**The Outcome of Porcine Foetal Infection with Bungowannah Virus Is Dependent on the Stage of Gestation at Which Infection Occurs. Part 2: Clinical Signs and Gross Pathology**

#### **Deborah S. Finlaison \* and Peter D. Kirkland**

Virology Laboratory, New South Wales Department of Primary Industries, Elizabeth Macarthur Agricultural Institute, Menangle, NSW 2568, Australia; peter.kirkland@dpi.nsw.gov.au

**\*** Correspondence: deborah.finlaison@dpi.nsw.gov.au

Received: 6 July 2020; Accepted: 3 August 2020; Published: 10 August 2020

**Abstract:** Bungowannah virus is a novel pestivirus identified from a disease outbreak in a piggery in Australia in June 2003. The aim of this study was to determine whether infection of pregnant pigs with Bungowannah virus induces the clinical signs and gross pathology observed during the initial outbreak and how this correlates with the time of infection. Twenty-four pregnant pigs were infected at one of four stages of gestation (approximately 35, 55, 75 or 90 days). The number of progeny born alive, stillborn or mummified, and signs of disease were recorded. Some surviving piglets were euthanased at weaning and others at ages up to 11 months. All piglets were subjected to a detailed necropsy. The greatest effects were observed following infection at 35 or 90 days of gestation. Infection at 35 days resulted in a significant reduction in the number of pigs born alive and an increased number of mummified foetuses (18%) and preweaning mortalities (70%). Preweaning losses were higher following infection at 90 days of gestation (29%) and were associated with sudden death and cardiorespiratory signs. Stunting occurred in chronically and persistently infected animals. This study reproduced the clinical signs and gross pathology of the porcine myocarditis syndrome and characterised the association between the time of infection and the clinical outcome.

**Keywords:** Bungowannah virus; foetus; pestivirus; porcine

#### **1. Introduction**

Bungowannah virus is a novel pestivirus identified from an outbreak of a disease in a piggery in New South Wales, Australia, in June 2003 [1]. It is genetically distinct from the other recognised pestiviruses of pigs, namely, classical swine fever virus (CSFV) and atypical porcine pestivirus (APPV) [2,3], with its closest genetic relationship to the recently identified Linda virus [4]. The disease was referred to as the porcine myocarditis syndrome, or PMC, because histological changes in the first affected animals consisted almost exclusively of multifocal, non-suppurative myocarditis, with myonecrosis observed in some cases. The outbreak initially presented as sudden death in 2–3-week-old weaning-age pigs, but soon after the onset, there was a marked increase in the birth of stillborn foetuses and a slight increase in the occurrence of mummified pigs [1]. Cumulative losses in some weeks exceeded 50% of pigs born, and it is estimated that as many as 50,000 pigs died in the initial outbreak. Due to the reproductive effects and disease becoming apparent almost exclusively in the first 2–3 weeks of life, it was presumed to be predominantly the consequence of in utero infection. This hypothesis was supported by the detection of elevated serum immunoglobulin G levels in up to 50% of stillborn pigs, and by the absence of disease in pigs after the immediate postweaning period or in sows farrowing affected litters [1].

The pestiviruses are well-recognised reproductive pathogens with the outcome of infection being dependent on a number of factors, including the stage of gestation that the infection occurs in relation to organogenesis and the development of immune competence [5–12]. CSFV has historically been the main pestivirus recognised to infect pigs (although infections with bovine viral diarrhoea virus (BVDV) and border disease virus (BDV) have been identified), and low virulence strains of CSFV may cause in utero infection and reproductive losses in the absence of disease in the rest of the herd [13–18].

Subsequent to the discovery of Bungowannah virus, two more pestiviral infections of pigs have been described: APPV [3] and Linda virus [4]. Both viruses have been associated with congenital tremors in piglets [4,10,19].

While there is strong epidemiological, virological [20] and preliminary experimental [21] evidence that Bungowannah virus is the causative agent of the PMC syndrome, the disease has not been reproduced experimentally. In an accompanying report [22], we described the virological and serological characteristics following the successful infection of the porcine foetus at four different stages of gestation. Regardless of the stage of gestation at which infection occurred, Bungowannah virus was detected in the serum/body fluid and excretions of infected pigs at birth and this was unrelated to the presence or absence of a precolostral Bungowannah virus-specific antibody in these animals. Persistent infections, as described for other pestiviruses, were observed following infection of the dam at 35 days of gestation, as was a chronic infection state where animals that had been presumed to be persistently infected (PI) following infection of the dam at 55 days later seroconverted. Viraemia, virus detection in tissues and viral shedding cleared more rapidly the later in gestation that infection occurred.

The aim of the current study was to determine whether infection of pregnant pigs with Bungowannah virus at different stages of gestation induces the clinical signs and gross pathology of PMC in the progeny that were observed during the initial outbreak. An in-depth description of the histopathological findings will be reported separately.

#### **2. Materials and Methods**

#### *2.1. Study Design*

Twenty-four pregnant pigs (22 gilts and two parity-1 sows) with known joining dates were obtained from a piggery known to be free of Bungowannah virus. Full details of the management of these pigs are described in a companion paper [22]. Pregnancy was confirmed using an ultrasound examination prior to selection for the study. The pigs were challenged intranasally at approximately 35 (34–36), 55 (55–58), 75 (72–76) or 90 (90–92) days of gestation (*n* = 6 per group—referred to as D35, D55, D75 and D90; Table 1). These time-points were selected as they were similar to those used in a previous study where foetuses were directly inoculated [21] and they span the gestational age at which the pig foetus is considered to become immunocompetent (70 days). All pregnancies were allowed to proceed to 113 days of gestation; then, all pigs were induced to farrow on day 114 to optimise the collection of blood from piglets prior to suckling.


*Viruses* **2020**

, *12*, 873

At birth, all pigs were weighed, their crown–rump lengths recorded, individually identified with ear tags and had their litter identity recorded. They were given an iron supplement intramuscularly between 2 and 4 days of age.

Farrowing outcomes, including the number of progeny born alive, stillborn or mummified, were recorded for each litter. Signs of disease observed throughout the study were documented.

Pigs from infected litters were weaned when they were 21–25 days old. As the presence of Bungowannah virus could still be detected at weaning in several challenge groups, as many pigs as the secure containment facilities could hold were weaned and retained until at least 8 weeks old. Selected pigs from D35 and D55 were kept in the study for a longer period to monitor clinical signs and virological and serological parameters [22]. Throughout the study, pigs that were moribund, not feeding or did not appear to be viable were euthanased.

