A Novel Strategy of US3 Codon De-Optimization for Construction of an Attenuated Pseudorabies Virus against High Virulent Chinese Pseudorabies Virus Variant

In this study, we applied bacterial artificial chromosome (BAC) technology with PRVΔTK/gE/gI as the base material to replace the first, central, and terminal segments of the US3 gene with codon-deoptimized fragments via two-step Red-mediated recombination in E. coli GS1783 cells. The three constructed BACs were co-transfected with gI and part of gE fragments carrying homologous sequences (gI+gE’), respectively, in swine testicular cells. These three recombinant viruses with US3 codon de-optimization ((PRVΔTK&gE-US3deop−1, PRVΔTK&gE-US3deop−2, and PRVΔTK&gE-US3deop−3) were obtained and purified. These three recombinant viruses exhibited similar growth kinetics to the parental AH02LA strain, stably retained the deletion of TK and gE gene fragments, and stably inherited the recoded US3. Mice were inoculated intraperitoneally with the three recombinant viruses or control virus PRVΔTK&gEAH02 at a 107.0 TCID50 dose. Mice immunized with PRVΔTK&gE-US3deop−1 did not develop clinical signs and had a decreased virus load and attenuated pathological changes in the lungs and brain compared to the control group. Moreover, immunized mice were challenged with 100 LD50 of the AH02LA strain, and PRVΔTK&gE-US3deop−1 provided similar protection to that of the control virus PRVΔTK&gEAH02. Finally, PRVΔTK&gE-US3deop−1 was injected intramuscularly into 1-day-old PRV-negative piglets at a dose of 106.0 TCID50. Immunized piglets showed only slight temperature reactions and mild clinical signs. However, high levels of seroneutralizing antibody were produced at 14 and 21 days post-immunization. In addition, the immunization of PRVΔTK&gE-US3deop−1 at a dose of 105.0 TCID50 provided complete clinical protection and prevented virus shedding in piglets challenged by 106.5 TCID50 of the PRV AH02LA variant at 1 week post immunization. Together, these findings suggest that PRVΔTK&gE-US3deop−1 displays great potential as a vaccine candidate.


Introduction
Pseudorabies is an acute infectious disease caused by a highly virulent and contagious herpesvirus, pseudorabies virus (PRV). Although the natural host is the pig, it can also infect many different mammals, including cattle, goats, cats, dogs, sheep, and wild animals, in which it causes often-fatal disease of the central nervous system [1,2]. Despite immunization with classic vaccines, such as Bartha-K61, there has been ongoing virus evolution with apparent antigenic variation in PRV in China since 2011, which has resulted in several outbreaks of swine pseudorabies in pig farms [3][4][5]. Thus, vaccine development is necessary to combat newly arising variants.
In our previous study, we constructed a PRV variant TK/gE double gene deletion strain (PRV ∆TK&gE-AH02 ), which was safe for 1-day-old PRV antibody-positive piglets and 4-to 5-week-old PRV-negative piglets. Moreover, the strain provided robust protection to a recent PRV variant. However, it showed some pathogenicity in 1-day-old PRV antibodynegative piglets [6]. Consequently, further reducing the virulence of this particular PRV strain while preserving the immunogenicity seemed necessary.
The basic amino acids (except methionine and tryptophan) are often encoded by two or more synonymous codons. Further, evolution has led to different codon usage frequencies in different species or various tissues of the same species [7]. In this context, codon de-optimization has emerged as a tool for replacing synonymous codons in the existing virus sequences with suboptimal codons [8]. The rationale of the technique is that suboptimal codons are used less frequently in the host cell, reducing the mRNA stability, affecting translation efficiency, and thus significantly reducing protein expression levels [8][9][10]. Moreover, it should be noted that codon de-optimization does not result in changes to the amino acid sequence of the viral proteins, which retains the same antigenic epitopes as the wild-type virus and thus remains completely immunogenic [9,11]. Owing to these aspects, codon de-optimization-based modification of viruses to attenuate virulence has become a popular trend to produce vaccine candidates. Several attenuated viruses have been generated using various codon de-optimization strategies [12][13][14][15][16]. We here focused on the US3 gene, which encodes a serine/threonine kinase. Previous studies indicated that the US3 gene was dispensable for virus growth in cells but was a critical virulence factor in vivo [17]. Motivated by these aspects, we constructed three recombinant viruses with US3 codon de-optimization based on a PRV TK/gE double gene deletion strain (PRV ∆TK&gE-AH02 ) by replacing partial synonymous codons with suboptimal codons in the first, central, and terminal segments of the US3 gene with reference to the porcine codon usage frequency table and codon adaptation index (CAI). We then systematically evaluate their safety and immune potency in mice and piglets.

