Host Range Expansion of Pseudomonas Virus LUZ7 Is Driven by a Conserved Tail Fiber Mutation

Abstract

Background: When subjected to phage infection, bacteria can rapidly become resistant by changes in the phage receptors at the bacterial surface. Phages thus require adaptive mechanisms to circumvent this type of resistance.

Methods: LUZ7 phage with an altered host range were isolated and analysed for mutations and their effect.

Results: We find that Pseudomonas virus LUZ7 has an unusually high number of mutants (0.01–0.1% of the population) that drive host range expansion. Interestingly, all tested mutants have a single D737Y mutation in the tail fiber. This mutation allows the phage to adsorb to P. aeruginosa strains that are not natively recognized by the wild-type phage.

Conclusion: The high number and specificity of mutants suggests the presence of an uncharacterized mechanism that drives these mutations. This mechanism enables the phage to better evade host resistance at the surface level and expand its host range in general, a feature that could be valuable in phage therapeutic settings or for phage engineering.

Introduction

A continuous evolutionary arms race exists between bacteriophages and their hosts, leading to a wide array of mechanisms by which bacteria become resistant to phages. Examples of this include CRISPR-Cas systems, abortive infection mechanisms, restriction modification, and the more recently described BREX and DISARM systems.^1–5 Aside from these dedicated antiphage defenses, perhaps the most straightforward way by which bacteria become resistant to phage is the alteration, disappearance, or masking of phage receptors. Indeed, mutations in genes that encode for receptor proteins (such as pili and outer membrane proteins) or receptor biosynthesis proteins (such as those from LPS biosynthesis) are widely reported to confer resistance to phage infection.^6–8 To cope with these receptor alterations, phages require sufficient variability in their receptor binding proteins. Here we find that Pseudomonas virus LUZ7 has such a naturally occurring variability that presents itself as a single amino acid substitution in the tail fiber. This variability improves its infectivity on non-/less susceptible strains and could prove a valuable feature for phage therapy applications.

Materials and Methods

Phage amplification and isolation of mutants

For amplification of LUZ7, Pseudomonas aeruginosa strain US449 was grown in Lysogeny Broth (LB) at 37°C to an optical density at 600 nm (OD₆₀₀) of 0.3 and infected with a LUZ7 stock.⁹ Phage lysates were treated and purified with polyethylene glycol exactly as described elsewhere.¹⁰ LUZ7 mutants were isolated by directly picking up clear plaques of a P. aeruginosa PAO1 agar overlay (see also Results section) and inoculate them in a PAO1 culture for amplification as already described.

Phage spot test

A double agar overlay was made with indicated strains embedded in the top (0.75%) agar layer. A 100-fold dilution series of the phage in phage buffer (10 mM Tris-HCl, pH 7.5; 10 mM MgSO₄; 150 mM NaCl) was spotted on the agar overlay. Plates were incubated at 37°C overnight.

Infection curves

Cultures of indicated P. aeruginosa strains were grown in LB at 37°C to OD₆₀₀ of 0.3 at which they were infected with LUZ7 at multiplicity of infection (MOI) of 20. OD₆₀₀ of the culture was tracked throughout infection.

Adsorption assay

Adsorption of phage was tested according to the protocol of Kropinski.¹¹ Cells were grown in LB medium at 37°C and infected at OD₆₀₀ of 0.3 with an MOI of 0.001.

Genome sequencing

DNA was isolated from phage particles by phenol/chloroform extraction. Illumina libraries were prepared with a MiSeq Reagent Nano Kit v2 with barcoding. It was sequenced on an Illumina MiSeq using a paired-end approach (2 × 150 bp). Quality control was done with FastQC¹² and BBDuk. Quality controlled reads were mapped using BWA¹³ and resulting alignment was processed using Bowtie2.¹⁴ Variants were called using FreeBayes¹⁵ and effect of mutation was by using snpEff.¹⁶ Sanger sequencing of orf56 mutation was done by Eurofins Genomics (Germany) with primer 5′-cagacacctgcgttgttc-3′.

Results

Pseudomonas virus LUZ7 was originally isolated on P. aeruginosa strain US449.^9,17 This strain gets infected very efficiently, as seen by formation of clear plaques when LUZ7 is spotted on an LB double-agar overlay with this strain (Fig. 1A). In contrast, infection on the standard laboratory strain PAO1 results in turbid plaques, indicating impaired infection (Fig. 1A). However, a small number of clear plaques (∼0.01–0.1% of total plaques) were observed among the turbid plaques. In liquid LB medium, the discrepancy between infection of LUZ7 on US449 and PAO1 strains is even more apparent. Infection of PAO1 with a wild-type LUZ7 stock at high MOI ( = 20) results in a markedly slow reduction of cell density, with first signs of lysis occurring after >3 h. This is contrary to the efficient clearing of the US449 culture, visible within 30–40 min (Fig. 1B). The slow lysis on PAO1 cultures could be because wild-type LUZ7 is unable to properly infect and instead a small amount of mutant phage (visible as clear plaques in the spot test) gradually takes over the culture. To validate this, a fully lysed culture after single passage of LUZ7 on PAO1 was filtered over 0.22 μm and the resulting phage lysate was spotted on both US449 and PAO1 (Fig. 1 A). No more turbid plaques could be observed and all plaques demonstrated the same clear morphology as wild-type LUZ7 on both strains. Hence after single passage, LUZ7 acquired the capacity to infect both strains efficiently, indicating that the mutant phage population indeed took over the wild-type population.

FIG. 1.