Pigs from the three litters that did not become infected with Bungowannah virus (on the basis that the virus could not be detected in serum or body fluid for any of the pigs in the litter in a real-time reverse-transcription polymerase chain reaction (qRT-PCR) assay and negative precolostral serology results) were kept as control pigs until 14 to 21 days old, at which time they were euthanased. Because of the logistical considerations associated with the management of breeding sows in a secure containment facility, no mock-inoculated treatment group was included.

Individual pig IDs have been retained in the text where relevant to facilitate comparison between histopathology and virology and serology findings described in related manuscripts [22]. The first number relates to the litter ID and the second relates to the animal ID within the litter (generally in order of birth), e.g., 8-01 indicates the first pig to be born in litter 8.

All piglets were subjected to a detailed necropsy.

The animal studies were approved by the Animal Ethics Committee of the Elizabeth Macarthur Agricultural Institute, AEC Reference No. M09/02 (6 March 2009).

#### *2.2. Inoculum*

The pregnant pigs were challenged with a total of 5.3–5.8 log10 TCID50 of Bungowannah virus in 5 mL of phosphate-buffered gelatin saline (2.5 mL per nostril), with the exception of litter 11, where the sow only received approximately 3.8 log10 TCID50 of the virus. This inoculum was prepared as previously described [22,23].

#### *2.3. Statistics*

The confounding effect of treatment due to the separate cohorts was considered. The proportions of live, stillborn, mummified (of total born), weaned and lost/died pigs were analysed using a generalised linear model (GLM) as quasi-binomial (allowing for additional variation between sows). Each model included the time of challenge as a fixed effect. For the crown–rump length and birth weight data, a mixed model was fitted, with the time of challenge as a fixed effect, as well as random sow effects.

All calculations were performed with the R statistical software [24] using the glm function to fit the GLM model or the ASReml-R library [25] to fit the mixed models for length and weight.

#### **3. Results**

All 24 pigs challenged became infected. One pig from D55 failed to farrow and was not pregnant. As previously reported, the foetuses of 20/23 (87%) of the infected dams became infected with Bungowannah virus, as determined by the detection of Bungowannah virus nucleic acid using qRT-PCR in serum or body fluid at birth [22]. Within the 20 infected litters, 226 pigs (including stillborn and mummified foetuses) were born, of which, 225 had been infected in utero [22]. As all infected pigs were viraemic at birth, where individual animals are mentioned, reference is only made to their precolostral antibody level at birth. Three litters (*n* = 42 pigs) were uninfected at birth (as determined by negative qRT-PCR and precolostral serology results) and remained uninfected for the duration they remained in the study (14–21 days).

There were no significant differences in the crown–rump length (*p* = 0.594) and birth weight (*p* = 0.180) of pigs in the four treatment groups and uninfected litters.

#### *3.1. Incidence of Stillbirths, Preweaning Deaths and Mummified Pigs*

The day of gestation that the pregnant animals were challenged and their reproductive outcomes, including preweaning losses are summarised in Table 1. The greatest effects of infection were observed in the litters of dams infected at 35 or 90 days of gestation. While the use of cohorts confounded results, significant differences were associated with the time of challenge for the proportion born alive (of total) (*p* = 0.029) and proportion mummified (*p* = 0.008), where these values are lower and higher for the D35 group, respectively. For live-born pigs, there were significant differences between challenge groups in the proportion of preweaning losses (*p* = 0.004). The preweaning losses were significantly higher for D35 (there was also a numerically higher mean for D90 compared with D55 and D75; however, this was not statistically significant).

#### *3.2. Clinical Signs*

Table 2 summarises the reproductive effects, the likelihood of development of persistent infections and the most common clinical signs and necropsy findings observed in the progeny of sows following in utero infection with Bungowannah virus at different stages of gestation.


**Table2.**Summaryofthereproductiveeffects,andmostcommonclinicalsignsandnecropsyfindingsfollowinginuteroinfectionwithBungowannahvirus.

1 ↑—mild increase,

↑↑—moderate

 increase,

↑↑↑—marked

 increase,

↓↓↓—marked

 decrease. 2 [22].

#### 3.2.1. Group D35

The exposure of the pregnant animals at around 35 days of gestation resulted in a high percentage of infected pigs that were seronegative (96%) and presumably persistently infected at birth [22]. Clinically, a high proportion of stillborn (24%) and mummified pigs (18%) and a very high level of preweaning mortalities (70%) resulted (Table 1). The live-born pigs in these litters were often very weak, had difficulties moving from the rear of the sow after birth and some showed a limited ability to suckle. Many died soon after birth or were overlain, presumably as a consequence of their weakness or were euthanased due to their inability to feed. Additional abnormalities observed at birth in both stillborn and live-born pigs included purpura extending over the skin (*n* = 15) (Figure 1a,b) and/or subcutaneous oedema, particularly extending over the head, neck and ventral abdomen (*n* = 9) (Figure 1c). Pigs born with subcutaneous oedema had poor viability. All nine died soon after birth, were weak and were overlain or were euthanased on humane grounds within 3 days of birth. Mild subcutaneous oedema was noted in an additional two pigs at necropsy at 3 and 11 days of age. Of the four pigs with purpura that survived to be weaned, the purpura resolved with age and three of the four appeared to grow normally until weaning, although two developed conjunctivitis at approximately 3 weeks of age. They subsequently died or were euthanased when 26, 34, 35 and 57 days old. Several pigs exhibited neurological signs at birth, including walking into walls (14-04) and screaming, due to an inability to find the udder despite being able to suckle (most of litter 12). One pig (14-04) also developed haemorrhagic diarrhoea in the first week of life that resolved following antibiotic treatment. Three pigs developed diarrhoea in the first week of life, two of which were cross-fostered from their original mother due to milk supply issues, presumably due to poor udder stimulation from weak piglets and were ultimately euthanased (12-02, 12-05) by 2 weeks of age, and a third that had diarrhoea for 3–4 days that resolved following antibiotic treatment and then died suddenly 6 days later (16-02). One 2-week-old pig was euthanased as it failed to grow.