Test Animals
The 4-to 6-week-old healthy Institute of Cancer Research (ICR) female mice (18-22 g) were purchased from Nanjing Qinglong Mountain Animal Breeding Farm (Nanjing, China) and reared in the mice room of the North Animal House of Jiangsu Academy of Agricultural Sciences (Nanjing, China). The 1-day-old PRV-negative piglets and sows were self-raised in the laboratory and reared in the same pens of a pig farm of Jiangsu Academy of Agricultural Sciences. The piglets, aged 28-35 days old, were self-raised and fed twice daily (at 8:00 and 17:00) in the pig farm of Jiangsu Academy of Agricultural Sciences. It should be noted that all the test pigs were negative for PRV, swine fever virus, cerebrospinal virus, porcine reproductive and respiratory syndrome virus, porcine parvovirus, and porcine circovirus type 2.

Construction of Bacterial Artificial Chromosomes (BACs) with US3 Codon De-Optimization
The primer sequences used in this study are shown in Table 1. The primers were synthesized by the Tsingke Biotechnology Co. Ltd. (Nanjing, China). The codon de-optimized US3 (US3 deop -1, US3 deop -2, US3 deop -3 in the Supplementary Information S1) was obtained after the codon de-optimization of the US3 gene into the first, central, and terminal segments with reference to the porcine codon usage frequency table and CAI. The recombinant plasmids, T-US3 deop -1, T-US3 deop -2, and T-US3 deop -3, were constructed by ligating the synthesized US3 deop -1, US3 deop -2, and US3 deop -3 with the linearized PMD19-T simple vectors. Then, plasmids T-US3 deop -1, T-US3 deop -2, and T-US3 deop -3 were digested with restriction endonucleases Cla I, Btg I, and BspE I, respectively. The target fragments were ligated with kanamycin resistance gene, respectively, and then transformed into DH5α cells to obtain recombinant plasmids. Finally, PCR and sequencing were performed to identify whether the kanamycin resistance gene was correctly ligated in recombinant plasmids.
The obtained BAC PRV∆TK/gE/gI-US3deop−1&K+ , BAC PRV∆TK/gE/gI-US3deop−2&K+ , and BAC PRV∆TK/gE/gI-US3deop−3&K+ were subjected to a second step of Red recombination to remove the kanamycin resistance genes. The BAC DNA was initially extracted using the double plate resistance screening approach and further characterized with PCR, and subsequently sequenced using primers PRV US3 check F/R, respectively. The three PCR-positive BAC DNA were selected and digested by Kpn I for RFLP analysis. Finally, the cloned strains that matched with the predicted profiles were named BAC PRV∆TK/gE/gI-US3deop−1 , BAC PRV∆TK/gE/gI-US3deop−2 , and BAC PRV∆TK/gE/gI-US3deop−3 , respectively.

