Development of an Anti-Acinetobacter baumannii Biofilm Phage Cocktail: Genomic Adaptation to the Host

ABSTRACT

The need for alternatives to antibiotic therapy due to the emergence of multidrug resistant bacteria (MDR), such as the nosocomial pathogen Acinetobacter baumannii, has led to the recovery of phage therapy. In addition, phages can be combined in cocktails to increase the host range. In this study, the evolutionary mechanism of adaptation was utilized in order to develop a phage adapted to A. baumannii, named phage Ab105-2phiΔCI404ad, from a mutant lytic phage, Ab105-2phiΔCI, previously developed by our group. The whole genome sequence of phage Ab105-2phiΔCI404ad was determined, showing that four genomic rearrangements events occurred in the tail morphogenesis module affecting the ORFs encoding the host receptor binding sites. As a consequence of the genomic rearrangements, 10 ORFs were lost and four new ORFs were obtained, all encoding tail proteins; two inverted regions were also derived from these events. The adaptation process increased the host range of the adapted phage by almost 3-fold. In addition, a depolymerase-expressing phenotype, indicated by formation of a halo, which was not observed in the ancestral phage, was obtained in 81% of the infected strains. A phage cocktail was formed by combining this phage with the A. baumannii phage vB_AbaP_B3, known to express a depolymerase. Both the individual phages and the phage cocktail showed strong antimicrobial activity against 5 clinical strains and 1 reference strain of A. baumannii tested. However, in all cases resistance to the bacterial strains was also observed. The antibiofilm activity of the individual phages and the cocktail was assayed. The phage cocktail displayed strong antibiofilm activity.

INTRODUCTION

The Gram-negative bacterium Acinetobacter baumannii causes bacteremia, pneumonia, skin and soft tissue infections, and meningitis, and its ability to become resistant to almost all known antibiotics leads to serious problems (1). The ability of the bacterium to form biofilms and persist in dry environments, along with the acquisition of multidrug resistance (MDR), makes treatment of the associated infections a challenge in hospitals and intensive care units (ICUs) (2).

Bacterial biofilms are structured communities of encapsulated bacteria held in a matrix composed by extracellular polymeric substances (EPS) (3). The ability of bacteria to form biofilms is considered an important virulence factor that confers resistance to antibiotics, the host immune system, and environmentally harsh conditions. Biofilms represent a serious health problem as they are associated with device-related infections and non-device-related chronic inflammatory conditions such as cystic fibrosis, obstructive pulmonary disease, otitis, and prostatitis, as well as with wound and skin infections (1, 3). Many infections involve biofilm formation, and approximately 80% of all bacterial infections in humans are thought to be derived from a biofilm (4).

The main treatments for infections are currently based on the application of antibiotic therapy. However, the emergence of MDR makes necessary the use of other antimicrobial agents such as bacteriophages (viruses that infect bacteria) and their derived lytic proteins, lysins, and depolymerases (5, 6). Phage therapy is based on the targeted application of phages that infect and kill bacteria thereby controlling the infection (7). Phage therapy has some limitations, including the narrow host range, the existence or emergence of bacterial resistance, and the inaccessibility to the bacteria caused by the formation of biofilms. Some bacteria are surrounded by a capsule, conformed by capsular polysaccharides (CPS), that act as a protective barrier. The capsule can be digested by some phages carrying polysaccharide lytic enzymes named depolymerases, thus getting access to the inaccessible bacterial surface (8). All these limitations can be overcome by the use of evolved phages that are adapted to the resistant bacteria and also by the use of cocktails of different phages, thus increasing the host range and reducing the emergence of resistant bacteria (9, 10).

Phages and bacteria coexist in a continuous race in antagonistic coevolution, which drives the selection of resistance and infectivity (11). The coevolution implies the alteration of several features, of both phages and bacteria, involved in phage–bacteria interactions. Such alterations can lead to phages overcoming or minimizing bacterial resistance (12). Bacteria can evade phages by several mechanisms: prevention of adsorption by blocking phage receptors or by producing an extracellular matrix or competitive inhibitors; cleaving the injected DNA mediated by the restriction-modification systems, the CRISPR-Cas systems, and the DISARM system; and abortive infection systems, which are frequently encoded by mobile elements such as prophages (13–15). Resistance mechanisms can be evaded by phages through genetic mutations that lead to the production of proteins that overcome bacterial resistance mechanisms. This can be used in phage therapy by adapting phages to resistant bacteria; the adapted phages are active against resistant variants and have an increased host range (12, 16). The increase in the host range is brought about by the acquisition of point mutations in receptor binding proteins (RBP), which allows phages to target new receptors rather than adapting to the modified receptor (14). The host range is also influenced by host factors such as the resistance mechanisms, environmental factors, and the features encoded by the phage, thus enabling adaptation to these mechanisms (17).