From the five infected litters, eleven pigs survived to weaning and were monitored until approximately 11 weeks old. Of these 11 pigs, nine were considered PI with Bungowannah virus based on ongoing viraemia and shedding of the virus throughout the study until death or euthanasia and negative precolostral serology at birth (14-01, 14-02, 14-04, 14-07, 14-09, 16-07, 17-01, 17-03, 17-04) [22]. The PI pigs grew poorly compared to their two cohorts (17-02, 17-06), which were not PI, based on their precolostral seropositive status at birth [22]. The PI pigs became stunted from soon after weaning (Figure 1d) and exhibited generalised skin pallor from 7 weeks of age. The first pig (17-03) was euthanased within 4 days of weaning after the sudden onset of recumbency; mouth breathing with a small amount of purulent nasal discharge; and purplish skin discolouration of the snout, ears, ventral abdomen and hind quarters (Figure 1e,f). Conjunctivitis was an ongoing problem in two pigs (17-01, 17-04). In addition, three pigs developed firm subcutaneous masses under their eyes between 5 and 9 weeks of age, which forced the lower eyelid dorsally, leaving them unable to open the affected eye (16-07, 17-01, 17-04) (Figure 1g); another developed an erosive lip lesion that resolved following antibiotic treatment at 6 weeks old (14-09). Investigation of the pallor diagnosed anaemia and only the PI pigs were affected. When sampled at 8 weeks old, the packed cell volume of the five affected animals ranged from 8% to 21% (reference interval for weaner pigs 26–41%; [26]). The low levels of circulating reticulocytes (0–0.4%) (corrected reticulocyte percentages <1%) and the normal mean corpuscular volume indicated that the anaemias were non-regenerative and normocytic. In addition, three animals were leukopaenic [26] and three were thrombocytopaenic. While one PI pig survived to 75 days of age (14-01), the remaining PI pigs that did not succumb to other problems were ultimately euthanased due to severe anaemia.

**Figure 1.** Clinical signs observed in D35 pigs: (**a**) purpura, subcutaneous oedema and absence of facial whiskers (12-09, stillborn); (**b**) purpura (17-11, stillborn); (**c**) subcutaneous oedema—most evident ventrally (12-10, 1 day old); (**d**) remaining pigs at 7 weeks of age with stunting of 6 smaller pigs compared with the 2 larger, non-PI pigs; (**e**) 26-day-old pig with purplish discolouration of the ears and snout (17-03); (**f**) same pig as (e) with purplish discolouration ventrally and on hind-quarters; and (**g**) 16-07 (62 days old) with mass under left eye, resulting in closure of the eyelids (arrow).

#### 3.2.2. Group D55

There were no clear reproductive effects observed following infection at approximately 55 days gestation (Table 1), with 82% of the infected pigs being seropositive at birth. Two of the stillborn pigs that were seronegative [22] exhibited mild subcutaneous oedema, although the significance is unclear as they were also moderately autolysed. In the immediate postnatal period (first 2 days of life), deaths were attributed to savaging (*n* = 3), overlay (*n* = 3) and runts (*n* = 1). Commencing at about one week after birth, in two litters, four pigs had dark faeces on a rectal swab that was consistent with melaena, suggesting upper gastrointestinal tract bleeding (7-10, 7-11, 10-04, 10-12) and one pig passed bloody liquid at the time of sampling (10-10) (the precolostral antibody titres for 10-04 and 10-10 were 10 and 80 respectively; the three other pigs had suckled prior to the initial sampling). Melaena was detected again on a rectal swab at 14 days of age in one of these pigs (7-11). This pig exhibited pallor, grew poorly compared to its littermates and was euthanased when it failed to gain weight and was unable to compete for food. Another pig (10-12) also grew poorly after melaena was detected at 6 days of age and was euthanased rather than weaned. The third pig (10-04) with melaena grew normally until 13 days old but then grew little in the following week, and at 20 days old, it exhibited pallor and mild jaundice. It was weaned at 28 days, and at this time, it was still growing poorly and exhibiting mild hind limb ataxia. At 19 days old, one pig (7-06; precolostral antibody titre of 1280) was pyrexic (40.7 ◦C), mildly ataxic with a slight tremor and showed retarded growth compared to its littermates. Despite antibiotic treatment, it was still pyrexic (40.2 ◦C) 2 days later and continued to exhibit neurological signs, including poor balance with a wide base stance; there was no obvious head tilt or nystagmus. It was noticed that one pig from litter 8 (8-01; seronegative at birth) appeared weaker than the other pigs in its litter in the first couple of days of life and was sometimes slow to move to the teat to feed.

In litters 7 (*n* = 4), 8 (*n* = 3, including 8-01 and 8-05) and 10 (*n* = 5, including 10-01), the pigs with the highest viral loads in serum and oropharyngeal secretions at 2 weeks of age, as well as all pigs surviving to weaning age from litter 11 (*n* = 10), were weaned to clarify whether the ongoing viraemia was due to a persistent infection, given that the dams were infected prior to the expected timing of foetal immunocompetence [22]. All weaned pigs were raised until at least 10 weeks old. Disease was not observed during this period, with the exception of occasional diarrhoea, and at 6.5 and 8 weeks of age, 10-04 had two episodes of pyrexia and lethargy, which responded well to antibiotic treatment on both occasions. The most dramatic clinical difference was the variation in size of the littermates from litter 8 (Figure 2) and litter 10 when 11 weeks old. Across both of these litters, five pigs were live born and seronegative for Bungowannah virus antibody [22]. The two small pigs (Figure 2; 8-01 and 8-05) appeared to be PI on the basis of qRT-PCR results and the absence of antibodies against Bungowannah virus at birth and were markedly stunted compared to their non-PI littermates (seropositive at birth). When 6 to 7 months old, Bungowannah virus RNA could no longer be detected in the serum of these stunted animals; they concurrently developed high antibody titres and their growth rate improved markedly and they are described here as chronically rather than persistently infected [22].

One of the stunted pigs was reared until 6 months of age (10-01), and another two until 11 months of age (8-01 and 8-05). At approximately 3.5 months old, 8-01 and 10-01 experienced an episode of diarrhoea due to a *Salmonella* infection, but otherwise, no disease was observed. At 189 days of age, pig 10-01 weighed 105 kg; at 11 months of age, 8-01 and 8-05 weighed 183 and 153 kg respectively.

**Figure 2.** Littermates from litter 8 in D55 at 77 days old—note the stunting of two chronically infected littermates (8-01 and 8-05) compared with the larger non-persistently or chronically infected animal (8-03).

#### 3.2.3. Group D75

Losses in the preweaning period were the result of savaging (*n* = 18), overlay (*n* = 2), splay leg (*n* = 1), runt (*n* = 1) and poor mothering/milk production (*n* = 2) and occurred in the first week of life (Table 1). Pigs surviving to weaning did not exhibit any signs of disease and 98% of the infected pigs were seropositive at birth [22].

After weaning, several pigs lost condition compared to others in the group and one developed a skin infection. Otherwise, no disease was observed. No postweaning mortalities occurred and 23 pigs in this group were reared up to 8 weeks of age.

#### 3.2.4. Group D90

Preweaning losses (29%) in this group were higher compared to the D55 and D75 groups (Table 1) and were clinically different from the D35 group. Only 50% of the infected pigs were seropositive at birth, presumably because of the short interval between infection and birth [22]. While some losses due to savaging (*n* = 4) and overlay (*n* = 1) were recorded, some pigs also exhibited tachypnoea and dyspnoea prior to death or euthanasia. Sudden death was recorded for seven apparently healthy pigs between 2 and 5 days of age without any prior signs (19-03, 19-07, 20-04, 22-03, 22-05, 22-06 and 22-08).

