Introduction

H9N2 avian influenza viruses (AIVs) have become increasingly prevalent in domestic ducks [6, 20]. Because of the wide host range of H9N2 AIVs, reassortment between H9N2 and other influenza viruses is possible. For instance, H5N1 viruses, which infected humans in Hong Kong in 1997, are presumed to have gained six genes from H9N2 viruses (Qa/HK/G1/97) [11]. This potential for reassortment remains a public health concern.

Wild aquatic birds play an important role in the transmission of AIVs as a natural reservoir and carrier of these viruses [1, 25]. Domestic ducks serve as intermediaries between wild aquatic birds and terrestrial poultry. Thus, effective control of H9N2 viruses in domestic ducks is important to prevent AIV infection in poultry and protect public health.

Vaccination is one of the most effective ways to prevent and control influenza outbreaks in poultry. However, H9N2 AIV vaccines for ducks are scarce. At present, a commercial vaccine of inactivated virus mixed with mineral oil adjuvant is available to control H9N2 AIVs in ducks in China; however, it remains unstable. The hemagglutination inhibition (HI) antibody geometric mean titer induced in ducks by the inactivated vaccine is about 1:64, however, the titers can widely vary and some are as low as 1:8 (unpublished data). Moreover, compliance with vaccination schedules is sometimes poor because ducks are asymptomatic after infection. Therefore, the immunization rate of H9N2 vaccine in ducks is very low, and a large proportion of ducks remain susceptible to H9N2 AIV viruses. Domestic ducks in Southern China are raised in a high-density and free-range manner. As the world’s largest domestic duck population is in China, an efficient vaccine for ducks is urgently needed.

Duck plague, caused by duck enteritis virus (DEV), is an acute, contagious, and lethal disease of ducks, geese, and swans, which results in high mortality and decreased egg production in domestic and wild waterfowl [27]. DEV is one of the main diseases restricting the development of the duck industry in China. Vaccination with an attenuated vaccine is an efficient way to control DEV infection, and about 1 billion doses of DEV live vaccine are used in China annually. DEV taxonomically belongs to species Anatid herpesvirus 1, genus Mardivirus, subfamily Alphaherpesvirinae, family Herpesviridae (http://www.ictvonline.org), and possesses a large genome, approximately 160 kb, composed of a unique long (UL) region and a unique short (US) region, with the US region flanked by inverted repeated sequences (IRS and TRS) [15, 31–34]. Similar to other herpesviruses [3, 5, 13, 18, 21, 22, 26], DEV may be an efficient vector candidate for a live vaccine. Recently, DEVs have been used to construct recombinant viruses to express foreign antigens from other pathogens [14, 16, 17, 29, 35, 36].

In this study, we constructed a recombinant DEV by inserting the hemagglutinin (HA) gene from duck original H9N2 AIV in to the UL2 gene by homologous recombination. Our results showed that HA insertion had no effect on viral replication and suggested that the UL2 gene was nonessential for virus growth in vitro. In vivo tests further demonstrated that the recombinant virus induced antibody responses against DEV and H9N2 AIV virus, and provided solid protection against challenge with virulent strains of both viruses.

Materials and methods

Viruses and cells

The attenuated DEV strain C20E85 was kept in our laboratory; the virulent DEV strain AV1221 was obtained from the China Veterinary Culture Collection (Beijing, China). H9N2 (A/duck/Guangdong/08) AIV was isolated from ducks in China and propagated in the allantoic cavities of 10-day-old, specific-pathogen-free (SPF), embryonated chicken eggs. Primary chicken embryo fibroblasts (pCEF) were prepared in our laboratory according to standard procedures, maintained in M199 medium supplemented with 5 % fetal bovine serum (M199-FBS) and used to propagate DEV strain C20E85.

Preparation of DNA

Viral DNA was extracted from infected pCEF with sodium dodecyl sulfate -proteinase K as described previously [19].

Construction of recombinant plasmids

The plasmid pT-GFP-gpt was kept in our laboratory, and contained a Sal I restriction site, an EGFP expression cassette flanked by the immediate-early promoter of human cytomegalovirus (PHCMV-IE), and a simian virus 40 polyadenylation signal (A +SV40 ).