In this study, the mutant lytic phage Ab105-2phiΔCI (5) was adapted to a clinical strain of A. baummannii, giving rise to an adapted phage, Ab105-2phiΔCI404ad, with an increased host range and a depolymerase-expressing phenotype. The genome of the adapted phage was sequenced, and several mutations were observed in the region corresponding to the tail morphogenesis module, which we refer to as a “hot adaptation site.” With the objective of reducing the resistance generated by this phage, it was combined with the A. baumannii phage vB_AbaP_B3 (18), which expresses a depolymerase. The antibiofilm activity of this cocktail was tested against the biofilms produced by 5 clinical strains and 1 type strain of A. baumannii, and synergistic antibiofilm activity was observed in 5 cases.

RESULTS

Phage host range and adaptation of the phage Ab105-2phiΔCI.

The host range of phages B3 and Ab105-2phiΔCI, a lytic mutant phage that was obtained after deletion of the regulator gene CI, a repressor of the lytic regulator Cro, thereby inducing the entrance of the phage in a lysogenic state (19), from the temperate phage Ab105-2phi (5), was determined using 30 isolates of A. baumannii (Table 1). The host range was first checked by spot test and then confirmed by EOP. The results showed that although phage B3 infected 43% of the strains in the spot test, this was reduced to 36% after taking the EOP into consideration. A halo was observed around 81% of the infected strains; this halo is characteristic of the depolymerase activity previously described for this phage (8, 18). It is produced by the action of a depolymerase that freely diffuse degrading the polysaccharide and is time-dependent (8). The mutant lytic phage Ab105-2phiΔCI infected 13% of the strains, as indicated by the spot test and confirmed by the EOP, and no halo was observed. In order to increase the host range, the phage was adapted by applying Appelmans method using the clinical strain Ab404_GEIH-2010 as host. This strain was selected because of the presence of a halo when infected with the phage B3 in double agar plate and because this strain was resistant to the phage Ab105-2phiΔCI. After two rounds of adaptation, complete lysis was obtained and the adapted phage Ab105-2phiΔCI404ad was recovered. The adapted phage produced a halo surrounding the spot and that increased over time (even when the plates were stored at 4°C), indicating the presence of a depolymerase (20–23) (Fig. 1A). The host range of the adapted phage was tested by spot test and EOP, showing an infective capacity of 36%, in addition to the presence of a halo in 81% of the infected strains. The clear plaques observed in all infected strains indicated that the infection was produced by lytic phages in all cases; besides, it should be noted that the halo was present in the same strains for both phages.

TABLE 1.

Host range determined for phages vB_AbaP_B3 and Ab105-2phiΔCI and the adapted phage Ab105-2phiΔCI404ad^d

		Bacteriophages
A. baumannii strain	ST	vB_Abap_B3	Ab105-2phiΔCI	Ab105-2phiΔCI404ad
Ab007_GEIH-2010	2	1.085a	NDc	0.900a
Ab008_GEIH-2010	2	1.123a	NDc	0.876a
Ab019_GEIH-2010	254	0.769a	NDc	0.833a
Ab034_GEIH-2010	2	1.027a	NDc	0.980a
Ab05_GEIH-2010	186	NDc	NDc	NDc
Ab09_GEIH-2010	297	NDc	NDc	NDc
Ab104_GEIH-2010	2	0.010b	NDc	NDc
Ab105_GEIH-2010	2	0.003b	1b	0.0002b
Ab13_GEIH-2010	79	NDc	NDc	NDc
Ab155_GEIH-2000	2	NDc	NDc	NDc
Ab174_GEIH-2010	2	NDc	NDc	NDc
Ab177_GEIH-2000	2	NDc	1.550b	0.001b
Ab192_GEIH-2010	79	1.100a	0.220b	1.060a
Ab22_GEIH-2010	52	NDc	NDc	NDc
Ab226_GEIH-2010	181	NDc	NDc	NDc
Ab24_GEIH-2010	255	NDc	NDc	NDc
Ab290_GEIH-2010	264	NDc	NDc	NDc
Ab291_GEIH-2010	265	NDc	NDc	NDc
Ab307_GEIH-2010	266	0.250a	NDc	0.414a
Ab309_GEIH-2010	267	NDc	NDc	NDc
Ab310_GEIH-2010	268	NDc	NDc	NDc
Ab32_GEIH-2010	2	NDc	NDc	NDc
Ab404_GEIH-2010	80	1.050a	0.0002b	1a
Ab440_GEIH-2010	187	NDc	NDc	NDc
Ab456_GEIH-2010	269	NDc	NDc	NDc
Ab459_GEIH-2010	182	0.112a	NDc	0.104a
Ab461_GEIH-2010	270	NDc	NDc	NDc
Ab462_GEIH-2010	271	NDc	NDc	NDc
Ab464_GEIH-2010	272	NDc	NDc	NDc
NIPH2061	ND	1a	NDc	1.080a
EOP %		36%	13%	36%