(A) Plaque morphology on US449 and PAO1 strains from wild-type LUZ7 (left) or after passage on a PAO1 strain (right). A 100-fold dilution series of the phage stock was spotted from left to right. (B) Infection by wild-type LUZ7 in liquid LB at MOI 20 on either PAO1 or US449 strains in comparison with an uninfected control. OD₆₀₀ of each culture was tracked for a 200 min timeframe (n = 3). Arrow indicates the first sign of lysis on PAO1.

To determine which genomic changes occurred compared with wild-type LUZ7, phages from three individual plaques were isolated and sequenced by Illumina. The wild-type LUZ7 used for generating mutants was also resequenced as control. Interestingly, all three mutant LUZ7 (LUZ7-m) isolates had the same single mutation in their genome with no other mutations observed throughout the genome (Fig. 2). This guanine to thymine mutation results in an aspartate to tyrosine change (D737Y) in the putative tail fiber protein of the phage. Two additional isolates for which this gene was sequenced with Sanger sequencing also showed this same mutation. Since tail fibers are generally required for recognition and binding of the host cell,¹⁸ the poor infectivity of LUZ7 on PAO1 might be due to a failure to adsorb efficiently. In addition, the D737Y mutation results in a significant alteration of this amino acids' sidechain properties, from negatively charged to hydrophobic. This change could be the direct cause of the improved infectivity by LUZ7-m phages. To verify whether the differences in infectivity are due to aberrant adsorption, the adsorption efficiency of LUZ7 to US449 and PAO1 strains was tracked for a 10-min timeframe. No adsorption to PAO1 could be observed, whereas there is a clear adsorption to the US449 strain (Fig. 3). Moreover, LUZ7-m phages adsorb to PAO1 at the same rate as wild-type LUZ7 to US449. These mutant phages have thus gained the ability to adsorb to PAO1 without losing their ability to infect US449 by means of a single tail fiber mutation.

FIG. 2.

Location of orf56 mutation in the LUZ7 tail cassette. Arrows represent the different open reading frames in the tail cassette of LUZ7 with their respective orientation in the genome. Putative LUZ7 mutants (LUZ7-m) displaying an increased infectivity on PAO1 were sequenced. The whole genome of three mutants (isolated from different plaques) was sequenced by Illumina (m1–m3). This showed a single mutation in the orf56 gene (nucleotide position 30,426 in the genome), of which a short sequence stretch on the coding strand is given here, together with its translation (mutation in bold). Two additional mutants (m4 and m5) were sequenced only in the orf56 gene with Sanger sequencing. The mutants were compared with the reference sequence of LUZ7 on NCBI (NC_013691) and a freshly resequenced wild-type LUZ7.

FIG. 3.

Differences between adsorption of LUZ7 and LUZ7-m on a US449 and a PAO1 strain. (A) Adsorption of LUZ7 (at MOI 0.001) on PAO1 and US449 is indicated as the amount of viable (nonadsorbed) phage in the sample relative to the start amount (P/P0) over a time course of 10 min (n = 4 for LUZ7, n = 2 for LUZ7-m).

Discussion

Single point mutations in the tail fiber that alter or expand the host range have been described for other Pseudomonas phages.¹⁹ However, the number of LUZ7 mutants (∼0.01–0.1%) found here is significantly higher than anticipated from just random mutations. The wild-type phage can plaque efficiently on strain US449, which was used for its propagation, and thus there should not be a selection for broader host range LUZ7 mutants. All tested mutants also have the same amino acid substitution. The high rate and specificity of mutation suggests that there is a targeted yet unknown mechanism from LUZ7 for their generation. In the genus of phiKMVvirus, unrelated to LUZ7, such a system is known to exist. In these phages, a strong increase of SNP occurrences in structural, mostly tail fiber, proteins was observed. This was caused by error-prone DNA polymerase activity that seems to be higher in regions required for adsorption.²⁰

From a biological perspective, the introduction of mutations in the tail fiber allows the phage to increase its chances of generating progeny that is able to infect a broader/different set of hosts. LUZ7 is thought to target LPS for attachment to the host.⁹ P. aeruginosa strains have a large versatility of LPS types with 20 known O-antigen serotypes.²¹ Moreover, the bacterium readily adapts when challenged with phages and the alteration of the receptor occurs to provide resistance against these phages.²² To prevent that a new generation of LUZ7 progeny can no longer infect on these altered bacteria, LUZ7 seems to have a mechanism to generate variability in the tail fiber. This mechanism resembles phase variation systems of bacteria, in which rapid changes in mainly surface structures allows avoiding immune responses or phage infection.²³ Although the effects of the tail fiber mutation have only been shown in a single strain here, the high number of mutants that is naturally present in LUZ7 stocks due to this mechanism can be advantageous for phage therapy applications. It expands the range of hosts that LUZ7 preparations can efficiently infect and should allow it to better circumvent resistance that occurs during treatments.

Data Availability Statement

The sequencing data of Pseudomonas virus LUZ7 and its three mutants are available in the NCBI Sequence Reads Archive database under accession numbers SRR11409997, SRR11409996, SRR11409995, and SRR11409994, respectively (BioProject PRJNA614983).

Authorship Confirmation Statement

M.B. and D.H. conceived and performed experiments. R.L. conceptualized and directed the research. C.L. performed data analysis. M.B wrote the article. D.H, C.L., R.L. and V.v.N. revised the article. All authors have reviewed and approved the article before submission. This article has not been previously nor simultaneously submitted to another journal.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

M.B. and R.L. were funded by KU Leuven as part of the Geconcerteerde Onderzoeksacties (GOA) program “Phage biosystems” [3E140356]. D.H. and C.L. received funding from Research Foundation Flanders (FWO) [1S02520N and 1S64720N, respectively].