monomer or/and an oligomer(s) of the K units. The specific interaction of the APK2_gp43 depolymerase with both K2 and K93 CPSs was suggested to be due to the similarity of their K units, which have identical main chains and differ only in their side chain structures (Fig. 7). Particularly identical are the linkages between the K2 and K93 K units that are cleaved by the APK2_gp43 depolymerase.
Worthy of note is that phages APK2, w AB6, and vB_AbaP_B3, with the same specificity to K2 A. baumannii strains, were independently isolated in remote geographic locations: Russia, China, and Portugal, respectively. The same situation has been observed for phages that infect A. baumannii strains of the K9 capsular type. These findings suggest that representatives of these K types are widely spread around the world and that the phages should adopt their receptor-binding/recognizing proteins to successfully infect them.
In conclusion, the comprehensive characterization of new lytic phages and phageencoded TSDs expand our knowledge of virus-bacterial host interaction strategies. Numerous A. baumannii phages isolated by different research groups, in various geographic locations, have demonstrated "broad" or "narrow" lytic spectra. As each TSD specifically recognizes and enzymatically digests only certain CPS, we declare that the spectrum range of a depolymerase-carrying A. baumannii phage depends on the  distribution of A. baumannii strains with appropriate CPS structure among all those tested in each particular case. That means the different phages, independently on broad or narrow spectra, could be efficiently used as antibacterial tools, depending on the clinical situation.

MATERIALS AND METHODS
Phage isolation, propagation, and purification. For phage isolation, a collection of various A. baumannii strains with defined CPS structures (listed in Table 1), kindly provided by the members of research groups from different countries (see Acknowledgements), was used. Bacteriophages were isolated from sewage and environmental (river water) samples, in accordance with a previously reported procedure (33). The samples were cleared by low-speed centrifugation (7,000 Â g for 30 min), the supernatants supplemented with LB medium were incubated for 16 to 18 h in the presence of growing A. baumannii strains belonging to capsular types of interest at 37°C, and then chloroform was added. Bacterial debris was pelleted by centrifugation at 7,000 Â g for 30 min, followed by filtration of the supernatants through 1.20-and 0.45-mm-pore-size membrane filters. The purified filtrates were concentrated by ultracentrifugation at 85,000 Â g at 48C for 2 h. The spot test method, as well as the plaque assay (34), was used to  screen for the presence of lytic phage activity in the resulting concentrated preparations. The plates were incubated overnight at 37°C and examined for zones of lysis or plaque formation. Single plaques formed on the lawns of sensitive A. baumannii strains were picked up in SM buffer (10 mM Tris-HCl, pH 7.5, 10 mM MgSO 4 , 100 mM NaCl) and replated three times, in order to obtain pure phage stock.
The phages were propagated using a liquid culture of the A. baumannii host strains (optical density at 600 nm [OD 600 ] of 0.3) at a multiplicity of infection (MOI) of 0.1.
The phage particles were precipitated by polyethylene glycol (PEG) 8000 (to a final concentration of 10% and 500 mM NaCl). The final purification was executed by cesium chloride density gradient centrifugation at 100,000 Â g (SW50.1 Ti rotor; Beckman Coulter Inc., Brea, CA, USA) for 24 h (35).
Phage host specificity determination. The host specificity of the phages was tested against A. baumannii strains belonging to 56 different K types (presented in Table 1), using the double-layer method (34). For this, 300 ml of A. baumannii bacterial cultures grown in LB medium at 37°C to an OD 600 of 0.3 was mixed with 4 ml of soft agar (LB broth supplemented with 0.6% agarose). The mixture was plated onto the nutrient agar. Then, the phage suspensions (;10 9 PFU per ml) or purified recombinant depolymerases, and their several dilutions, were spotted on the soft agar lawns and incubated at 37°C for 18 to 24 h.
Phage infection inhibition assay. The effect of purified TSD proteins APK2_gp43, APK32_gp46, APK37_gp44, APK44_gp44, APK48_gp43, APK87_gp48, APK89_gp46, and APK116_gp43 on the phage infection of A. baumannii ACICU, LUH5549, NIPH146, NIPH70, NIPH615, LUH5547, LUH5552, and MAR303 host cells was assessed as following. A titer of 2 Â 10 6 PFU/ml for phages APK2, APK32, APK37, APK44, APK48, APK87, APK89, and APK116 was chosen for the competition experiments. A. baumannii host strains were grown in LB medium at 37°C to an OD 600 of 0.3. Then, 20 mM concentrations of corresponding TSD solutions were added to 100-ml aliquots of the cell cultures and incubated for 20 min at 37°C. One-hundred-microliter  aliquots of the A. baumannii host cells without anything and with 20 mM BSA incubated for 20 min at 37°C served as controls. After the incubation, several dilutions of corresponding phages and 4 ml of soft agar were added to the mixtures and plated onto the nutrient agar. The plates were incubated overnight at 37°C and assayed for the number of lysis plaques. The experiment was performed in triplicate. Electron microscopy. The phage was examined by negative-contrast electron microscopy, using the following procedure. Three microliters of purified and concentrated phage preparation was applied to the carbon-coated 400 mesh copper grids and subjected to glow discharge using the Emitech K100X apparatus (Quorum Technologies, UK). Grids were then negatively stained with 1% uranyl acetate for 30 s, air dried, and analyzed using a JEOL JEM-2100 200-kV transmission electron microscope. Images of negatively stained phage particles were taken with a Gatan Ultrascan 1000XP charge-coupled-device (CCD) camera (14 mm pixels) and Gatan Digital Micrograph software with the following parameters: Â30,000 magnification, 0.5 to 1 mm defocus, 40-mm objective aperture, 2k by 2k pixel size unbinned image size, 3.4-Å pixel size.