Acquisition of Recombinant Viruses with US3 Codon De-Optimization
The DNA of PRV ∆TK&gE-AH02 was used as a template to amplify the gI and partial gE gene fragments with homologous arms (gI+gE'), using the primers PRV gI+gE' F/R. Then, the gI+gE' fragments were co-transfected with BAC PRV∆TK/gE/gI-US3deop−1 , BAC PRV∆TK/gE/gI-US3deop−2 , and BAC PRV∆TK/gE/gI-US3deop−3 DNA, respectively, on ST cells using Lipofectamine ® 3000 following the manual of supplier. The mini-F sequences of BAC PRV∆TK/gE/gI-US3deop−1 , BAC PRV∆TK/gE/gI-US3deop−2 , and BAC PRV∆TK/gE/gI-US3deop−3 were replaced with gI+gE' fragments. After 24 h of transfection, the culture medium was discarded and supplemented with DMEM containing 10% FBS and 0.5% methylcellulose for 24-48 h. Then, the viruses with no green fluorescence emission were picked under the excitation of Ultraviolet (UV) light at a wavelength of 488 nm and inoculated with fresh ST cells. After picking for several rounds, the DNAs from the three purified viruses were extracted, and the gE gene was identified through PCR and sequencing with a pair of primers (PRV gE site check F/R). Finally, the adequately identified strains were denoted PRV ∆TK&gE-US3deop−1 , PRV ∆TK&gE-US3deop−2 , and PRV ∆TK&gE-US3deop−3 , respectively. A schematic diagram of the construction of PRV ∆TK&gE-US3deop−1 as an example is shown in Figure 1.

Detection of US3 Gene Expression from the Virus Background
To detect the US3 gene expression level from the virus background, PRV ∆TK&gE-US3deop−1 , PRV ∆TK&gE-US3deop−2 , PRV ∆TK&gE-US3deop−3 , and PRV ∆TK&gE-AH02 , as well as parental AH02LA strains, were inoculated with a multiplicity of infection (MOI) of 10 into a monolayer of ST cells. ST cells were harvested at different times (2, 6, and 10 h) after infection. Then, total RNA was extracted and subjected to reverse transcription. After reverse transcription, the expression of the US3 gene in different virulent strains was detected using fluorescence qPCR. The primers are shown in Table 1.

Cytopathic Effect
To determine the characteristics of cytopathic effects (CPE), PRV ∆TK&gE-US3deop−1 , PRV ∆TK&gE-US3deop−2 , PRV ∆TK&gE-US3deop−3 , PRV ∆TK&gE-AH02 , and parental AH02LA strains were inoculated into the 6-well plates seeded with freshly grown monolayers of ST cells at a MOI of 0.01. Cells were incubated with a cell maintenance solution to 0.6 mL at 37 • C. After 1 h of incubation, the inoculum was removed and DMEM containing 2% FBS was added. Further, CPE were observed under an inverted microscope at 24, 48, and 72 h post-infection, respectively.

Plaque Assay
PRV ∆TK&gE-US3deop−1 , PRV ∆TK&gE-US3deop−2 , PRV ∆TK&gE-US3deop−3 , and PRV ∆TK&gE-AH02 , along with parental AH02LA, were inoculated into the 6-well plates seeded with fresh ST cells at a MOI of 0.01. DMEM was added as a cell maintenance solution to 0.6 mL and incubated at 37 • C for 1h. Further, the inoculum was removed and a cell maintenance medium containing 1% low melting point agarose and 2% FBS was added. After solidifying the medium, the plate was further incubated for 24 h. A total of 100 images of each virus were taken under the same doubling microscope for each plaque, whose surface area was determined by Image J software (LOCI, University of Wisconsin, Madison, WI, USA). The plaque areas of other strains were compared with that of the parental AH02LA strain, which was set to 100%. All experiments were repeated thrice, independently.
The healthy ICR female mice (n = 20) aged 4-6 weeks were randomly divided into 4 groups of 5 mice each. Further, the mice were intraperitoneally injected with 0.2 mL of PRV ∆TK/gE-US3deop−1 , PRV ∆TK/gE-US3deop−2 , PRV ∆TK/gE-US3deop−3 , or PRV ∆TK&gE-AH02 (10 7.0 TCID 50 /0.2 mL), respectively. At 5 days post-infection, all mice were sacrificed and major organs, i.e., brain and lung samples, were collected. The tissue samples were prepared by weighing 0.1 mg of samples separately and placing them in the grinding beads. After grinding, the samples were centrifuged. Then, the supernatant was used to extract DNA, and the PRV gB gene was used as the target. The viral load was quantified with real-time quantitative PCR using primers gB F/R. Notably, the gB standard curve was used for the linear regression analysis of copy numbers. In addition, histopathological sections of the brain and lung tissues were prepared, and the changes were observed.
Based on our previous study [18], after 2 weeks of immunization in the 10 7.0 TCID 50 dose group, mice were administered with 100 LD 50 of PRV AH02LA strain. Further, the clinical signs and mortality rate of mice were monitored daily for 14 days after the challenge. At the end of the experiment, all surviving mice were sacrificed and disposed of harmlessly.