Six pigs developed a notable increase in respiratory rate and effort (19-02, 20-06, 21-02, 21-03, 21-09 and 23-11). The earliest onset of respiratory signs was at 4 days and the latest at 7 days old. Three pigs were euthanased due to the severity of their clinical signs. Two pigs, one of which was thinner than its littermates, died 5 days after the detection of an increased respiratory rate. Finally, one pig continued to exhibit an increased respiratory rate when stressed and was thinner than its littermates through to weaning and subsequently to 5 weeks age.

Twenty pigs from this group were weaned and reared up to 5 weeks of age. No disease was evident in these animals.

#### 3.2.5. Uninfected Litters

No disease was observed with the exception of diarrhoea in one pig. The only preweaning losses were due to overlay (*n* = 2), splay leg (*n* = 3), runt (*n* = 1) and accident (*n* = 1) (Table 1). While there was no statistical difference in birth weight between the infected and uninfected litters, the mean weight for the uninfected litters at approximately 7 and 14 days of age was 0.5–0.8 kg and 0.6–1.1 kg greater, respectively.

#### *3.3. Necropsy Findings*

Table 2 summarises the most common gross pathology findings observed following in utero infection with Bungowannah virus at different stages of gestation. The viral load in tissues at necropsy was dependent on the stage of gestation that infection occurred in, whether the pig was able to mount an immune response in utero and the age of the pig at necropsy [22].

#### 3.3.1. Group D35

Subcutaneous oedema was observed in nine pigs and was most commonly observed ventrally and around the head and neck, giving rise to palpebral oedema and the impression of a Roman nose (Figure 1a,c). Petechiae were observed on the hearts of two stillborn pigs. Purpura (pigs 0 to 3 days old) were most easily identified on the ears, head, ventral abdomen and sometimes on the dorsal body (Figure 1a,b). Petechiae were observed on the pleural surface of the lungs of an overlain pig. White pin-point foci were observed throughout the brain of one stillborn pig (12-09) and a region of presumptive necrosis was observed in the left cerebral hemisphere of another (17-11). Other observations at necropsy in the preweaning period included haemorrhages on the tonsils (*n* = 2) and tongue (*n* = 1), mottled orange liver (*n* = 3) and the absence of facial whiskers (*n* = 1). Red mottled lungs were commonly observed and presumptively attributed to hypostasis. Thoracic and abdominal effusions in stillborn pigs were associated with autolysis. Fibrinous adhesions were observed in the thorax and or abdomen of two pigs that died at 12 and 15 days (both had previously been treated for diarrhoea).

A range of lesions was observed in the PI pigs that survived past weaning. These animals remained viraemic throughout their life, shed high levels of the virus and had high quantities of the virus in all tissues examined at their necropsies [22]. All became severely stunted, with weights ranging at time of euthanasia from 4.0 kg at 34 days to 13.2 kg at 77 days. Two animals (14-04, 14-09) showed evidence of a bleeding disorder, including the presence of clots in the abdominal cavity and serosanguinous fluid/effusion in the thoracic and/or abdominal cavities; blood streaks were observed in the stomach contents of one of these two pigs. Other changes that indicated a bleeding disorder or vasculitis in the PI pigs included red/haemorrhagic inguinal, mesenteric and/or lumbar lymph nodes in four animals; petechiae were occasionally observed on a number of organs, including the heart (*n* = 3), kidney (*n* = 2) (Figure 3a), liver (*n* = 1), serosal surface of the intestines/colon (*n* = 2), peritoneal surface (*n* = 1), cerebellum (*n* = 1) and oropharynx (*n* = 1). Other abnormalities observed included ulcerated tonsils (16-07) (Figure 3b), increased pericardial fluid (*n* = 4; up to 20 mL), cobblestone pattern to the liver (*n* = 1), mild pulmonary interlobular oedema (*n* = 4) and fibrin tags on intestinal serosa (*n* = 1). The facial mass observed in three pigs (Figure 1g) was firm with an intimate connection between the skin and the subcutis. There was no evidence of wounds in the skin or oral cavity connecting with the mass. Mild subcutaneous oedema was detected in the neck of one pig.

No abnormalities were detected in the two pigs that were not PI (17-02 and 17-06) [22]. At 75 days of age, their mean weight was 30.75 kg. Only low quantities of Bungowannah virus were detectable in these two animals at necropsy [22].

**Figure 3.** Necropsy findings from D35: (**a**) petechiation of a kidney (17-03—same pig as Figure 1e,f) and (**b**) ulcerated tonsil (16-07).

#### 3.3.2. Group D55

Serosanginous pleural and abdominal fluid, as well as mild subcutaneous oedema, was noted in two stillborn pigs that exhibited autolytic changes. Pericardial petechiation was observed in one stillborn pig. Mild interlobular pulmonary oedema and congestion was observed in the lungs of six animals (ages 2–22 days) and was considered to be due to hypostasis. One animal had fibrinous pleural and abdominal adhesions, and another, fibrinous adhesions between the spleen and the liver; both suffered a preweaning death. Other lesions included a cobblestone pattern on the liver, kidney enlargement, oedematous spiral colon and dark blood clots in the colon.

The two chronically infected pigs (8-01 and 8-05) [22] were reared to 11 months of age. Few abnormalities were observed at necropsy. The testes of 8-01 were soft on palpation, fluid collected from the epididymis was of low opacity and sperm were not observed via a microscopic examination, suggesting infertility. The highest viral load detected throughout this study was in the epididymal semen of this animal (9.8 log10 copies/mL) [22]. Pig 8-05 was female and penned with 8-01 from 4 months old. From 8 months old, she was observed in oestrus approximately every 3 weeks and did not become pregnant. This was confirmed at necropsy, and an examination of her ovaries showed many follicles and multiple corpora lutea, indicating reproductive maturity.

#### 3.3.3. Group D75

No significant gross pathology was observed for this group. Occasionally petechiae in the heart wall (*n* = 2; stillborn pigs) and consolidated lung lobules (*n* = 6; presumptively due to hypostasis) were observed. Serosanginous pleural and abdominal fluid was noted in two stillborn pigs that exhibited autolytic changes. Bungowannah virus was cleared relatively rapidly from this group, with only low viral loads generally detected after 21–28 days of age [22].

#### 3.3.4. Group D90

Serosanguinous thoracic or abdominal effusions were observed in stillborn pigs with autolytic changes (*n* = 3). Occasional petechiation of the pericardium was also observed (*n* = 1). Two stillborn pigs had fibrin tags in the peritoneal cavity and equivocally enlarged hearts (18-07, 18-08), while another had red tonsils.