^aThe spot test results are shown as, clear spots with a halo.

^bClear spots.

^cNo spots. Efficiency of plating (EOP) is shown as numerical values for those infections that previously produced a spot. High: EOP ≥ 0.5; Medium: 0.1 ≤ EOP < 0.5; Low: 0.001 < EOP < 0.1; Invalid: EOP ≤ 0.001; ND, not determined (65). Each phage host has a value of 1. The sequence type (ST) of each of the strains is indicated.

^dThe sensitivity of 29 clinical strains of A. baumannii and 1 reference strain to the phages was tested.

FIG 1.

Genome analysis of the adapted phage Ab105-2phiΔCI404ad. (A) Spots of the three phages tested in the study in an agar plate with the host Ab404_GEIH-2010. The halos around the spot of the phages vB_AbaP_B3 and Ab105-2phiΔCI404ad are shown. (B) Comparison of the genome of the ancestral phage Ab105-2phiΔCI (the ORF17 was previously deleted in this phage) and the adapted phage Ab105-2phiΔCI404ad. The deleted genes are shown in red, and each of the genes that recombined and the resultant genes are shown in different colors. (C) Homology scheme of the genes derived from the recombination events in the phage Ab105-2phiΔCI404ad and those ancestral genes of the phage Ab105-2phiΔCI. The colors correspond to the scheme in section B; the homology region is represented in solid colors, and the percentage homology is indicated. (D) Schematic representation of the domains identified in the ORF69 and ORF74 in phage Ab105-2phiΔCI404ad.

Storage stability of phages.

The storage stability was tested for both phages B3 and 404ad at two temperatures, 4°C and −20°C for 30 days. When both phages were stored at 4°C, no significant differences in the titers were obtained for any of the phages when compared with the same phage before the storage (B3 = 1 × 10¹⁰ CFU/mL; 404ad = 3.53 × 10⁰⁹ CFU/mL). When the titers obtained after the storage at –20°C were quantified, no differences were obtained for the phage B3, but a significant reduction (P < 0.05) of 28% was obtained for the phage 404ad.

Genome analysis of the adapted phage Ab105-2phiΔCI404ad.

The whole genome sequence of the adapted phage 404ad was obtained and compared with the genome sequence of the viral database of the GenBank, showing identity with 11 phage genomes, including the phage Ab105-2phi, which (as already mentioned) is the phage employed to delete the CI regulator (ORF17) (5), thus obtaining the ancestral phage Ab105-2phiΔCI. A phylogenetic tree was constructed showing the maximum similarity with the phage 404ad with the mutant lytic phage Ab105-2phiΔCI (Fig. S1 in the supplemental material).

The genome analysis revealed that the genome of phage 404ad was 53211 bp in length, while the genome of the lytic mutant phage Ab105-2phiΔCI was 60629 bp in length (Fig. 1B). In addition, phage 404ad contains 82 ORFs, while the phage AB105-2phiΔCI contains 92 ORFs. Ten genes were lost in the adaptation process, corresponding to the loss of 7418 bp. Comparison with the genome of the phage B3 did not reveal any significant homology.

The comparison showed several genomic rearrangements that include a region of recombination and deletion between ORF57 and ORF85 and a deletion in ORF14 of the original phage Ab105-2phiΔC (Fig. 1B).

In ORF14, a deletion of 63 pb (6797-6860 pb) corresponding to 21 aa occurred. Although this region was not present in ORF14 of phage Ab105-2phiΔC404ad, the homology analysis showed that both genes were homologous and encoded a putative membrane protein.