Two fragments of 1.2 kb (114037-115221, 115750-117051) (GenBank accession number KF487736) were amplified with UL2 gene primers (Table 1). Both fragments were subsequently cloned into the pMD-18T vector (TaKaRa, Dalian, China) to generate pT-UL2-ud, with Sal I restriction sites at both ends. The EGFP expression cassette was inserted into this vector, generating the plasmid pT-UL2-GFP. Finally, the HA gene was amplified from the H9N2 virus (A/duck/Guangdong/08) with H9 primers (Table 1), and used to replace the EGFP gene in the plasmid pT-UL2-GFP to generate the recombinant plasmid pT-UL2-HA (Fig. 1).

Table 1 Primers for PCR
Full size table
Fig. 1
figure 1

Construction of the recombinant plasmid. (A) Map of the Duck Enteritis Virus (DEV) genome, which consists of long (UL) and short (US) unique regions with inverted repeat sequences (IRS,TRS) flanking the US region. (B) Regions upstream (UL2-u) and downstream (UL2-d) of UL2 were amplified by PCR. (C) Construction of UL2-GFP. UL2 was replaced by an expression cassette encoding EGFP flanked by the immediate-early promoter of human cytomegalovirus (PHCMV-IE) and the simian virus 40 polyadenylation signal (A +SV40 ). (D) Construction of UL2-HA. The EGFP expression cassette was replaced by the HA gene

Full size image

Generation of the recombinant virus

Confluent pCEF monolayers were transfected with DEV genomic DNA and recombinant plasmid pT-UL2-GFP using Lipofectamine 2000 (Invitrogen, CA, USA), according to the manufacturer’s instructions. When cytopathic effect was observed the total supernatant and cells were harvested. The infected virus was diluted and then plated on fresh pCEF monolayers and overlaid with M199-FBS containing 1 % agarose. When green fluorescent plaques were observed plaque purification was performed to obtain a green fluorescent plaque population termed rDEV-∆UL2-GFP. rDEV-∆UL2-GFP and the plasmid pT-UL2-HA were then used to transfect pCEFs in the manner described above. Plaques without fluorescence were then selected, and after several rounds of selection, the purified recombinant plasmid, designated rDEV-∆UL2-HA, was obtained.

Detection of insertion and expression of the HA gene in cells infected with rDEV-∆UL2-HA

Polymerase chain reaction analysis

Polymerase chain reaction (PCR) was used to verify the insertion of the HA gene in rDEV-∆UL2-HA-infected cells. Primary CEF grown on coverslips in 25-cm2 flasks were infected with rDEV-∆UL2-HA. At 48 to 72 hours (h) post-infection, the cells were frozen and thawed twice to obtain the virus. Total DNA was extracted, as previously described [19], and a PCR with ORF C17 primers (Table 1, Fig. 1B) was performed. PCR products were commercially sequenced by Invitrogen (Shanghai, China).

Indirect immunofluorescence assay

HA gene expression in recombinant DEVs was confirmed by indirect immunofluorescence assay (IFA). Primary CEF grown on coverslips in 48-well plates were infected with rDEV-∆UL2-HA at a multiplicity of infection of 0.2. When cytopathogenicity was evident (about 48 h post-infection), cells in 48-well plates were fixed with cold (4 °C) methanol for 15 min at room temperature, then washed three times with phosphate-buffered saline (PBS). A chicken anti-DEV or anti-H9N2 polyclonal antibody (1:200 dilution, prepared in our laboratory) was added, followed by incubation for 1 h at 37 °C. After washing three times with PBS including 0.1 % Tween-20, cells were incubated with a fluorescein isothiocyanate (FITC)-conjugated goat anti-chicken IgG (1:200, Sigma, Missouri, USA) for 1 h at 37 °C.

Stability of rDEV-∆UL2-HA

To evaluate the genetic stability of foreign gene insertions in recombinant viruses they were passaged 20 times in pCEF. Viral DNA was extracted from infected cells for PCR analysis every five generations. The inserted HA gene was detected by PCR with ORF C17 primers. HA expression was also confirmed by IFA as described above.