Of the preweaning mortalities, four of the five pigs observed to have increased respiratory rates exhibited cardiomegaly (20-06, 21-02, 21-03, 21-09) based on subjective observation of an increased base to apex height and increased sternal contact (Figure 4a), and two 5 mm white foci were observed in the superficial myocardium of 20-6. The fifth pig had a marked haemorrhagic pericardial effusion (>50 mL) and blood clots in the abdominal cavity (19-02) (Figure 4b). Thoracic and/or abdominal effusions (*n* = 2), fibrin tags in the thoracic and/or abdominal cavity (*n* = 2), liver enlargement (*n* = 4), reddened inguinal and mesenteric lymph nodes (*n* = 3) and oedema of the spiral colon (*n* = 3) were also observed amongst these five cases. The lungs of the most severely affected pigs did not deflate and a large amount of fluid/froth drained from the lungs after death. Viral RNA was detected in the hearts of all pigs sampled in the first 10 days of life [22]. The pigs that died suddenly (19-03, 19-07, 20-04, 22-03, 22-05, 22-06 and 22-08) appeared to be behaving normally and eating well prior to death. Bruising to suggest they had been overlain was not observed. A range of changes was observed, including subcutaneous oedema of the ventral neck (*n* = 2), pericardial effusion (*n* = 1), fibrin tags in thoracic and/or abdominal cavities (*n* = 2), blood clots in the abdominal cavity (*n* = 2), petechiae and haemorrhages on the kidney (*n* = 1), red/haemorrhagic mesenteric and/or inguinal lymph nodes (*n* = 3), haemorrhages on the spiral colon and caecum (*n* = 2) and oedema of the spiral colon (*n* = 1).

**Figure 4.** Necropsy findings from D90: (**a**) cardiac and hepatic enlargement (20-06; 7 days old) and (**b**) free blood in the unopened pericardial sac (arrow) and abdomen (19-02; 12 days old).

From weaning to termination of the monitoring at 5 weeks of age, few gross lesions were observed. Bungowannah virus was cleared most rapidly from this group with virus shedding and the viral loads in tissues were only at low levels from 11 days of age or often not detected at all [22]. Interlobular pulmonary oedema was observed in two animals and fine fibrin tags were observed in the peritoneal cavity of five animals. The pig that exhibited an increased respiratory rate after exercise and survived to 5 weeks old (23-11) also showed mild interlobular pulmonary oedema and a smaller than expected thymus, but did not have definite cardiac enlargement. However, the ventricular walls appeared thinner than normal, suggesting possible cardiac dilation.

#### 3.3.5. Uninfected Litters

No significant abnormalities were observed at necropsy in pigs from this group. The ages of pigs examined ranged from stillborn to 21 days old. The most frequent changes observed were reddening of the lungs (presumptively due to hypostasis), with some interlobular oedema in four pigs. Other changes observed included petechial haemorrhages on the pericardium, tonsil and skull of an overlain pig, a moderate increase in the amount of thoracic fluid of stillborn pigs, circular skin deficits over the left carpus of two stillborn pigs in litter 9 and moderately enlarged inguinal lymph nodes in a splay-legged pig.

#### **4. Discussion**

In this study, the clinical signs and gross pathological lesions observed during the outbreak of the porcine myocarditis syndrome were reproduced in the progeny of pregnant pigs experimentally infected with Bungowannah virus. The disease and lesions associated with in utero Bungowannah virus infection at different gestational ages were further characterised. Infection of the foetus at 35 or 90 days gestation resulted in the most severe clinical and pathological effects.

Losses in the field due to PMC were characterised by increased stillbirths, preweaning losses and to a lesser extent, mummified foetuses. At the peak of the outbreak on the farm, on the most severely affected production unit (a gilt unit), preweaning losses approached 50%, stillbirths 40% and mummified foetuses 10–15%, although losses did vary across production units [1,27]. The results of the current study indicated that Bungowannah virus is capable of causing these effects and, depending on the presence of sows at critical stages of pregnancy, could cause significant reproductive losses if introduced into a naïve piggery. The two most critical time-points for exposure of the pregnant pig in relation to disease appears to be around 35 and 90 days of gestation. Our observations following infection at these stages of gestation are consistent with events observed during the early stages of the PMC outbreak, which suggests that the introduction of the virus occurred at a single time-point. The first losses observed were sudden death at 2–3 weeks of age, which is consistent with the outcomes of experimental infection at around 90 days gestation. These initial losses were followed approximately 4 weeks later by a marked increase in the number of stillborn pigs, which would indicate infection at around 40–50 days gestation if 90 days is considered the critical time-point for sudden death at 2–3 weeks of age [20,27].

The description of porcine myocarditis syndrome arose from the fact that early in the field outbreak, affected animals often had an enlarged/dilated heart with evidence of congestive heart failure and non-suppurative myocarditis [1]. Based on the results of this study, it appears that the cardiac enlargement previously reported occurs principally following infection in late gestation (approximately 90 days) and results in a moderate increase in preweaning losses. This was the only challenge group where the previously described [1] multifocal, non-suppurative myocarditis with myonecrosis was observed [28]. Based on these findings, it is clear that the outcome of infection with Bungowannah virus and the associated pathological changes are more extensive than the original field description "porcine myocarditis syndrome" would suggest. Clinically, similar to the field cases, sudden death or increased respiratory rate and effort were observed. Unlike the field situation, cyanosis of the snout and ears and excessive vocalisation before death were not observed during the course of this study. The cardiorespiratory signs and sudden deaths observed have not been reported for CSFV, APPV or Linda viruses, the other pestiviruses primarily infecting pigs. Postnatal deaths are uncommon following infection after 90 days gestation with low virulence strains of CSFV [13,29].

The reproductive effects of in utero infection with Bungowannah virus were most severe following infection around 35 days gestation and are presumably a consequence of the foetus being unable to mount an immune response to the virus at this stage of gestation [22]. These effects included a marked increase in preweaning mortalities (particularly in the first few days of life) and mummified foetuses, as well as a moderate to marked increase in stillbirths, all of which suggests that these effects in the field were the result of infection around 35 days of gestation, reducing by day 55. These clinical findings are similar to those observed following in utero infections with low virulence strains of CSFV that, prior to 70 days gestation, are characterised by high prenatal mortality due to stillborn and mummified foetuses, and increased postnatal mortality with weak-born pigs, although congenital deformities were not observed for Bungowannah virus [13,29–31]. The increased levels of stillbirths following exposure of the dam at 55, 75 or 90 days gestation during this study, when compared with industry targets, suggests that Bungowannah virus may have also caused a mild to moderate increase in the stillbirth rate at these time-points, although a greater number of pregnant animals would need to be studied to confirm this finding.