Analysis of the lost proteins (encoded by ORF76 to ORF85) revealed that they correspond to 3 proteins of the tail module, 4 hypothetical proteins, and 3 regulator proteins. We also observed that four recombination events occurred, in ORF57-ORF69, ORF69-ORF85, ORF58-ORF73, and ORF58-ORF74-ORF75 of phage Ab105-2phiΔC. These recombination events gave rise to the inversion of two flanked genomic regions and to 4 new genes in the adapted phage Ab105-2phiΔC404ad: ORF56, ORF69, ORF70, and ORF74 (Fig. 1B and C). Thus, the tail fiber protein, encoded by ORF56 (404ad), was derived from the recombination of ORF57 and ORF69, and a tail tip protein, encoded by ORF74 (404ad), was derived from the recombination of ORF69 and ORF85 from the phage Ab105-2phiΔCI. Recombination of the ORF58, ORF74, and ORF75 genes (Ab105-2phiΔCI), which encoded 3 tail tape measure proteins, gave rise to another tail tape measure protein encoded by ORF69 (404ad). Finally, recombination of the ORF73 and ORF58 genes (Ab105-2phiΔCI) gave rise to a new gene, ORF70 (404ad), encoding a tail tape measure protein.

As the adaptation process mainly occurs by deletion and recombination events in the tail module (24, 25), there is a high probability of mutations in this region, which we referred to as a “hot adaptation site.”

Analysis of the amino acid sequence of the 4 new proteins (Fig. 1D) led to identification of 3 domains in the protein encoded by ORF69: a tape measure domain, characteristic of the tail tape measure proteins; a NlpD superfamily, i.e., a murein DD-endopeptidase activator containing a LysM domain; and a DUF5401 domain with unknown function. The two first domains were also present in the ancestral phage protein encoded by ORF75 (Ab105-2phiΔCI), and the third domain was present in the ORF74 (Ab105-2phiΔCI). A domain corresponding to a GP38 superfamily was identified in ORF74 (404ad), which encodes a tail fiber adhesin. This domain was not present in any of the proteins in phage Ab105-2phiΔCI, homologous to the protein encoded by the ORF74 (404ad).

Phage and phage cocktail antimicrobial activity.

The antimicrobial activity of phage B3, phage 404ad, and the cocktail composed of both was assayed by growth infection curves, which revealed high antimicrobial activity in all cases 6 h after infection.

Infection curves constructed for the host strain Ab404_GEIH-2010 revealed high antimicrobial activity at all MOI tested with all phages. However, the best results were obtained for infection with the adapted phage 404ad and with the phage cocktail (Fig. 2A). After 4 or 5 h, depending on the MOI, emergence of phage resistant mutants was observed as an increase in the OD_600nm for each phage and the phage cocktail, although higher for the phage B3.

FIG 2.

Infection curves. (A) Infection curves for phages vB_AbaP_B3, Ab105-2phiΔCI404ad, and the cocktail composed by both phages, in a planktonic culture of the strain Ab404_GEIH-2010 at MOI 0.1, 1, and 10. (B) Infection curves for phages vB_AbaP_B3, Ab105-2phiΔCI404ad, and the cocktail composed by both phages, at MOI 10 in planktonic culture of 6 strains of A. baumannii. Measurements were made in triplicate and the standard deviation represented as bars.

Infection curves were also constructed for 6 selected strains in which the adapted phage yielded a spot surrounded by a halo (Fig. 2B). Growth of almost all strains was reduced and remained low for 6 h. As an exception, no infection was detected for strain Ab007_GEIH-2010.

Frequency occurrence of phage resistant mutants.

The rate of resistant mutants was established for the 6 A. baumannii strains tested (Table 2). Both phages, B3 and 440ad, as well as the phage cocktail generated resistance in all strains tested, but at low rate. The phage B3 was able to induce more resistance than the phage 404ad and the phage cocktail, in all strains. Finally, the clinical strain Ab404_GEIH-2010 produces the lower rate of resistant mutants, and the higher rate was produced by the clinical strain Ab007_GEIH-2010.

TABLE 2.

The frequency rate of the phage resistant mutants for phages B3 and 404ad and the cocktail was calculated for 5 clinical and 1 reference strain of A. baumannii

	Bacteriophages
A. baumannii strain	vB_Abap_B3	Ab105-2phiΔCI404ad	Cocktail
Ab404_GEIH-2010	6.02 ± 4.32 × 10−6	3.74 ± 1.86 × 10−6	3.73 ± 1.86 × 10−6
Ab019_GEIH-2010	5.73 ± 0.97 × 10−5	3.60 ± 3.11 × 10−6	7.09 ± 5.6 × 10−6
Ab008_GEIH-2010	7.22 ± 2.73 × 10−4	8.05 ± 1.82 × 10−5	6.73 ± 3.86 × 10−5
Ab034_GEIH-2010	2.89 ± 0.1 × 10−4	1.61 ± 0.12 × 10−5	8.14 ± 0.48 × 10−5
Ab007_GEIH-2010	2.35 ± 0.44 × 10−3	3.63 ± 1.65 × 10−3	4.83 ± 0.78 × 10−4
NIPH2061	8.3 ± 4.88 × 10−5	4.2 ± 3.44 × 10−5	7.3 ± 3.72 × 10−5

Removal of biofilm.