Growth properties of rDEV-∆UL2-HA

To compare the growth of rDEV-∆UL2-HA with its parental virus, one-step growth analyses were performed in pCEF, as described previously [34].

Animal experiments

SPF ducks were obtained from the Harbin Veterinary Research Institute, CAAS, China, and raised in an SPF facility at our laboratory. A total of 100 4-week-old SPF ducks were used for these studies. The experimental protocols were approved by the Animal Care and Protection Committee, China Institute of Veterinary Drug Control (No. [2015]00151). All animal experiments were carried out in accordance with the requirements of the Regulations of Experimental Animal Administration of the P. R. China. All efforts were made to minimize suffering.

Five groups of ducks (n = 10 per group) were inoculated intramuscularly with 102–106 50 % tissue culture infectious dose (TCID50) of rDEV-∆UL2-HA virus, respectively, and two groups of ducks (n = 10 per group) were intramuscularly inoculated with 103 TCID50 of parental DEV or 1 ml of PBS as controls. At 2, 3, and 4 weeks post-vaccination (p.v.) sera were obtained from all ducks to monitor neutralizing (NT) antibody against DEV, and HI antibody against H9N2 AIV virus, according to the methods described in OIE standard protocols [8]. At 4 weeks p.v. all ducks were challenged with 108 50 % egg infectious dose (EID50) H9N2 virus (A/duck/Guangdong/08) by intravenous injection. Oropharyngeal swabs were collected from H9N2 virus challenged ducks each day, from 1 to 4 days post-challenge (p.c.), to detect viral shedding.

To evaluate the clinical protection of rDEV-∆UL2-HA against DEV challenge, three groups of ducks (n = 10 per group) were inoculated intramuscularly with 103 TCID50 (the recommended dose of DEV vaccine in China is 103.5 TCID50) of rDEV-∆UL2-HA, parental DEV, or 1 ml PBS. Ducks were then challenged with 1000× minimum lethal dose of DEV strain AV1221 by intramuscular injection at 14 days p.v.. Ducks were observed for symptoms of disease or death for 14 days after challenge.

Results

Construction and purification of recombinant virus

As described in the “Materials and methods”, we first constructed two recombinant plasmids by inserting a GFP expression cassette or the HA gene into the homologous arms of the DEV UL2 gene. Recombinant viruses were generated by conventional homologous recombination between the parental virus and transfer plasmids in transfected pCEF. In rDEV-∆UL2-GFP, an EGFP expression cassette was inserted in the UL2 gene. This reporter gene facilitated isolation of the recombinant virus. Viral plaques with green fluorescence were isolated to purify the recombinant virus. In rDEV-∆UL2-HA, the EGFP gene was replaced by the IAV H9 HA gene. Non-fluorescent viral plaques were then isolated after six rounds of selection, and purified rDEV-∆UL2-HA was obtained.

Expression of the HA gene in cells infected with rDEV-∆UL2-HA

Total DNA was extracted from pCEF infected with rDEV-∆UL2-HA and used for PCR with ORF C17 primers to yield a 2800-bp fragment (data not shown). Subsequent sequencing confirmed that the HA expression cassette was successfully inserted into the recombinant DEV.

IFA was performed to confirm HA expression. As expected, cells infected with rDEV-∆UL2-HA reacted with chicken antiserum to HA protein, as well as antiserum against DEV. In contrast, cells infected with parental virus only reacted with antiserum against DEV, but did not react with chicken antiserum to H9 (Fig. 2).

Fig. 2
figure 2

Detection of HA protein in pCEF infected with rDEV-∆UL2-HA by immunofluorescence. Confluent pCEF were infected with rDEV-∆UL2-HA (b, e) or parental DEV (c, f). Cells in a, b, and c were incubated with chicken anti-H9N2 AIV polyclonal antibody, and cells in d, e, and f were incubated with chicken anti-DEV polyclonal antibody, then stained with a goat anti-chicken IgG-FITC conjugate, and examined using fluorescence microscopy. a and d indicate negative controls