Subcutaneous oedema that spread principally over the head and thorax was observed in a proportion of stillborn pigs during the PMC outbreak [1]. Experimentally, this clinical sign was only observed following Bungowannah virus infection in early gestation (approximately 35–55 days). While purpura was not reported in the field, it was observed in 31% of live and stillborn pigs born to dams infected at around 35 days gestation in the current study. Purpura/petechiae/haemorrhages and subcutaneous oedema have also been described following in utero infection with CSFV in foetuses infected before 70 days gestation [13,29–31].

Other clinical similarities between infection with Bungowannah virus at 35 days gestation and in utero CSFV infection prior to 70 days include weak-born pigs that may exhibit or develop poor sucking ability, alopecia, ascites/hydrothorax, petechiae/haemorrhages in organs other than the skin, pallor, diarrhoea, conjunctivitis, anorexia, lethargy or cyanosis of the snout and ears, with some of these changes being observed postweaning [13,29–31]. In contrast, the clinical signs following experimental Bungowannah virus infection at 55 days more closely parallel what is observed when in utero infection with CSFV occurs between 70 and 90 days gestation, which is associated with fewer prenatal deaths [29–31], although some melaena and diarrhoea was observed in the preweaning period. It is not clear whether the diarrhoea and melaena were increased compared to D75 and D90 pigs due to in utero Bungowannah virus infection or is unrelated.

Following the initial postnatal losses in the D35 group, the next major clinical effect noted following Bungowannah virus infection at either 35 or 55 days of gestation was variability in pig size within a litter postweaning. This disparity in size was noted in those pigs that were born seronegative to Bungowannah virus and, on the basis of PCR results, were considered to be persistently infected [22]. Persistent infections with CSFV are recognised following in utero infection before approximately 70 days gestation and with decreasing frequency up to 90 days gestation [13,29,30]. Clinically, CSFV PI pigs that survive past weaning may present similarly to chronic cases of classical swine fever (CSF), where after an initial, albeit slow recovery, they eventually relapse and die. Initially, they may be indistinguishable from uninfected pigs, but at a variable time (usually weeks to months) after weaning, they often develop severe runting and growth retardation, or "late-onset CSF", characterised by increasing anorexia and lethargy, pyrexia, conjunctivitis, dermatitis, locomotion disturbances and intermittent diarrhoea [13,15,30,31]. These clinical signs were generally consistent with those observed in pigs infected with Bungowannah virus at D35 that were considered PI. With the exception of a slightly slower growth rate, these pigs often appeared normal while still suckling, but at a variable time after weaning, developed severe growth retardation. Pallor (due to anaemia) and conjunctivitis were the most frequently observed clinical signs. Ataxia and locomotion disturbances were not

observed. In contrast, those pigs that were infected at D55 and were born viraemic, seronegative and survived past weaning, while clearly stunted compared to their littermates, were generally disease-free. Whether this would be the case in the field, where there is greater exposure to pathogens, is not clear.

Two of the pigs (8-01 and 8-05) in the D55 group that were initially considered PI were raised until 11 months of age. Unexpectedly, these two animals seroconverted at approximately 6 months of age and cleared their viraemia, at which point, their growth rate increased [22]. After seroconversion, the gilt started cycling at around 8 months of age, which is older than expected and was presumably delayed due to her stunted growth. These animals are similar to the chronically infected pigs described when pigs were experimentally infected in utero with BDV or BDV-like viruses, or BVDV [14,16]. Similar to our study they also noted an improved growth rate after seroconversion. The reason for this improvement is not clear but it is suggested that the effect on growth is the result of a continuing process and not of an irreversible lesion acquired in early development [14].

Stunting is also observed in cattle PI with BVDV, although some animals may appear normal for several years before succumbing to a disease predominantly induced as a result of the persistent BVDV infection. While not observed during the course of this study, it is important to determine whether pigs persistently or "chronically" infected with Bungowannah virus can remain clinically normal for prolonged periods, as these animals are an ongoing source of infection in the herd and their selection as breeding animals could result in disastrous consequences if they were to be introduced into a naïve herd.

Congenital tremors were not observed in this study or in the field following Bungowannah virus infection. Aside from the absence of congenital tremors, the outcomes of infection of Bungowannah virus and CSFV after foetal infection in early gestation are remarkably similar and presumably have a similar pathogenesis. To date, Bungowannah virus has only been identified in Australia, but as its origin remains unknown, it should be considered in the differential diagnosis in reproductive investigations, especially if CFSV is suspected but has been excluded by laboratory testing.

It was not possible to weigh live animals once they reached 5 kg; therefore, all weights postweaning were obtained at the time of necropsy. In retrospect, weekly weights would have been useful to compare animal weights within litters and in relation to infection and serological status at birth. Any comments on stunting in this manuscript are unfortunately subjective rather than objective and relative to that challenge group cohort/littermates only. Additionally, while qRT-PCR results were typically available within a couple of days after sample collection, the serology (peroxidase-linked assay) was batch-tested at a convenient time [22]. The availability of these results may have been useful when selecting animals to follow postweaning, particularly for the D55 group, to better understand the impact of these "chronic" infections on growth rates. As each treatment group was reared individually and at separate times, direct comparisons regarding weight are difficult; in addition, direct comparisons against commercial growth targets are not appropriate.

#### **5. Conclusions**

This study demonstrates that in utero infection with Bungowannah virus was able to produce the clinical signs and gross pathology observed during the field outbreak and has the potential to be a significant reproductive pathogen of the pig, as well as impacting on postnatal mortality, principally in the first few weeks of life. The outcome of infection was dependent on the stage of gestation at which infection occurs. Maternal infection at around 35 days gestation resulted in high pre- and postnatal mortalities, which reduced when infected at about day 55. Reproductive effects and clinical signs were similar to those observed following in utero infection with low virulence strains of CSFV. Increased preweaning mortalities, cardiorespiratory clinical signs and gross lesions in the heart of the progeny of sows infected at around 90 days gestation suggests that infection in late gestation was required for the development of cardiac pathology. Persistent infections resulted in the development of clinical signs and lesions similar to "late-onset" CSFV.

**Author Contributions:** Conceptualization, D.S.F. and P.D.K.; methodology, D.S.F. and P.D.K.; investigation, D.S.F.; resources, D.S.F. and P.D.K.; data curation, D.S.F.; writing—original draft preparation, D.S.F.; writing—review and editing, D.S.F. and P.D.K.; visualization, D.S.F. and P.D.K.; supervision, P.D.K.; project administration, D.S.F. and P.D.K.; funding acquisition, P.D.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Australian Pork Limited (Project No. 2188), the Australian Biosecurity Cooperative Research Centre for Emerging Infectious Disease and NSW Department of Primary Industries.