The antibiofilm activity of phage B3, phage 404ad, and the cocktail of both phages was tested (Fig. 3). As the crystal violet can stain the capsule EPS, the degradation of the biofilm was also tested by biofilm CFU quantification. When the biofilm was measured by crystal violet staining, significant differences were obtained, in all strains, when compared with the biofilm in the control and those treated with individual phages and the cocktail, with the exception of the strain Ab019_GEIH-2010. These significant differences were maintained in all strains, when the biofilm CFU were quantified. It was also noticeable that except for the strain NIPH2061, the biofilm was significantly reduced when treated with the cocktail and compared with the biofilm treated by the individual phages, measured by both techniques.

FIG 3.

Biofilm removal capacity of phages vB_AbaP_B3, Ab105-2phiΔCI404ad, and the cocktail composed by both phages for the biofilms produced by 6 strains of A. baumannii, 5 clinical isolates, and 1 reference strain. The biofilm removal for each phage was determined by crystal violet staining (OD595) and by CFU quantification (log CFU/biofilm). Statistically significant differences were determined by a Student's t test (GraphPad Prism v.9.3.0). *, P < 0.05; **, P < 0.01.

DISCUSSION

Phage therapy can be considered a valuable alternative to antibiotic therapy, which may fail owing to the emergence of MDR. Phages may not always be suitable alternatives to antibiotic therapy, but both can be used synergistically (5). Nevertheless, phage therapy has some advantages over antibiotic therapy: phages have a narrow spectrum of activity that protects the microbiota; they can replicate at the infection site; they are abundant and easily isolated in nature; and they can overcome bacterial resistance via continuous coevolution (5, 26–28). The ability of phages to sensitize resistant bacteria is a characteristic advantage that is utilized in the selection of phages for therapy. This characteristic can be used to produce phages for bacterial strains that have become resistant to a phage or to increase the host range of a phage or a phage cocktail (29).

In this study, an adapted phage, 404ad, with a broad host range relative to the ancestral phage, Ab105-2phiΔCI, was obtained by Appelmans protocol. Analysis of the adapted phage 404ad revealed a region of recombination and deletion in the tail morphogenesis module, referred to as a “hot adaptation site” (Fig. 1B). Four important genomic rearrangements occurred in the “hot adaptation site” and led to the deletion of 10 genes, as well as the recombination and fusion of 7 genes (giving rise to 4 different genes) and also the inversion of the regions flanked by these genes (Fig. 1C). The genomic rearrangements that involve recombinations, inversions, and deletions are described in phages among other organisms. It was described that the adaptation to a resistant host mainly affects genes encoding tail proteins and commonly is a result of a recombination event (24, 25). The ancestral phage Ab105-2phiΔCI is derived from the temperate phage Ab105-2phi, which characteristically contains several proteins involved in the recombination to integrate its genome into the host genome. In this phage several proteins related to the recombination and to the genomic rearrangements, were identified as an integrase, a resolvase, and a HNH endonuclease (5, 30–32).

The specificity of the interaction between the phage and its host depends on the phage RBPs, which are located in the tail structures, so that mutations in the genes encoding the tail affect host recognition. The increase and the differences observed in the host range of the phage 404ad relative to the ancestral phage Ab105-2phiΔCI, in a manner that EOPs are inverted between them, are due to mutations in the RBPs, which are described as proteins with a high genetic plasticity that are employed for the phages to infect resistant hosts (33). The mutations in the RBPs are derived from the genomic rearrangements that took place in the “hot adaptation site,” giving rise to 4 genes encoding tail structures and also deletion of the genes encoding 3 tail structures. ORF56 encodes a tail fiber protein and ORF74 encodes a tail tip protein, which have been described as RBPs (33) and are closely related to host recognition. ORF69 and ORF70, both of which encode tail tape measure proteins that are related to the length of the tail and also to access to the bacterial cytoplasm. In some cases, these ORFs can degrade the peptidoglycan layer and penetrate the membrane, because they may contain a virion-associated lysins (VALs) domain with muralytic activity (34), such as the murein DD-endopeptidase domain identified in the tail tape protein encoded by ORF69 (Fig. 1D).

