Construction and characterization of bacterial artificial chromosomes harboring the full-length genome of a highly attenuated vaccinia virus LC16m8

Abstract

LC16m8 (m8), a highly attenuated vaccinia virus (VAC) strain, was developed as a smallpox vaccine, and its safety and immunogenicity have been confirmed. Here, we aimed to develop a system that recovers infectious m8 from a bacterial artificial chromosome (BAC) that retains the full-length viral genomic DNA (m8-BAC system). The infectious virus was successfully recovered from a VAC-BAC plasmid, named pLC16m8-BAC. Furthermore, the bacterial replicon-free virus was generated by intramolecular homologous recombination and was successfully recovered from a modified VAC-BAC plasmid, named pLC16m8.8S-BAC. Also, the growth of the recovered virus was indistinguishable from that of authentic m8. The full genome sequence of the plasmid, which harbors identical inverted terminal repeats (ITR) to that of authentic m8, was determined by long-read next-generation sequencing (NGS). The ITR contains x 18 to 32 of the 70 and x 30 to 45 of 54 base pair tandem repeats, and the number of tandem repeats was different between the ITR left and right. Since the virus recovered from pLC16m8.8S-BAC was expected to retain the identical viral genome to that of m8, including the ITR, a reference-based alignment following a short-read NGS was performed to validate the sequence of the recovered virus. Based on the pattern of coverage depth in the ITR, no remarkable differences were observed between the virus and m8, and the other region was confirmed to be identical as well. In summary, this new system can recover the virus, which is geno- and phenotypically indistinguishable from authentic m8.

Fig 1. Scheme to generate a VAC-BAC plasmid, pLC16m8-BAC.

A recombinant VAC, m8-EGFP-BAC, was generated by homologous recombination accelerated by a CRISPR-Cas9 system (A). pVAC-BAC11 harbors an EGFP expression cassette driven by an early and late VAC promoter and a mini-F cassette, which is essential for plasmid maintenance in E. coli. The cassettes were flanked by the front (B5R-F) and rear (B5R-R) part of the VAC B5R gene. Note that the gRNA targeted the sequence on the m8 genome but not on the pVAC-BAC11. m8-EGFP-BAC was used to infect RK13 cells (B). The virus genome formed replication intermediates, concatemers, some of which were circularized was purified from the infected RK13 cells with the addition of IβT. An electrocompetent E. coli was transformed by a circularized virus genome (arrows) of the concatemerized genome, and the chloramphenicol-resistant cells were cloned (C). The VAC-BAC plasmids obtained from the clones were named pLC16m8-BAC.

Fig 2. Growth of the recovered virus from pLC16m8-BAC named vLC16m8-BAC.

RK13 cells were infected with either authentic m8, m8-EGFP-BAC, or vLC16m8-BAC at an MOI of 0.01. The infected cells along with the culture media at the indicated d.p.i. were collected and freeze-thawed. The amount of the virus was determined by a standard plaque assay.

Fig 3. Generation of the m8-BAC system, which enables recovery of the bacterial replicon-free m8.

The scheme to modify pLC16m8-BAC to pLC16m8.8S/8.OS-BAC and then to pLC16m8.8/8.O-BAC is shown. The first recombination was performed to insert PCR products of I-SceI site and a gene expression cassette conferred kanamycin resistance (Kan^R) with the rear part of B5R (B5R-R) harbored by flanks homologous to parts of B5R-F and EGFP (α and β in the rectangles) cassette into pLC16m8-BAC (A). The flank α contained either of the nucleotide sequences that resulted defective (m8 type) or functional (mO type) B5R gene product. The second red recombination, which was achieved by induction of I-SceI digestion system with the addition of L-arabinose was performed to make both pLC16m8.8S-BAC and pLC16m8.OS-BAC by excision of the Kan^R. Either m8 type or mO type B5R gene product was expressed, when the virus was recovered from either the pLC16m8.8S-BAC or pLC16m8.OS-BAC, and the EGFP and mini-F cassettes are expected to be self-excised by the homologous recombination while recovering the viruses (B). pLC16m8.8-BAC and pLC16m8.O-BAC were made from pLC16m8.8S-BAC and pLC16m8.OS-BAC, respectively (C). The first Recombination was performed to remove the B5R-R from either pLC16m8.8S-BAC or pLC16m8.OS-BAC by substitution of a PCR product. The PCR product harbored a gene expression cassette conferred zeocin resistance (Zeo^R), which was flanked by homologous to part of B5R-R and the untranslated region besides B5R-R (γ and δ in the rectangles). Both pLC16m8.8-BAC and pLC16m8.O-BAC were made by excision of the Zeo^R by the second red recombination. The virus recovered from pLC16m8.8-BAC or pLC16m8.O-BAC is expected to stably harbor the EGFP and mini-F cassettes. The target region of primer A, B, and C used for following the PCR are indicated.

Fig 4. Restriction fragment analysis of BAC plasmids.

