Successful Rescue of Synthetic AcMNPV with a ~17 kb Deletion in the C1 Region of the Genome

Baculoviruses have been widely used as expression vectors. However, numerous genes in the baculoviral genome are non-essential for cellular infection and protein expression, making the optimisation of baculovirus expression vectors possible. We used a synthetic biological method to reduce the number of genes in a partial region of the autograph californica multiple nucleopolyhedrovirus (AcMNPV), the most widely used baculovirus expression vector. The C1 region of the AcMNPV is 46.4 kb and is subdivided into B1, B2, and B3 fragments. We first designed modified B1, B2, and B3 fragments by deleting the non-essential genes, and then synthesised complete viral genomes containing either individual modified B fragments or joint modified B fragments through transformation-related recombination in yeast. The synthetic genomes were then transfected into Sf9 cells to rescue the progeny viruses and test their infectivity. The design-build-test cycle was repeated until the ultimately rescued virus could produce progeny viruses efficiently. Finally, AcMNPV-Syn-mC1-1.1 by deleting approximately 17.2 kb, including 20 ORFs, in the C1 region, was obtained. This is essential to the synthesis of a minimal AcMNPV genome that can generate infectious progeny viruses and can be further used to optimise the foundation of baculovirus expression vectors.


Introduction
Baculoviruses are insect-specific pathogens with a circular double-stranded DNA (dsDNA) genome and have been widely used as biological control agents against fundamental agricultural and forestry pests [1,2]. In addition, they are potent expression vectors owing to their high expression, high abundance, and relatively complete post-translational modification of proteins. Since the first report of human IFN-β expression using the baculovirus expression vector system (BEVS) in 1983 [3], many proteins have been expressed via this system. Moreover, certain biological products have been approved by the Food and Drug Administration and marketed. For example, FluBlok ® and Cervarix ® vaccines were produced using baculovirus expression systems to guard against influenza and human papilloma virus, respectively, and have been commercialised [4]. In addition, an AAV vector was produced from Glybera, the first gene therapy product approved in Europe, using BEVS [5].
Insects are the natural hosts of baculoviruses. Because insects appear periodically, baculoviruses have developed a unique life cycle with two different phenotypes of progeny viruses: occlusion-derived viruses (ODVs) and budded viruses (BVs). ODVs are occluded in occlusion bodies (OBs) which serve as shelters until the OBs are digested by insect larvae, and the ODVs are released from OBs to initiate primary infection. BVs are produced from infected midgut epithelial cells and spread the infection to other insect tissues. Therefore, ODVs are responsible for initiating and spreading infections in insect populations, whereas BVs are responsible for systemic infection within infected larvae [6]. Although these two phenotypes are necessary for the natural life cycle of baculoviral infections, the BEV mainly relies on the productivity of BV.
The genomes of more than a hundred different baculoviruses have been sequenced; they range from 80 to 180 kb and encode 90 to 180 genes [7]. Numerous genes in the baculovirus genome are non-essential for BV production in vitro. For instance, baculovirus encodes 10 per os infectivity factors (PIFs), which are essential for oral infection in larvae. However, the deletion of most PIFs has no impact on BV production in cell culture [8]. In fact, the deletion of certain genes, such as chitinase and cathepsin, enhances the properties of the baculoviral expression system [9,10]. Therefore, the deletion of non-essential genes from the baculovirus genome to make it an optimized vector with a larger capacity is desirable.
Autograph californica multiple nucleopolyhedrovirus (AcMNPV) is the most wellstudied baculovirus and has been widely used to generate baculoviral vectors. It contains a genome of approximately 134 kb and encodes 156 genes [11]. We previously synthesised an AcMNPV genome (AcMNPV-WIV-Syn1) based on the three-step transformation-associated recombination (TAR) in yeast [12]. TAR based on co-transformation of overlapping gene fragments and vector DNAs into Sacharomyces cerevisae, where these fragments and the vector DNAs are joined by homologous recombination. This method was used to synthesis a Mycoplasma genitalium genome [13]. During synthesis, the genome was first amplified in 45 A-level fragments (A1-A45) of~3 kb, then assembled sequentially into 9 B-level fragments (B1-B9) of~15 kb, three C-level fragments (C1-C3) of~45 kb, and finally ã 145 kb full genome. The rescued synthetic virus showed biological properties similar to those of the wild-type parental virus [12]. Recently, we systemically studied the impact of deleting individual AcMNPV genes on BV production and proposed a schematic design for generating a minimal AcMNPV genome [14]. These developments allow for the construction of a minimised AcMNPV genome that deletes all unnecessary genes for BV production and retains suitable reproduction in cell culture.
In this study, as the first step to generate a minimised AcMNPV genome, we focused on a 46.4 kb C1 region of the AcMNPV genome and used synthetic biology to reduce the region. The classical design-build-test synthetic approach was adopted. We first designed modified B1-B3 fragments by deleting all non-essential ORFs and synthesised the first version of modified B fragments by TAR in yeast. The modification effect was then tested by constructing an AcMNPV bacmid and rescuing the virus. After a few rounds of the design-build-test cycle, a synthetic AcMNPV-Syn-mC1-1.1 genome with a~17 kb deletion that had the ability to produce infectious BVs after rescuing the transfected insect cells was constructed.