The increase in the host spectrum may also be related to the phenotypic detection of a depolymerase because, as observed in the spot test and in the EOP assays (Table 1), the efficiency of infection was always highest when a halo appeared around the lytic plaque (Fig. 1A), which is indicative of the presence of polysaccharide lytic enzymes (20–22, 35). Several studies have shown the same host range for the phages and for the depolymerase, suggesting that these enzymes are involved in host recognition (36–38).

The EPS and CPS act as a barrier that inhibits adsorption of the phage. In some phages these structures can also act as primary receptors, leading to binding and degradation of the polysaccharides by the action of depolymerase and thus facilitating access to the secondary receptor present in the surface of the host (8, 39). The presence of a depolymerase that degrades the type K2 capsule has previously been described for phage B3 (18, 40). Although the characteristic phenotype suggests the presence of a depolymerase in the adapted phage 404ad, analysis of the genome did not reveal a depolymerase domain.

Phage cocktails are commonly used in phage therapy to increase the host range and to reduce the emergence of resistance. In this study, a cocktail of phages B3 and 404ad was used to study the combined antimicrobial activity in infection curves (Fig. 2). The infection curves revealed high antimicrobial activity against all the strains tested, except for Ab007_GEIH-2010, at all MOIs tested and with the individual phages and the phage cocktail. However, resistance emerged in all cases, after 5 h of infection with the individual phages and also with the phage cocktail. The interaction between both phages was neutral, as it did not improve or reduce the infectivity of the single phage with the best antimicrobial profile, i.e., phage 404ad. This is a common effect, as described by Niu et al. (2021) in a study testing phage cocktails composed of up to 5 phages, with a neutral effect in 77% of the combinations (41).

Biofilm formation is a virulence factor that hampers access of antimicrobial agents to the bacteria. Biofilms are generally recalcitrant to antibiotics (4), and phages can again be used in alternative therapy. Although bacteria can develop resistance to phages via the production of protective structures such as capsules and biofilm, phages have evolved to overcome these physical barriers (34). Antibiofilm activity of phages has been demonstrated in different bacterial pathogens such as A. baumannii, P. aeruginosa, and Staphylococcus aureus (36, 42–46).

Tests of biofilm removal with phages (Fig. 3) revealed a great antibiofilm activity with individual phages and the phage cocktail. In addition to viable cells, the crystal violet can stain dead cells and EPS, which is the cause of the poor reduction in the biofilm when this technique was employed (47), an effect that was not obtained when only biofilm viable cells were quantified. However, the phage cocktail efficiently degraded the biofilm produced by all strains except one. In a study in which antibiofilm activity of 3 phages was tested in 5 P. aeruginosa strains, it was found that the resistance to the phages emerged 6 h after the infection, with higher production of the biofilm in these mutants 24 h after infection (42). The phage cocktail can evade the resistance mechanisms developed by the host for the individual phages, as infection with different phages can reduce the frequency of bacterial mutations (42, 48, 49).

In this study, an A. baumannii adapted phage, 404ad, derived from a mutant lytic phage, Ab105-2phiΔCI, showed four important recombination events that resulted in the production of four new genes, inversion of two genomic regions, and deletion of 10 genes. These genomic rearrangements mainly affected the tail morphogenesis module, the “hot adaptation site,” host recognition, and consequently the host range. The host range is also determined by the depolymerase-expressing phenotype presented by 81% of the infected strains. Due to the increased lytic spectrum of this phage and its antimicrobial activity, its combination with phage B3, which has a similar host range and a good infectivity, resulted in a phage cocktail of potential value as an antibiofilm phage therapy against multiple strains of the MDR pathogen A. baumannii. The study of genomic features of the adaptation of the phages to the host is important to improve this innovative therapy for personalized treatment.

MATERIALS AND METHODS

Bacteriophage and bacterial strains.

Three bacteriophages were used in this study. The first was bacteriophage Ab105-2phiΔCI (belonging to the family Siphoviridae), a lytic mutant bacteriophage derived from the temperate phage Ab105-2phi (GenBank: KT588075) isolated from the A. baumannii clinical isolate Ab105_GEIH-2010 (5). The second phage, Ab105-2phiΔCI404ad (404ad) (GenBank: MZ514874), is the result of adaptation of the first phage to A. baumannii clinical strain Ab404_GEIH-2010. The third phage used was the podovirus vB_AbaP_B3 (B3) (GenBank: MF033348), previously isolated from sewage samples (18).