Full size image

Characterization of rDEV-∆UL2-HA

After 20 passages in pCEF, the HA gene in all of the recombinants was detectable by PCR amplification (Fig. 3A). Additionally, HA expression in recombinant virus-infected cells was retained after 20 passages (Fig. 3B)

Fig. 3
figure 3

Stability of rDEV-∆UL2-HA. (A) Detection of HA gene insertion in rDEV-∆UL2-HA by PCR. M, DNA Marker IV; P, parental virus (negative control). The lane numbers represent the passage numbers of the recombinant virus. (B) Detection of HA expression in rDEV-∆UL2-HA-infected CEFs by immunofluorescence. a, negative control; b, the 10th passage (100×); c, the 20th passage (100×)

Full size image

We also investigated the growth kinetics of the recombinant virus in pCEF. One-step growth analyses showed that rDEV-∆UL2-HA virus and its parental virus exhibited comparable growth kinetics (Fig. 4), suggesting that the insertion of a foreign gene in UL2 gene did not affect DEV replication.

Fig. 4
figure 4

One-step growth curves of the recombinant virus and its parental virus in pCEF. Infected cells and supernatants were collected and viral titers were determined at the indicated time points post-inoculation

Full size image

Neutralizing antibody against DEV

All ducks inoculated with rDEV-∆UL2-HA remained healthy during the observation period, demonstrating the safety of the recombinant virus in ducks. Sera were obtained from all vaccinated ducks to screen for NT antibody. The NT antibody titers of the PBS-inoculated group were lower than 1:4 and considered negative. By contrast, the NT antibody titers of ducks in rDEV-∆UL2-HA vaccinated groups exceeded 1:15 at 2 weeks p.v., which was similar to the parental DEV-vaccinated group. Thereafter, the titers of both the parental and recombinant virus vaccinated groups started to decline. However, these titers remained higher than those of the negative control group (Table 2). Furthermore, there was no significant difference (p < 0.001) between the NT antibody titers primed by 103 TCID50 rDEV-∆UL2-HA and the parental virus, indicating that the insertion of the HA gene did not change the immunogenicity of the parental virus.

Table 2 Geometric mean titers of neutralising antibody against DEV in ducks
Full size table

Protective efficacy against lethal DEV challenge in ducks

After challenge, all ducks in the PBS-inoculated control group showed signs of disease, including listlessness, ruffled feathers, and anorexia, and died within 6 days. In contrast, all ducks inoculated with 103 TCID50 of rDEV-∆UL2-HA or parental virus remained healthy during the 2-week observation period (Fig. 5). These findings suggest that the insertion of the HA gene did not alter the protective efficacy of the parental virus.

Fig. 5
figure 5

Protective efficacy of the recombinant virus against lethal DEV challenge. Groups of ducks (n = 10 per group) were vaccinated intramuscularly with 103 TCID50 of recombinant virus or parental virus and challenged with lethal DEV at 14 days post-vaccination. Ducks inoculated with PBS were used as a negative control. Ducks were monitored daily for 14 days after challenge

Full size image

Antibody responses against H9N2 virus induced by rDEV-∆UL2-HA in ducks

To understand the antibody response induced by varying the dose of the recombinant virus, five groups of ducks (n = 10 per group) were inoculated with five doses ranging from 102–106 TCID50 of rDEV-∆UL2-HA. Sera from all ducks were collected at 2, 3, and 4 weeks p.v. for HI antibody detection. The HI titers gradually increased from 1:16 at 2 weeks p.v. to 1:256 at 4 weeks p.v., irrespective of dose concentration (Fig. 6).

Fig. 6
figure 6

HI antibody titers against H9N2 virus in ducks inoculated with recombinant DEV. Groups of ducks (n = 10 per group) were inoculated intramuscularly with five doses of recombinant virus or inoculated with PBS or parental virus as negative controls. Sera were collected at 2, 3, and 4 weeks post-vaccination for HI antibody titer assays

