Engineering Modular Viral Scaffolds for Targeted Bacterial Population Editing

Abstract

Bacteria are central to human health and disease, but existing tools to edit microbial consortia are limited. For example, broad-spectrum antibiotics are unable to accurately manipulate bacterial communities. Bacteriophages can provide highly specific targeting of bacteria, but assembling well-defined phage cocktails solely with natural phages can be a time-, labor- and cost-intensive process. Here, we present a synthetic-biology strategy to modulate phage host ranges by engineering phage genomes in Saccharomyces cerevisiae. We used this technology to redirect Escherichia coli phage scaffolds to target pathogenic Yersinia and Klebsiella bacteria, and conversely, Klebsiella phage scaffolds to target E. coli by modular swapping of phage tail components. The synthetic phages achieved efficient killing of their new target bacteria and were used to selectively remove bacteria from multi-species bacterial communities with cocktails based on common viral scaffolds. We envision that this approach will accelerate phage-biology studies and enable new technologies for bacterial population editing.

Figure 1. Yeast platform for phage engineering

(A) Schematic illustrating the workflow to capture and reboot phages using our yeast platform. An entire phage genome or PCR products spanning an entire phage genome are transformed into yeast cells along with a linearized yeast replicon fragment from the yeast artificial chromosome, pRS415. In yeast, the phage genome is assembled and captured in the YAC by gap-repair cloning. The resulting YAC-phage DNA is extracted and transformed into host bacteria. Active phages are produced from the YAC-phage DNA and generate plaques on a lawn of host bacteria. (B) High-titer phage lysates (>10⁹ PFU/ml) were spotted onto E. coli 10G lawns, which were sensitive only to T7 and T3 phages. (C) Rebooting T7 and T3 phages from purified phage genomes (10 ng) after electroporation into 10G, recovery in LB, chloroform sterilization, and plating with E. coli BL21. (D) All tested phage genomes (10–200 ng), including non-E. coli phages, could be rebooted by electroporation into E. coli 10G cells and plating supernatants with host bacteria. Host bacteria: IJ1668 K-12 hybrid; K1 capsule for K1E, K1F, and K1-5 phages, IJ612 Salmonella typhimurium LT2 for SP6 phage, Pseudomonas aeruginosa PAO1 for LUZ19 phage, Pseudomonas putida C1S for gh-1 phage, and Klebsiella sp. 390 for K11 phage. (E) An example of capturing and rebooting a phage through the yeast platform. An excised YAC pRS415 amplicon and the T7 genome were co- transformed in yeast cells. The T7 genome was captured in the YAC by gap-repair cloning in yeast. Progeny phages were produced from YAC-T7 DNA via E. coli 10G and generated plaques on E. coli BL21.

Figure 2. Creation of synthetic phages with engineered host range

(A) We prepared multiple PCR fragments encoding the wild-type T7 phage genome (T7_WT) (PCR fragments 1 – 4; 10.0, 10.0, 10.0, and 10.1 kbp, respectively), T7 phage with the entire T3 phage tail fiber (T7_T3(gp17)) (PCR fragments 1 – 3 and 5 – 7; 10.0, 10.0, 10.0, 4.7, 1.7, and 3.7 kbp, respectively), and T7 phage with a hybrid T7-T3 tail fiber (T7_T3(C-gp17)) (PCR fragments 1 – 3, 8, 9, and 7; 10.0, 10.0, 10.0, 5.2, 1.3, and 3.7 kbp, respectively). All fragments were co-transformed and assembled in yeast along with YAC DNA (3.0 kbp). (B) Phage A with its primary host determinant, gene a, infects bacteria A, but cannot infect bacteria B. Phage B with its primary host determinant, gene b, infects bacteria B, but cannot infect bacteria A. We hypothesized that swapping these host determinants between phages would switch their respective host ranges. (C) Creation of synthetic T7 phage with phage 13a tail fiber (encoded by gene 17). We synthesized 13a’s gene 17 and assembled it with the rest of the T7 genome via overlapping PCR products in yeast (PCR fragments 1 – 6 and YAC amplicon; 10.0, 10.0, 10.0, 4.8, 1.7, 3.7, and 3.0 kbp, respectively). The YAC-phage DNA was extracted and used for transformation.