The potential hosts used to test the antimicrobial activity of these phages included 29 clinical strains of A. baumannii, belonging to different multilocus sequence types and isolated from Spanish hospitals during the GEIH-REIPI Spanish Multicenter Acinetobacter baumannii Study II 2000–2010 (GenBank Umbrella project PRJNA422585 [https://www.ncbi.nlm.nih.gov/bioproject]) (50), and also the reference A. baumannii strain NIPH2061 (Table 1).

From these isolates, clinical strains Ab177_GEIH-2010 and Ab404_GEIH-2010 were found to be hosts of phages Ab105-2phiΔ and 404ad, respectively. Reference strain NIPH2061 was found to be a host of phage B3.

Bacteriophage isolation, propagation, and purification.

The bacteriophages were propagated using the corresponding host strains. Briefly, the host strains were incubated at 37°C and 180 rpm in Lysogeny Broth (LB) supplemented with 10 μM CaCl₂. The optical density (OD) was measured at a wavelength of 600 nm (OD_600nm), and once a value of 0.2 was reached, the infection was initiated by inoculating the corresponding phage in each culture at a 10 multiplicity of infection (MOI). The cultures were maintained static for 15 min to allow phage adsorption, and they were then incubated at 37°C and 180 rpm until clearance of each culture occurred. The cleared cultures were centrifuged at 3000 rpm for 15 min, and the supernatants were recovered, treated with chloroform 1%, and filtered (20 or 45 μm).

Phage filtrates were washed and concentrated in a 100K AmiconUltra-15 filter device (Merck Millipore Ltd.) as previously described (51). The phage isolate was suspended in SM buffer (0.1 M NaCl, 1 mM MgSO4, 0.2 M Tris-HCl, pH 7.5).

Bacteriophage host range (spot test and efficiency of plating).

The host range for each phage was determined by the spot test (52). Thirty strains of A. baumannii were tested as putative hosts for each of three phages (Table 1). Briefly, double agar layer plates were prepared with the putative host strain mixed with the soft upper agar layer (0.4% agar). Once solidified, a 10-μL drop of the isolate phage was deposited in the top layer of agar, and the plates were incubated at 37°C for 20 h. The presence of a clear spot indicated that the strain is a potential host.

The host range determined by spot test was confirmed by Efficiency of Plating (EOP). The phages were titrated against those strains yielding a positive spot, and the EOP was calculated as the ratio of titer for the tested strain and for the host strain (53).

Phage strain adaptation by a modified Appelmans method.

In order to adapt a phage to a host strain, Appelmans protocol was applied, with minor modifications (16). Briefly, 10 μL of a serial dilution of the Ab105-2phiΔCI phage was added to 90 μL of LB medium in the wells of a 96-well microtiter plate, with a culture of the strain Ab404 at 0.1 OD_600nm. The plates were incubated at 37°C for 24h, and the supernatant was recovered and treated with 1% chloroform. The infection procedure was repeated with the recovered lysate until a clear well was obtained. Finally, the supernatant was recovered, treated with 1% chloroform, and filtered. The new adapted phage was quantified using the double agar overlay plaque assay (54).

Storage stability of phages.

The stability of the phages in storage conditions was tested by maintaining both phages, 404ad and B3, at 4°C and −20°C for 1 month. The phages were maintained in SM buffer when stored at 4°C and diluted 1:1 in sterile glycerol when stored at −20°C (55). Before and after the storage period, the phage titer was calculated by serially diluting the phage stock and quantified in double agar layer plates. A Student's t test (GraphPad Prism v.9.3.0) was used to determine any statistically significant differences between the titers.

Phage DNA extraction and whole genome sequencing.

The phage genomic DNA was extracted following the phenol:chloroform:isoamyl alcohol/sodium dodecyl sulfate (PCI/SDS) DNA extraction method (https://phagesdb.org/media/workflow/protocols/pdfs/PCI_SDS_DNA_Extraction_2.2013.pdf). Once extracted, the DNA quality and concentration were measured in a Nanodrop ND-10000 spectrophotometer (NanoDrop Technologies, Waltham, MA, USA). Genomic libraries were prepared using the Nextera XT Library prep kit (Illumina), and the distribution of fragment lengths was checked using the Agilent High Sensitivity DNA kit, in an Agilent 2100 Bioanalyser. Libraries were purified using Mag-Bind RXNPure plus magnetic beads (Omega Biotek), and the pool was then sequenced in Miseq platform (Illumina Inc., USA). The quality of the FASTQ files was checked using FastQC software (56) and summarized using MultiQC (57). De novo sequence assembly of 300 bp paired-end reads of each isolate was conducted with Spades V.3.15.2 (58). The assembly was annotated using Patric 3.6.9 (59), RAST 2.0 (60), BLASTX (http://blast.ncbi.nlm.nih.gov), HMMER 3.3 (http://hmmer.org), and HHpred (61).