Full size image

Protective efficacy against H9N2 AIV challenge in ducks

To evaluate the protective efficacy of the recombinant virus against H9N2 AIV challenge, all ducks inoculated with varying doses of the recombinant virus were intravenously injected with 108 EID50 H9N2 AIV at 28 days p.v.. Oropharyngeal swabs were collected each day from 1 to 4 days p.c. to isolate the H9N2 AIV virus. All vaccinated and control ducks were asymptomatic and survived during the observation period. Ducks inoculated with 103–106 TCID50 of recombinant virus were fully protected from challenge, as evidenced by the finding that no virus was recovered from oropharyngeal swabs collected from vaccinated birds. Conversely, ducks inoculated with 102 TCID50 of rDEV-∆UL2-HA were only partially protected against H9N2 virus challenge. All ducks survived challenge, yet the virus was recovered from oropharyngeal swabs of 6/10 ducks. The virus was detected in all oropharyngeal swabs collected from control ducks inoculated with DEV or PBS, although they all survived after challenge (Table 3). These results indicate that a single dose of 103 TCID50 of rDEV-∆UL2-HA can induce robust protection against H9N2 AIV challenge.

Table 3 Protective efficacy of rDEV-∆UL2-HA against H9N2 virus challenge in ducksa
Full size table

Discussion

To date, recombinant DEVs expressing HA of H5N1 AIVs, duck Tembusu virus, and infectious bronchitis virus have been reported [4, 14, 16, 17, 29, 30, 35, 36]. In this study, we generated a recombinant DEV, rDEV-∆UL2-HA, by inserting the HA gene of an H9N2 AIV virus in to UL2 gene, which could stably express the HA protein in pCEF. The insertion had no effect on the replication phenotype of DEV in the cells and did not alter the immunogenicity of DEV in ducks. When compared to parental DEV, rDEV-∆UL2-HA induced similar NT antibody titers and protective efficacy in ducks. More importantly, 103 TCID50 of recombinant DEV induced a high HI titer, leading to complete protection against challenge with H9N2 AIV virus. These results demonstrated that recombinant DEV may be a novel vaccine candidate to efficiently provide protection against both DEV and H9N2 AIVs in ducks.

The HI antibody titers against H9N2 virus induced by rDEV-∆UL2-HA in ducks were over 1:256 at 4 weeks p.v., three of which reached 1:4096. Interestingly, all vaccinated ducks generated comparable antibody levels. In contrast, the commercial inactivated vaccine H9 AIV induced an antibody titer of about 1:64. Moreover, the titer varied widely, with some ducks only generating titers ranging from 1:8–1:16 (unpublished data). Ducks inoculated with 102 TCID50 of rDEV-∆UL2-HA induced similar antibody levels; however, they showed only partial protection against H9N2 AIV virus challenge (6/10). A recombinant live virus expressing the HA gene is able to induce humoral and cellular immunity [9, 12, 23, 36]. It is not ascertained whether the different protective rates may be associated with cellular responses.

Viral vectors have been widely developed for vaccine use. Fowlpox virus, turkey herpesvirus, Marek’s diseases virus, and Newcastle disease virus are the most extensively studied viral vectors, and some recombinant vaccines have even been granted licenses [2, 3, 7, 10, 22–24, 28]. DEV, as a member of the herpesvirus family, is a promising candidate viral vector. Since Wang et al. first reported a recombinant DEV expressing the HA of H5N1 AIV, in which the H5 HA protein was detected by IFA and western blotting [29], recombinant DEVs generated by inserting the H5 HA gene at different positions have attracted much more attention [16, 17, 30, 36]. A single dose of the recombinant virus can induce HI antibody and provide rapid protection against lethal H5N1 AIV virus challenge [16, 36]. However, it remains unknown whether differing insertion positions for foreign genes in DEV can illicit varying degrees of protective efficacy. Until now, DEV-vectored IAV H9 vaccines have not been reported. Therefore, in this study, we generated a recombinant virus, rDEV-∆UL2-HA, by replacing the UL2 gene with the HA gene cassette. rDEV-∆UL2-HA showed protective efficacy against lethal DEV and H9N2 virus challenge in ducks, indicating that the UL2 region of DEV is suitable for foreign gene insertion.

In summary, rDEV-∆UL2-HA demonstrated good immunogenicity and a protective effect in ducks. Thus, rDEV-∆UL2-HA holds great promise for use in the poultry industry to prevent and control both DEV and H9N2 influenza in ducks.