Construction and Synthesis of Modified Genomes
The C1 region in AcMNPV-WIV-Syn1 is 46.4 kb containing 55 ORFs. Based on the individual gene knockout results [14] and transcriptome analysis data [17], the initial design of deleting all non-essential genes was made by constructing the first version of modified B1-B3. For example, the modified sub-fragments of B1 were first constructed by overlap PCR using the primers listed in Table 1 and the previously constructed B1 [12] as the template. Then, modified B1 (mB1-1.0) was synthesised through TAR in yeast using the two PCR fragments and the pGF vector. The full genome Syn-mB1-1.0 was generated by co-transfection of mB1-1.0, B2, B3, C2, C3, and pGF-egfp vector DNA into S. cerevisiae VL6−48N and was synthesised via TAR. Similar methods were used to generate Syn-mB2-1.0, and Syn-mB3-1.0. In the meantime, three modified B-level fragments (mB1-1.0, mB2-1.0, and mB3-1.0), C2, C3, and pGF-egfp vector were used to synthesise Syn-mC1-1.0. Synthetic genomes were extracted from yeast cells and then electroporated into the EPI300 E. coli strain. After confirmation by PCR and restriction enzyme analysis, the correct genomes were further analysed. After evaluation using transfection and infection assays, the necessary redesigns were made, and new versions of the synthesised genome were generated via the aforementioned methods. Table 1. Primers used in this study.

Name
Sequence ( The designed overlapping sequence is underlined, while the others are the original genome sequences.

Transfection and Infection Assays
Bacmid DNA was extracted from E. coli using methods developed for large plasmids (Instruction Manual of Bac-to-Bac System/Life Technologies, Carlsbad, CA, USA). Sf9 cells (2 × 10 6 ) were transfected with 5 µg of each bacmid DNA using Cellfectin (Invitrogen Carlsbad, CA, USA). The cells were observed under a fluorescence microscope (Invitrogen, Carlsbad, CA, USA) at different time points post-transfection (p.t.). Supernatants of transfected cells were collected at 120 h p.t. and used to infect fresh Sf9 cells. At 120 h post-infection (p.i.), the supernatants were collected, and the titre of BV was determined via the fluorescence of the infected cells using an end-point dilution assay (EPDA). To characterise the BV productivity of the rescued viruses, Sf9 cells (1 × 10 6 ) were infected with the rescued viruses at a multiplicity of infection (MOI) of 1, supernatants were collected at 96 h p.i., and the BV titres were measured using EPDA.

One-
Step Growth Curve 2 × 10 6 Sf9 cells were infected with AcMNPV-Syn-mC1-1.1 or AcMNPV-WIV-Syn1 at an MOI of 1. Supernatants were collected at 0, 24, 48, and 72 h p.i., and the titres of progeny BVs were determined using EPDA. Each experiment was performed in triplicates. Student's t-test was used for significant difference analysis between the titres of the two viruses at each time point using the computer program Graphpad Prism v8 (Insightful Science, San Diego, USA,) [12].