Pathogenicity and Immunological Experiments in Piglets
Briefly, 1-day-old healthy piglets (n = 15), negative for PRV, were randomly divided into 3 groups of 5 piglets each. The piglets in the immunized group were intramuscularly inoculated with 1 mL of PRV ∆TK&gE-AH02 (10 6.0 TCID 50 /mL), PRV ∆TK/gE-US3deop−1 (10 6.0 TCID 50 /mL), and DMEM, respectively. High-dose PRV vaccines were used to evaluate their safety in 1-day-old piglets in clinical applications [6,18]. Further, body temper-atures, clinical signs, and the mortality rate of piglets in different treatment groups were recorded daily for 14 days. Finally, serum samples were collected from piglets at 7, 14, and 21 days post-inoculation and tested for neutralizing antibody index. Then, 100 µL of each serum sample (heat inactivated for 30 min at 56 • C) was mixed with an equal volume of virus (AH02LA) at a different dilution. The neutralization indexes were calculated as the TCID 50 of piglet serum in the test group divided by the TCID 50 of negative serum.
Piglets aged 28-35 days old (n = 15) were randomly divided into 3 groups of 5 piglets each and reared in isolated environments in group cages before treatment. The piglets were intramuscularly inoculated with 1 mL of PRV ∆TK&gE-AH02 (10 5.0 TCID 50 /mL), PRV ∆TK/gE-US3deop−1 (10 5.0 TCID 50 /mL), and DMEM, respectively. At 1 week post-immunization, all piglets were administered with 2 mL of PRV AH02LA strain (10 6.5 TCID 50 /2 mL) via nasal drip based on our previous study [6]. After challenging, the body temperatures, clinical signs, and mortality rate of piglets were recorded daily for 14 days. Nasal swab samples were collected daily from 0 to 14 days post challenge to detect virus shedding. After shaking and freeze-thaw cycles (−80 • C and 37 • C), samples were centrifuged (10,000 rpm) and the supernatants were used to determine the viral titers.

Statistical Analysis
The data were represented as mean ± standard deviation (SD) and analyzed with one-way analysis of variance (ANOVA) followed by Tukey's test using GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA, USA), considering a p-value less than 0.05 as statistically significant. * represents p < 0.05, ** indicates p < 0.01, *** signifies p < 0.001, and p > 0.05 refers to a not-significant result.

Generation of Recombinant BAC with US3 Codon De-Optimization
Initially, the codon de-optimized US3 containing the kanamycin resistance gene was subjected to the two-step Red-mediated homologous recombination in E. coli GS1783 containing BAC PRV∆TK/gE/gI , and the kanamycin resistance gene was knocked out. Further, the three BACs were identified using the double-plate resistance screening, and PCR and sequencing. Then, BAC DNA was cleaved with Kpn I and subjected to RFLP analysis. It was observed that the predicted PRV (GenBank: KM061380.1) profiles were essentially identical ( Figure 2A). Compared to BAC PRV∆TK/gE/gI-US3deop−1&K+ , BAC PRV∆TK/gE/gI-US3deop−2&K+ , BAC PRV∆TK/gE/gI-US3deop−3&K+ , BAC PRV∆TK/gE/gI-US3deop−1 , BAC PRV∆TK/gE/gI-US3deop−2 , and BAC PRV∆TK/gE/gI-US3deop−3 possessed an additional band of 5843 bp size and a missing band of 6881 bp size.