**Acknowledgments:** We gratefully acknowledge the staff of the Virology Laboratory at the Elizabeth Macarthur Agricultural Institute (EMAI) for technical assistance and assistance with sample collection, and both farm and laboratory staff, especially Glenda Macnamara, for their care of the animals used in this study. We thank Damien Collins for assistance with statistical analysis, and Sarah Gestier for constructive feedback on the gross pathology descriptions in the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Infection of Ruminants, Including Pregnant Cattle, with Bungowannah Virus**

#### **Andrew J. Read, Deborah S. Finlaison and Peter D. Kirkland \***

Virology Laboratory, Elizabeth Macarthur Agriculture Institute, Woodbridge Road, Menangle, New South Wales 2568, Australia; andrew.j.read@dpi.nsw.gov.au (A.J.R.); deborah.finlaison@dpi.nsw.gov.au (D.S.F.)

**\*** Correspondence: peter.kirkland@dpi.nsw.gov.au; Tel.: +61-2-4640-6331

Received: 2 June 2020; Accepted: 22 June 2020; Published: 26 June 2020

**Abstract:** Bungowannah virus is a pestivirus known to cause reproductive losses in pigs. The virus has not been found in other species, nor is it known if it has the capacity to cause disease in other animals. Eight sheep, eight calves and seven pregnant cows were experimentally infected with Bungowannah virus. It was found that sheep and calves could be infected. Furthermore, it was shown that the virus is able to cross the bovine placenta and cause infection of the foetus. These findings demonstrate the potential for species other than pigs to become infected with Bungowannah virus and the need to prevent them from becoming infected.

**Keywords:** Bungowannah virus; pestivirus F; ruminant infection

#### **1. Introduction**

In June 2003 a syndrome of sudden death in sucker pigs, followed by a marked increase in stillborn foetuses and pre-weaning losses, occurred on a large farm in NSW, Australia [1]. Myocarditis and myonecrosis were also observed in affected pigs. A novel pestivirus, known as Bungowannah virus, was subsequently identified [2]. A series of field and laboratory studies have provided strong evidence that Bungowannah virus is the aetiological agent in this disease [2–8]. Bungowannah virus contains all of the genomic and structural elements of classically described pestiviruses, yet phylogenetic analysis demonstrates that it is genetically remote from any of the other pestivirus species [9,10]. Bungowannah virus is the only extant isolate of *pestivirus F* species [10].

Pestiviruses were initially classified according to their host specificity. Whilst this classification was originally appropriate for classical swine fever virus (CSFV), it was soon shown that bovine viral diarrhea virus (BVDV) and border disease virus (BDV) could naturally infect a variety of ruminants, pigs and other mammals. Recently, CSFV has also been shown to naturally infect cattle [11]. In contrast, Bungowannah virus has only ever been detected in pigs. The origin of this virus is not known, nor what threat it may pose to other species. Bungowannah virus has been shown to replicate in ovine and bovine cells in vitro [12] and so the possibility that it may infect ruminants has been raised. This paper documents the outcome of experimental infections of sheep and cattle with Bungowannah virus. Patterns of virus shedding and pathology are described.

#### **2. Materials and Methods**

A series of inoculation experiments were conducted in both sheep and cattle. Cattle were either directly inoculated using intranasal instillation or by co-housing with pigs that were chronically infected with Bungowannah virus. Sheep were either directly inoculated using intranasal instillation or subcutaneous injection or by co-housing with pigs that were chronically infected with Bungowannah virus. The specific details are as follows:

#### *2.1. Virus Amplification*

The inoculum used for each of the direct inoculation experiments was derived from pooled pig foetal tissues that were passaged once in PK-15 cells (RIE5–1, Collection of Cell Lines in Veterinary Medicine, Friedrich-Loeffler-Institut, Insel Riems, Germany). The titre of infectious virus was also determined by titration in PK-15 cells using standard methods.

#### *2.2. Viral Transport Medium*

Swabs were collected into 3 mL of sterile phosphate buffered saline (137 mM NaCl, 8 mM Na2HPO4, 2.7 mM KCl and 1.5 mM KH2PO4, pH 7.4) containing 0.5% gelatin (*w*/*v*), 5000 IU penicillin/mL, 95,000 IU streptomycin, 50 μg/mL amphotericin B and 0.1% (*w*/*v*) phenol red (PBGS).

#### *2.3. Bungowannah Virus Real-Time Polymerase Chain Reaction (qRT-PCR)*

Bungowannah virus RNA was identified from samples using a real-time, reverse transcription PCR (qRT-PCR). The method has been previously described [3]. The fluorescence threshold was set manually at 0.05 and the background was automatically adjusted. qRT-PCR results were expressed as cycle threshold (Ct) values and classified as negative if no amplification was observed after the 45 cycles. For quantification, a 10-fold dilution series of Bungowannah virus RNA standards ranging from 107 to 102 RNA copies/5 μL [6] was included in the assay and the quantity of Bungowannah virus RNA in a sample was determined from the standard curve.

#### *2.4. Bungowannah Virus Neutralisation Test*

Antibody titres against Bungowannah virus were measured by virus neutralisation test (VNT). The VNT was performed as described previously [5]. Selected serum samples were tested in the VNT in a two-fold dilution series commencing at 1/4.

#### *2.5. Infection of Sheep*

Sheep used in these trials were obtained from a flock that was free of infection with ruminant pestiviruses and had not been vaccinated against pestiviruses. All sheep were tested for anti-pestivirus antibodies using a bovine viral diarrhea virus agarose gel immunodiffusion assay [13] and were found to be negative.

#### 2.5.1. Direct Inoculation

Six 3-month-old Merino lambs were infected intranasally with 2 mL of cell culture amplified Bungowannah virus (5.6 log10 TCID50/mL). Two other sheep were inoculated with the same dose subcutaneously while another two other sheep were held as uninfected controls. The inoculated sheep were held in two 11 m2 rooms (four intranasally infected sheep in one room, the remaining four infected sheep in the other room). The two uninfected sheep were held in a similar 11 m2 room and were not challenged.

Conjunctival, nasal, oral and rectal swabs, along with serum samples, were collected from all sheep prior to exposure to Bungowannah virus and daily for 14 days. Blood samples were subsequently collected approximately weekly until 6 weeks post-exposure. Clinical signs, including rectal temperatures, were also recorded daily for the first 14 days. The swabs and sera were tested for the presence of Bungowannah virus using real-time PCR (qRT-PCR). Serum samples were tested for the presence of antibodies against Bungowannah virus using a VNT.