Synthesis and Characterization of the Version 1.0 of Modified Genomes
Previously, AcMNPV-WIV-Syn1 was synthesised based on three C fragments (C1-C3) to form the entire AcMNPV genome [12]. Each C fragment was generated from three B fragments ( Figure 1A). To generate a functional genome with minimalized C1 region, a design-synthesis-test cycling approach was used in our study ( Figure 1B). According to previous studies [14,17], we designed the first version of modified B1-B3 (mB1-1.0, mB2-1.0, and mB3-1.0) by deleting non-essential genes but retaining other genes, including the required transcriptional signals. The fragments were generated by overlap extension PCR and TAR in yeast, and the correct fragments were used individually to synthesise the genomes of AcMNPV-Syn-mB1-1.0, AcMNPV-Syn-mB2-1.0, and AcMNPV-Syn-mB3-1.0. In addition, the three modified fragments: mB1-1.0, mB2-1.0, and mB3-1.0, were combined to synthesise the AcMNPV-Syn-mC1-1.0 genome, where the entire C1 sequence was modified. All the synthesised genomes were transfected to Sf9 cells, and the supernatants were collected for infection to test their capacity for BV production, which was then compared with that of the parental virus AcMNPV-WIV-Syn1. When the synthesised virus showed infectivity similar to that of AcMNPV-WIV-Syn-1, this version of the design was retained. Otherwise, the fragments were redesigned, synthesised, and tested ( Figure 1). The design-build-test cycling method to modify C1 fragment. The design of the modified C1 was based on the modification of B1, B2, and B3 fragments. Non-essential genes of each fragment were removed to generate modified fragments of mB1-1.0, mB2-1.0, and mB3-1.0. mB-1.0, mB2-1.0, and mB3-1.0 by overlapping PCR and used to generate modified genomes of AcMNPV-Syn-mB1-1.0, AcMNPV-Syn-mB2-1.0 and AcMNPV-Syn-mB3-1.0, respectively, using two steps of TAR in yeast. Meanwhile, AcMNPV-Syn-mC1-1.0 was generated using mB1-1.0, mB2-1.0, and mB3-1.0. To test, transfection and infection were performed; and the modified genome, which could not generate efficient progeny viruses, was redesigned and subjected to the design-build-test cycle again.

Restriction Enzyme Analysis and Genome Sequencing of AcMNPV-Syn-mC1-1.1
To verify the accuracy of the final sequence of the AcMNPV-Syn-mC1-1.1 genome, a restriction enzyme digestion assay was performed. The predicted physical maps for BglII, XhoI, and SacII digestion of AcMNPV-Syn-mC1-1.1 and AcMNPV-WIV-Syn1 are shown in Figure 4A. The electrophoresis profiles of the digested AcMNPV-Syn-mC1-1.1 genomic DNA showed the expected pattern ( Figure 4C), as predicted by computer simulation ( Figure 4B). In addition, genome sequencing of AcMNPV-Syn-mC1-1.1 was carried out using high-throughput sequencing, and the results showed that apart from the designed modification, there were no additional changes in the synthetic genome.

One-Step Growth Curve Analysis and Electron Microscopy of AcMNPV-Syn-mC1-1.1 in Comparison with the Parental Virus
To further compare the replication dynamics of AcMNPV-Syn-mC1-1.1 and that of the parental virus AcMNPV-WIV-Syn1, one-step growth curve assays were performed on Sf9 cells using an MOI of 1 ( Figure 5A). The propagation of AcMNPV-Syn-mC1-1.1 appeared to reach the stationary phase after 48 h p.i., as viral titres were 1.08 × 10 5 , 1.88 × 10 5 , and 6.81 × 10 5 TCID 50 /mL at 48 h, 72 h, 96 h p.i., respectively. The parental virus AcMNPV-WIV-Syn1 exhibited optimized productivity, with 1.15 × 10 5 , 2.10 × 10 7 , and 6.84 × 10 7 TCID 50 /mL at 48 h, 72 h, and 96 h p.i., respectively. Statistical analysis showed that the replication dynamics of the two viruses were significantly different (p < 0.001). The morphology of AcMNPV-Syn-mC1-1.1-infected cells was compared with that of AcMNPV-WIV-Syn1 cells by transmission electron microscopy ( Figure 5B). At 24 h p.i., most cells infected with AcMNPV-WIV-Syn1 showed typical viral infections, including nuclear enlargement and assembled nucleocapsids in the ring zone region and at the virogenic stroma in the nucleus, while very few nuclear enlargements and assembled nucleocapsids were observed in the samples of AcMNPV-Syn-mC1-1.1. At 48 h and 72 h p.i., AcMNPV-WIV-Syn1-infected cells showed typical phenomena of late-stage infection, where occlusion bodies with multi-nucleocapsid ODVs accumulated at the ring zone of the nucleus, while AcMNPV-Syn-mC1-1.1 showed enlarged nuclei and nucleocapsids assembled around the ring zone. Notably, the polh gene was deleted in AcMNPV-Syn-mC1-1.1, and, as expected, no occlusion bodies were observed. These results suggest that compared to the parental virus, the viral replication process of AcMNPV-Syn-mC1-1.1, was delayed.