DNA isolation and sequencing. Phage genomic DNAs were isolated from concentrated and purified high-titer phage stocks by incubation in 0.5% SDS, 20 mM EDTA, and 50 mg/ml proteinase K at 56°C for 1 to 3 h. The DNAs were extracted with phenol-chloroform and then precipitated with ethanol supplemented with sodium acetate (35).
Genome sequencing was performed on the MiSeq platform using a Nextera DNA library preparation kit (Illumina, San Diego, CA). The generated reads were assembled de novo into a single contig using SPAdes v.3.13 (36). Further, the resulting sequences were checked by mapping reads against the assemblies with DNASTAR's Lasergene sequence analysis software, version 11.1.0 (DNASTAR, Inc., Madison, WI, USA) (37). Direct terminal repeats were determined as regions of greater coverage of sequencing reads mapped on assembled viral genome contigs. The ends were then verified directly by Sanger sequencing with outward-directed primers located inside and outside putative repeats.
Phage genome analysis. Potential open reading frames (ORFs) were identified with the RAST automated annotation engine (38). Predicted proteins were searched against the NR (nonredundant) database of the NCBI and HHpred profile-profile search (39). The presence of tRNAs in the genome sequence was determined using tRNAscan-SE, version 2.0 (40). Comparative analysis of DNA genome sequences was performed using Easyfig (41).
Cloning, expression, and purification of the recombinant depolymerases. The DNA fragments corresponding to the deletion mutants lacking N-terminal domains of putative podophage  depolymerases were amplified by PCR, using oligonucleotide primers indicated in Table 12, and cloned into the pTSL plasmid (GenBank accession no. KU314761) (42). Expression vectors were transformed into chemically competent Escherichia coli B834(DE3) cells. Protein expression was performed in LB medium supplemented with ampicillin at 100 mg/ml. Transformed cells were grown at 37°C until the optical density reached a value of 0.6 at 600 nm. The medium was cooled to the temperature of 18°C, followed by expression induction by the addition of isopropyl-1-thio-b-D-galactopyranoside (IPTG) to a final concentration of 0.5 to 1.0 mM. After further incubation at 18°C overnight (approximately 16 h), the cells were harvested by centrifugation at 10,000 Â g for 15 min, at 4°C. The cell pellets were resuspended in buffer A (20 mM Tris [pH 8.0], 300 mM NaCl, 1/50 of the original cell volume), frozen, thawed, and disintegrated by sonication (Branson Ultrasonic, Danbury, CT, USA). The cell debris was removed by centrifugation at 15,000 Â g at 4°C for 20 min. The supernatant was applied to 5-ml Ni 21 -charged GE HisTrap columns (GE Healthcare Life Sciences, Chicago, IL, USA) equilibrated with buffer A, and the proteins were eluted with a 50-to-200 mM imidazole linear gradient in buffer A. Fractions containing the target protein were pooled and digested with tobacco etch virus (TEV) protease for 16 h at 20°C at a protease/protein ratio of 1/100 (wt/wt) to remove the His tag. The reaction mixture was simultaneously dialyzed against 10 mM Tris-HCl (pH 8.0) containing 1.0 mM 2-mercaptoethanol.
The cleaved proteins were clarified by filtration, loaded onto an ion-exchange MonoQ 10/100 GL column (GE Healthcare), and eluted with a 0-to-650 mM NaCl gradient in 20 mM Tris-HCl (pH 8.0) buffer. Protein-containing fractions were combined and concentrated to ;5 ml using Sartorius ultrafiltration devices (Sartorius AG, Germany), with a molecular-weight cutoff of 50,000 Da.
Capsular polysaccharides (CPSs) were isolated by extraction of A. baumannii cells with 45% aqueous phenol for 30 min at 65 to 68°C (43). The extract was cooled and dialyzed without layer separation; then, insoluble contaminations were removed by centrifugation (12,000 Â g, 20 min), and CPS preparations were purified as described previously (44). Briefly, aqueous 50% CCl 3 CO 2 H was added to a CPS solution in water at 4°C, a precipitate was removed by centrifugation, and the supernatant was dialyzed with distilled water and freeze-dried. The crude CPS preparations were heated with 2% HOAc (100°C, 3 h), and a lipid precipitate was removed by centrifugation (12,000 Â g, 20 min). Purified CPS samples were isolated from the supernatant by gel permeation chromatography on a XK 26-mm (depth) by 70-cm (height) column (gel layer, 560 mm) (GE Healthcare) of Sephadex G-50 Superfine (Amersham Biosciences, Sweden) in 0.05 M pyridinium acetate buffer, pH 4.5. The flow rate was 0.5 ml min 21 ; elution was monitored with a differential refractometer (Knauer, Germany). Control of retention of the intact structure upon mild acid treatment was performed by NMR spectroscopy.
Cleavage of the CPSs with tailspike depolymerases. Purified CPSs were solubilized in 20 mM Tris-HCl (pH 7.5) buffer, and 200 to 500 mg of the corresponding TSD proteins was added for digestion. The reaction mixture was stored at 37°C for 16 h.