Genomic analysis.

The genome sequence obtained and that of the ancestral phage were compared using BLASTn (http://blast.ncbi.nlm.nih.gov). The proteins were compared using BLASTp, HMMER 3.3, and Easyfig 2.2.5 (62).

Phylogenetic study.

A homology study of the phage 404ad was done comparing its genome with the viral database of the GenBank. Those genomes with an identity higher than 1% were selected and aligned using MAFFT v7.407 (63) with default options. The phylogenetic analysis was performed in RaxmlHPC-PTHREADS-AVX2 version 8.2.12 (64) under the GTRGAMMA model and 100 bootstrap replicates. FigTree (http://tree.bio.ed.ac.uk/software/figtree/) was used to visualize the phylogenetic tree.

Single phage and phage cocktail infection curves (antimicrobial activity).

Infection curves were constructed for phage 404ad, phage B3, and the combination of both (at a proportion of 1:1 in a phage cocktail) and at MOI values of 0.1, 1 and 10. The phage cocktail was prepared just before the infections.

An overnight culture of A. baumannii clinical strain Ab404_GEIH-2010 was diluted 1:100 in LB broth supplemented with 10 μM CaCl₂ and incubated at 37°C and 180 rpm until an early logarithmic phase was achieved (0.2 OD_600nm). Subsamples of the culture were then infected with each of the individual phages and the phage cocktail, and the suspensions were maintained static for 15 min to allow phage adherence to the bacteria. The infected cultures were then incubated at 37°C and 180 rpm for 5 h, and the OD of the culture was measured periodically at one hourly intervals.

The infection curves were constructed for each individual phage and the phage cocktail, at an MOI of 10, using as hosts 6 selected strains of A. baumannii in which the phages produced a halo: Ab007_GEIH-2010; Ab008_GEIH-2010; Ab019_GEIH-2010; Ab034_GEIH-2010; Ab404_GEIH-2010; and NIPH 2061. All measurements were made in triplicate.

Frequency of occurrence of phage-resistant mutant bacteria.

The frequency of occurrence of phage-resistant mutant bacteria was done as previously described by Merabishvili et al. (2014) (53), with some modifications. Briefly, an overnight culture of each strain of A. baumannii was diluted 1:100 in LB and grown to 0.5 OD_600nm. An aliquot of 100 μL of the culture containing 10⁷ CFU/mL was mixed with 100 μL of 10⁹ PFU/mL and plated in LB agar. The plates were incubated at 37°C for 24h to 48h, and the number of CFU was counted. The frequency of occurrence of phage resistant mutants was calculated by dividing the number of resistant bacteria by the total number of sensitive bacteria.

Phage biofilm removal assay.

The biofilm removal assays were conducted in the biofilm produced by 5 clinical strains and 1 reference strain of A. baumannii (the same as that indicated in the previous section) following a previously described procedure (5, 50). An overnight culture of the A. baumannii strain was diluted to OD_600nm of 0.1 and inoculated in 100 μL of LB in the wells of a 96-well microplate, with 16 wells for each strain and condition. After 24 h, the biofilm produced was treated with the individual phages and the phage cocktail (prepared just before the infection), which were added to the wells at an MOI of 10. The medium was removed and discarded and the wells were washed with phosphate-buffered saline (PBS) and filled with 100 μL of SM buffer, and each phage (B3 and 404ad) or the phage cocktail was added at an MOI of 10. Control wells were prepared by addition of LB (100 μL) supplemented with 10 μM CaCl₂. The plates were then incubated for 5 h. Finally, the supernatant was discarded and the wells were washed with PBS. Half of the wells were used to quantify the CFU forming the biofilm, and the other half were used to quantify the biofilm. For quantification of the biofilm by the crystal violet method, 100 μL of methanol was added to each well and was then discarded after 10 min. Once the methanol had completely evaporated, 100 μL of crystal violet (0.1%) was added and the plates were incubated for 15 min. Finally, the wells were washed with PBS, 150 μL of acetic acid (30%) was added to resuspend the crystal violet adhered to the biofilm, and the absorbance was measured at OD_{595 nm}. For the quantification of the CFU in the biofilm, 100 μl of PBS was added to the wells and the plates were agitated for 10 min and sonicated for another 5 min. The suspension was serially diluted and plated on LB plates.

A Student's t test (GraphPad Prism v.9.3.0) was used to determine any statistically significant differences in biofilm removal.