Discussion
Deleting non-essential genes from the baculovirus genome can improve the insertion capacity of BEVS. This study focused on reducing the C1 region of AcMNPV and success-fully rescued the infectious AcMNPV-Syn-mC1-1.1, which deleted approximately 17 kb. This is an important first step towards generating a minimised AcMNPV genome. The modified C1 fragment was built on the modified B1, B2, and B3 fragments using the designbuild-test approach. Our initial design (version 1.0) contained only essential genes, based on previous transcriptomics [17] and individual gene knockout assays [14]. However, the resulting AcMNPV-Syn-mC-1.0, could not produce infectious BV. This suggests that some of the non-essential genes may have synergistic effects in that they cannot be deleted together. Redesigning yielded AcMNPV-Syn-mC1-1.1 that was capable of producing infectious BVs.
Compared to AcMNPV-Syn-mC1-1.0, AcMNPV-Syn-mC1-1.1 contained eight extra genes (ac12, ac13, egt, odv-e26, ac17, ac26, iap-1, lef-6). Among these, five genes or their homologues have been found to affect BV production when deleted. For example, the deletion of Bm5, a homologue of ac13, caused a lower titre of BV [20]. Similarly, viruses with the deletion of Bm17 which is a homologue of ac26, showed a slow spread in cell culture [21]. The knockout of both odv-e26 and ac17 resulted in reduced BV levels [22]. In addition, late and very late transcription was delayed in cells infected with the lef-6-null Bac-mid [23]. Our results also suggest that these genes are essential to efficient BV production in the progeny. Three genes ac12, egt, and iap-1 were retained in AcMNPV-Syn-mC1-1.1 owing to the inconvenience of their deletion during synthesis, although they had no effect on BV production when deleted individually [14,24,25].
AcMNPV-Syn-mC1-1.1 removed a total of approximately 17 kb, including 20 genes as well as two homologous repeat regions hr1 and hr2 in the C1 region ( Table 3). The deleted genes included ac154, bro, ctl, ac4, ac5, orf603, polh, arif-1, ac22, F, ac29, ac30, sod, fgf, lef-12, gta, ac43, ac44, ac45, and odv-e66. All these genes have been found to be non-essential for BV production when deleted individually [14]. The deleted sequences covered approximately 21.5% of the C1 region (~46.4 kb), indicating the flexibility of the AcMNPV genome for minimisation and manipulation.  odv-e66 ODV-E66 B3 Oral infection [46] Non-essential [46] The current AcMNPV-Syn-mC1-1.1, may not be the most minimalised version of C1 in terms of BV production. The deletion of additional genes in C1 while maintaining the acquisition of an infectious virus may be possible. However, as the BV titre of AcMNPV-Syn-mC1-1.1 is already significantly lower than that of the parental virus, further deletion may not be necessary because our goal was to synthesise an entirely modified genome containing modified C2 and C3. The low growth rate of AcMNPV-mC1-1.1 may be due to the modification of mB2-1.1, because the titre of AcMNPV-mB2-1.1 was significantly lower than that of the parental virus (Table 2). We will continue to improve C1 modification and modify C2 and C3 fragments to construct a modified AcMNPV, which has a much smaller genome size but retains efficient infectivity.

Data Availability Statement:
The data is available within the manuscript.