Prediction of NcoI digested fragment patterns based on the sequences of pLC16m8-BAC and the derivatives were shown (A). Purified BAC plasmids, pLC16m8-BAC (m8-BAC), pLC16m8.8S/8.OS-BAC (8.8S or 8.OS) and pLC16m8.8/8.O (8.8/8.O) were digested with NcoI and separated on a 0.75% agarose gel (B). Color of the gel was inverted. Sizes of a molecular weight marker are shown as marker, and the sizes are given. Changes in the restriction pattern due to the presence of EGFP, mini-F cassettes or rear part of B5R are indicated with a red arrow.

Fig 5. The characteristics of the viruses recovered from the pLC16m8.8S-BAC or pLC16m8.OS-BAC.

The plaque size and EGFP expression were observed under phase-contrast and fluorescent microscopy. RK13 cells were infected with the recovered virus stocks from either pLC16m8.OS-BAC (vLC16m8.OS-BAC) or pLC16m8.8S-BAC (vLC16m8.8S-BAC). The plaque size and EGFP expression of vLC16m8.OS-BAC (plaque A and B) and vLC16m8.8S-BAC (plaque C and D) were observed at the same magnification under phase-contrast with fluorescent microscopy at 2 d.p.i.

Fig 6. Confirmation of self-excising EGFP and mini-F cassettes from the virus genome.

The viral genomic DNA purified from the RK13 cells infected with either vLC16m8.O-BAC (crude viruses) (8.O), vLC16m8.8-BAC (crude viruses) (8.8), vLC16m8.OS-BAC (crude or an EGFP-negative plaque cloned viruses) (8.OS), or vLC16m8.8S-BAC (crude or an EGFP-negative plaque cloned viruses) (8.8S) was amplified by conventional PCR using either the primer set A/B or A/C. The target position of the primers is shown in Fig 3A and 3B. The no template control (NTC) monitors unintentional PCR amplification to produce a false-positive result.

Fig 7

Growth of the recovered viruses in comparison with that of authentic m8 (A) and mO (B). RK13 cells were infected with either m8, vLC16m8.8-BAC (crude viruses), vLC16m8.8S-BAC (EGFP-negative plaque cloned), mO, vLC16m8.O-BAC (crude viruses), or vLC16m8.OS-BAC (EGFP-negative plaque cloned) as indicated at an MOI of 0.01.

Fig 8

Coverage depth of the pLC16m8.8S-BAC aligned to the consensus sequence of pLC16m8.8S-BAC generated by de novo assembly using HGAP software (A). The ITR regions were not arranged at the termini of the sequence for the purpose of accurate reference-based alignment. The ITR left to the right region flanked by the C23L genes is indicated by vertical dotted lines. Nucleotide position 1 is at the terminus of the ITR left. The structure of the ITR left region to the C23L gene of pLC16m8.8S-BAC (accession no. LC315596), VAC strain WR (accession no. AY243312), Copenhagen (accession no. M35027), and MVA (accession no. U94848) are shown (B). Gray rectangles indicate unique sequences in the ITR: non-repeated region (NR) I, II, III, and the C23L gene. The T in the ocher rectangle indicates a terminal hairpin. Dark or light blue rectangles correspond to 69 or 70 base pair tandem repeats. Red and yellow rectangles correspond to 53 or 54 base pair tandem repeats. Green rectangles correspond to 125 base pair repeats. Purple rectangles correspond 111 or 112 base pair repeats. The number of repeats in the left or right ITR is indicated below the repeats.

Fig 9. Coverage depth through the ITR left to the right region flanked by the C23L gene of pLC16m8.8S-BAC computed by HGAP (black line) or by Canu (red line).

Fig 10. Identity of the ITR region in authentic m8 and 8.8S-clone4, a clone that was plaque-purified from EGFP-negative and bacterial replicon-free vLC16m8.8S-BAC.

The coverage depth computed by the alignment of short-reads to the sequence of the left ITR region, which was extracted from the consensus sequence of pLC16m8.8S-BAC is shown (A and D). The coverage depth in the ITR left region from the terminus (position 1) to the C23L gene of m8 or 8.8S-clone 4 was plotted as either a black or a red line. The mean of the coverage depth by each of the regions (i.e., 70 base pair repeat, non-repeated region II to 125 base pairs, 54 base pair repeat, non-repeated region III, and C23L) was plotted (B, C, E, and F). One-way ANOVA with Bonferroni's multiple-comparison test was used to determine the level of statistical significance. The effect size (r), which indicates the magnitude of the mean difference between the regions, was also calculated. The mean of coverage depth at the 70 base pairs repeats were calculated from position 474 to 1423 to exclude the effect of low coverage depth artificially occurred at the terminus of the reference sequence. The number of 54 base pair tandem repeats in the reference sequence was artificially modified from x 41 (A-C) to x 23 (D-E). “n.s.” indicates "not significant”.