Safety and Immunogenicity in Piglets
Considering the safety and immunogenicity results in mice, we further explored the safety and immunogenicity of PRV ∆TK&gE-US3deop−1 in piglets. One-day-old piglets were inoculated intramuscularly with 10 6.0 TCID 50 of PRV ∆TK&gE-US3deop−1 or PRV ∆TK&gE-AH02 and the temperature changes as well as clinical signs were monitored for 14 days ( Figure 5A,B and Supplementary Table S2). After inoculation with PRV ∆TK&gE-US3deop−1 , it was observed that two piglets presented a normal body temperature and showed no other substantial clinical symptoms. However, three piglets exhibited a body temperature of 40.0-40.5 • C for 2-3 days, and two piglets showed mild respiratory symptoms. In contrast, after PRV ∆TK&gE-AH02 vaccination, four piglets displayed a body temperature of 40-41 • C for 3-6 days, and two piglets developed clinical symptoms, such as sneezing, coughing, and loss of appetite, while the other piglets appeared healthy. Moreover, serum was collected from piglets at 7, 14, and 21 days post-immunization to detect the neutralizing antibody index ( Figure 5C). At 7 days post-immunization, the neutralizing antibody index was low in the PRV ∆TK&gE-US3deop−1 and PRV ∆TK&gE-AH02 treatment groups, and it subsequently increased progressively. At 14 days post-immunization, the neutralizing antibody index was 50,118 in the PRV ∆TK&gE-US3deop−1 treatment group and 72,443 in the PRV ∆TK&gE-AH02 treatment group. At 21 days post-immunization, the neutralizing antibody index was 1,698,244 in the PRV ∆TK&gE-US3deop−1 group and 2,884,031 in the PRV ∆TK&gE-AH02 group. Notably, no significant difference in the neutralizing antibody index was observed between the PRV ∆TK&gE-AH02 and PRV ∆TK&gE-US3deop−1 groups at 7, 14, and 21 days post-immunization. In future studies, we will also detect the duration of antibodies in piglets immunized with PRV ∆TK&gE-US3deop−1 . For protective efficacy, all piglets in the challenge control group developed a reduced appetite and sneezing at 2 days post-challenge. At 3 days post-challenge, typical symptoms, such as nasal mucous discharge, salivation, abdominal breathing, depression, and loss of appetite, began to appear. The body temperature of all piglets in the challenge control group reached over 41 • C for 3-6 days. All five piglets in the challenge control group died during the test period. After the AH02LA challenge, no clinical signs or temperature responses were observed in piglets immunized with PRV ∆TK&gE-US3deop−1 and PRV ∆TK&gE-AH02 ( Figure 6A,B and Supplementary Table S3). After the challenge, nasal swabs were collected daily to determine the virus titers in the excreted nasal discharge. It was observed that all piglets in the control group started to shed the virus at 1 day post-challenge, and continued to do so until death. However, no virus was detected in piglets immunized with PRV ∆TK&gE-US3deop−1 and PRV ∆TK&gE-AH02 ( Figure 6C and Supplementary Table S4).