Figure 3. Efficiency of plating (EOP) of natural, reconstructed wild-type, and synthetic phages

EOP was determined with respect to a reference bacterium. Klebsiella sp. 390 was used as a reference for T7_{K11(gp11-12-17)}, K11, and K11_WT. Y. ptb IP2666 was used for T3_R(gp17). For other phages, BL21 was used as the reference host. EOP data were log₁₀-transformed and are presented as the mean of three independent experiments. Error bars represent SD. The “0” symbol indicates an EOP = 1, which marks the reference strain used for the EOP calculations within each phage. * indicate that the EOP was below the detection limit (EOP <10⁻⁸). Strain abbreviations: B, E. coli BL21; G, E. coli 10G; D, E. coli DH5α; W, E. coli BW25113; M, E. coli MG1655; 4, E. coli ECOR4; 13, E. coli ECOR13; 16, E. coli ECOR16; K, Klebsiella sp. 390; I, Y. ptb IP2666; Y, Y. ptb YPIII; N, E. coli Nissle 1917.

Figure 4. Creation of synthetic T3 phage with Yersinia phage R tail fiber

(A) We introduced mutations in T3 gene 17 by PCR to convert it into phage R gene 17 and assembled the resulting product with the rest of the T3 genome and YAC DNA in yeast (PCR products 1 – 6 and YAC amplicon; 10.0, 10.0, 10.0, 4.4, 0.2, 3.8, and 3.0 kbp, respectively). The YAC-phage DNA was extracted and used for transformation into E. coli. (B) Plaquing assay with T3_WT and T3_R(gp17) on E. coli BL21, Y. ptb IP2666, and Y. ptb YPIII shows that T3_R(gp17) can infect both E. coli and Y. ptb. Ten-fold serial dilutions of phage lysates were spotted on bacterial lawns and incubated for 4 h at 37°C for E. coli BL21 or 24 h at 30°C for Y. ptb strains. These pictures were cut out from Figure S2. Bottom panels show images of individual plaques. NP, no plaque. (C) Killing curves of Y. ptb IP2666 treated with T3_R(gp17). ~10⁸ CFU/ml bacteria and ~10⁷ PFU/ml phage were used (MOI ~0.1). The data are presented as the mean of three independent experiments. Error bars represent SD. Small error bars are obscured by symbols. The detection limit was 2.0 × 10³ CFU/ml.

Figure 5. Creation of synthetic T7 phage with Klebsiella phage K11 tail components as well as K11 phage with T7 tail components

(A) The tail complex of T7 phage is composed of two components, a tubular structure and tail fibers. The tubular structure consists of an upper dodecameric ring made of adaptor protein gp11 and a pyramidal hexameric complex of the nozzle protein gp12. The tail fiber protein gp17 interacts with the interface between gp11 and gp12 (Cuervo et al., 2013). Schematics illustrating the construction of synthetic hybrids between phages T7 and K11. Whole genomes were amplified as overlapping PCR amplicons. PCR fragments were co-transformed and assembled in yeast (PCR fragments 1 – 8 and YAC amplicon; 10.0, 10.0, 4.3, 3.0, 2.7, 4.8, 2.7, 3.7, and 3.0 kbp, respectively). YAC-phage genomes were extracted and used for transformation. We swapped K11 genes 11, 12 and 17 into T7 to create T7_{K11(gp11-12-17)} and T7 genes 11, 12 and 17 into K11 to create K11_{T7(gp11-12-17)}. (C) Plaquing of T7_{K11(gp11-12-17)} and K11_{T7(gp11-12-17)} on E. coli BL21 and Klebsiella sp. 390. Ten-fold serial dilutions of phage lysates were spotted on bacterial lawns and incubated for 4 h at 37°C. These pictures were cut from Figure S3. Bottom panels show images of individual plaques. NP, no plaque. (D) Killing curves of Klebsiella sp. 390 treated with T7_{K11(gp11-12-17)}. ~10⁸ CFU/ml bacteria and ~10⁷ PFU/ml phage were used (MOI ~0.1). The data are presented as the mean of three independent experiments. Error bars represent SD. Small error bars are obscured by symbols. The detection limit was 2.0 × 10³ CFU/ml.

Figure 6. Microbial population editing assay

A synthetic microbial community composed of E. coli Nissle 1917, Klebsiella sp. 390, and Y. ptb IP2666 was treated with various individual synthetic phages and a pairwise combination of phages. After adding ~10⁷ PFU/ml of each phage, the resulting samples were incubated at 30°C with shaking for 1 h. At each time point, bacteria were collected, washed in saline, serially diluted, and plated onto selective plates for viable cell counts after a 24 h incubation at 30°C. The data are presented as the mean of three independent experiments and the total numbers of cells (CFU/ml) are shown. The sizes of the pie charts reflect the total number of cells. Note that the chart does not allow the display of fractions smaller than ~1%. The detailed data and the SD are shown in Table S4A. The detection limit was 2.0 × 10³ CFU/ml.