Discussion
Since 2011, a virulent variant of PRV has developed into an epidemic in pig operations in China. It has increasingly been recognized that the traditional Bartha-K61 vaccine showed poor immune protection, lacking the ability to prevent virus shedding [19][20][21]. To deal with the hazardous nature of PRV mutant strains, several PRV gene-deleted and attenuated vaccines have been developed using mutant strains that can improve protection against PRV variant strains, and discriminate between immunized pigs and pigs infected with wild-type virus [6,[22][23][24]. Nevertheless, the safety of these attenuated strains for PRV antibody-negative neonatal piglets is particularly perturbing [6]. The further attenuation of these strains while maintaining immunogenicity is necessary to develop a safe and effective live PRV vaccine.
Codon de-optimization has emerged as one of the efficient strategies for the engineering of live-attenuated vaccines in many systems [9,12]. Notably, codon de-optimization is often achieved using suboptimal codons to recode virulence genes. This approach offers many advantages, as stated in the following. Regarding safety, codon de-optimization often relies on introducing many synonymous mutations in virulence genes, leading to the highly efficient attenuation of the virus by reducing viral gene expression [8,25]. Considering excellent immunogenicity, codon de-optimization did not change the amino acid sequence in the viral protein, retaining the same antigenic epitopes as the wild-type virus [8]. In terms of simple operation, the codon de-optimized viruses could be generated rapidly via gene synthesis and reverse genetics, and it is theoretically possible to control the degree of attenuation.
Although a lack of viral protein kinase activity results in reduced virulence, several kinases, including herpesvirus kinase, are often dispensable for growth in cell culture [17,18,26]. The US3 gene encodes a 334-390 amino acid protein serine/threonine kinase, and is a positive regulator of viral replication and pathogenicity [27,28]. Notably, the US3 gene is involved in viral particle formation, cytoskeletal rearrangement, and the escape of multiple host antiviral responses [29][30][31].
In the current study, the US3 gene was divided into three segments, and codon deoptimization was designed for each. It was observed that the protein expression levels of the recombinant eukaryotic plasmid were significantly reduced after the US3 codon de-optimization. Further, the virus growth curve assay showed that the US3 gene deoptimization showed no substantial effect on the virus proliferation. Nevertheless, several differences were evident when compared to the parental virus, such as the delayed time of plaque formation, a rare occurrence of cell fusion and smaller plaques. The genetic stability test revealed that there was no change in the codon de-optimized sequences, heralding a very low possibility of reversion to virulence. The three recombinant viruses with US3 codon de-optimization were tested for pathogenicity and immunogenicity in mice. The results showed that PRV ∆TK&gE-US3deop−1 decreased the virus load and attenuated pathological changes in the brain and lung of mice compared with PRV ∆TK&gE-AH02 . The protection efficiency of PRV ∆TK&gE-US3deop−1 was similar to PRV ∆TK&gE-AH02 in mice. Further, a piglet safety test showed that PRV ∆TK&gE-US3deop−1 was less pathogenic in piglets than PRV ∆TK&gE-AH02 , producing high levels of neutralizing antibody. Finally, an evaluation of the immunoprotective effect of PRV ∆TK&gE-US3deop−1 in piglets showed that it not only provided complete protection, but also prevented virus shedding against the virulent variant AH02LA challenge at 1 week post-immunization. It is clear that levels of antibodies correlate poorly with the decreased virus replication early after infection, and PRV-specific lymphocyte proliferation responses and a rapid influx of T lymphocytes at the site of viral replication play an important role in the clearance of PRV infection [32,33]. Therefore, the PRV-specific cell-mediated immune response may have been related to the prevention of clinical disease and virus shedding early after challenge. Future studies involving cellmediated immunity analysis are necessary to better understand the mechanisms of immune protection induced by live PRV vaccines.

Conclusions
In summary, a recombinant virus with US3 codon de-optimization (PRV ∆TK&gE-US3deop−1 ) was successfully constructed. With its high immune efficacy, it could be a suitable vaccine candidate, presenting its potential for eliminating newly arising PRV variants in pig farms. This study also provided a new theoretical basis and technical means for PRV vaccine development.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/vaccines11081288/s1, Supplementary Information S1: Nucleotide sequence of US3-1, US3-2, US3-3 and US3 deop -1, US3 deop -2, US3 deop -3. Nucleotide changes compared to original sequence are highlighted in red. Table S1: Data presents the immunogenicity of different virus strains in mice. Table S2: Data presents the pathogenicity of different virus strains in piglets. Table S3: Data presents the immunogenicity of different virus strains in piglets. Table S4: Determination of virus titers in nasal swabs collected daily in the challenge-infected piglets.

Institutional Review Board Statement:
This study was approved by the Animal Ethics Committee of Jiangsu Academy of Agricultural Sciences and followed the guidelines of animal experimentation outlined by Jiangsu Province (SYXK-Su-2020-0023).