#### 2.5.2. Exposure to Chronically Infected Pigs

Four 3-month-old Merino lambs were held in a small room (16 m2) with three pigs that were chronically infected with Bungowannah virus. The pigs (08-01, 08-05 and 10-01) had been infected in utero and were shown to be shedding Bungowannah virus (oropharyngeal secretions—6.9, 7.0 and 3.4 log10 copies/swab, respectively) 7 days prior to the trial [7]. The sheep and the pigs were co-housed for 48 h. During this time there were two periods of six hours of direct physical contact between the sheep and pigs. During the remainder of the time the pigs were separated from the sheep by a mesh partition which allowed for limited direct contact. The clinical examination and sampling were conducted as described above.

#### *2.6. Infection of Calves*

Calves used in the following trials were obtained from a herd that was free of infection with ruminant pestiviruses and had not been vaccinated against pestiviruses. All calves were tested for anti-pestivirus antibodies using a bovine viral diarrhea virus agarose gel immunodiffusion assay [13] and were found to be negative.

#### 2.6.1. Direct Inoculation

Eight 10-week-old Holstein–Friesian calves were infected intranasally with 2 mL of cell culture-amplified Bungowannah virus (5.6 log10 TCID50/mL). Two additional calves were held as uninfected controls and did not receive a challenge. The conditions under which they were held, the clinical examination and sampling were conducted as described above for the sheep.

#### 2.6.2. Exposure to Chronically Infected Pigs

Four 5-week-old Holstein–Friesian calves were held with two pigs (08-01 and 08-05; oropharyngeal secretions 5.5 and 7.1 log10 copies/swab respectively, 6 days prior to trial) [7] chronically infected with Bungowannah virus. The conditions under which they were held, the clinical examination and sampling were conducted as described above for the sheep.

#### *2.7. Infection of Pregnant Cows*

Five pregnant Holstein-Friesian and two Illawarra-Shorthorn cows were chosen for the trial. They were obtained from a herd known to be free of infection with ruminant pestiviruses and had not been vaccinated against pestiviruses. All cows were tested for anti-pestivirus antibodies using a bovine viral diarrhea virus agarose gel immunodiffusion assay [13] and found to be negative. They were infected by intranasal instillation of 2 mL of the cell culture amplified Bungowannah virus (4.5 log10 TCID/mL). Cows were between 53 and 65 days of pregnancy at the time of inoculation. Nasal and conjunctival swabs and serum samples were collected daily for 14 days following inoculation. The swabs and sera were tested for the presence of Bungowannah virus using qRT-PCR. Serum samples were then collected monthly until calving and were tested for antibodies against Bungowannah virus using a VNT. Pregnancy was monitored by ultrasound examination on a monthly basis. Cows were induced to calve between 255 and 276 days of pregnancy. Serum samples were collected from each calf after birth and prior to suckling. Conjunctival, nasal, oral and rectal swabs and serum samples were collected from calves every 48–72 h for 14 days. Vaginal swabs were also collected from the cows post-partum. A post-mortem examination of the calves was performed between 2 and 4 weeks of age. A wide range of tissues including brain, myocardium and skin (and testicle from a male) were tested for the presence of Bungowannah virus by qRT-PCR. Samples for qRT-PCR were collected by firmly rubbing a swab across the freshly cut surface of a section of the tissue and placed into 3 mL of PBGS. Skin biopsies were stored in 3 mL PBGS. All samples were stored at 4 ◦C until tested by qRT-PCR and virus isolation.

#### *2.8. Animal Ethics Approval*

The trials described in this paper were approved by the Elizabeth Macarthur Agricultural Institute Animal Ethics Committee. The specific approvals were AEC M09/17 "Studies of the biology of Bungowannah virus infections (PMC) in sheep and cattle" (7 December 2009) and AEC M10/16 "Effects of Bungowannah virus infection in pregnant cattle" (22 December 2010).

#### **3. Results**

#### *3.1. Infection of Sheep*

#### 3.1.1. Direct Inoculation

Bungowannah virus RNA was detected between days four and 11 in the nasal swabs of five of the six sheep exposed by intranasal instillation. Ct values ranged between 27.9 and 39.9 (Table 1). Low levels of Bungowannah virus RNA were sporadically detected between days three and 13 of infection in serum samples from four of the six intranasally inoculated sheep. Ct values ranged between 36.5 and 38.1. Two of these sheep had Bungowannah virus RNA detected in serum on three occasions (Table 1).


**Table 1.** Bungowannah virus qRT-PCR results from sheep intranasally and subcutaneously inoculated.

A dash indicates that Bungowannah virus RNA was not detected.

Bungowannah viral RNA was not detected in oral, conjunctival or rectal swabs of any of the inoculated or control sheep, nor in serum or nasal swabs for the control sheep or those inoculated subcutaneously. None of the infected sheep developed clinical signs during the course of the infection. The rectal temperatures remained within normal limits. All eight sheep directly inoculated (either intranasally or subcutaneously) developed Bungowannah virus-specific antibodies (Table 2).


**Table 2.** Bungowannah VN titres from sheep intranasally and subcutaneously inoculated.

A dash indicates that neutralising anti-Bungowannah virus antibodies were not detected.

#### 3.1.2. Exposure to Chronically Infected Pigs

Viral RNA was detected only in the nasal swabs of three of the four sheep during the period of co-mingling with the chronically infected pigs. The Ct values ranged from 36.1 to 36.4. Bungowannah virus RNA was not detected in serum or any other swabs during this time and was not detected in serum or any swabs after the pigs were removed from the room. None of the sheep developed Bungowannah virus-specific antibodies.

#### *3.2. Infection of Calves*

#### 3.2.1. Direct Inoculation

Bungowannah virus RNA was detected in nasal swabs of all eight calves on at least three occasions between days two and 11 (Table 3). Viral RNA was detected sporadically in serum samples from six of these eight calves between days five and 10 of infection and intermittently in the oral swabs from four calves between days four and eight. Bungowannah virus was not detected in any rectal swabs. All eight directly inoculated calves developed antibodies directed against Bungowannah virus by 14 days post-inoculation (Table 4). A very mild nasal discharge was observed in five of the challenged calves. The rectal temperatures remained within normal limits.


**Table 3.** Bungowannah virus qRT-PCR results from calves intranasally inoculated.


**Table 3.** *Cont.*

A dash indicates that Bungowannah virus RNA was not detected. Calves 9C and 10C were uninfected.


**Table 4.** Bungowannah VN titres from calves intranasally inoculated.

A dash indicates that neutralising anti-Bungowannah virus antibodies were not detected. Calves 9C and 10C were uninfected.

#### 3.2.2. Exposure to Chronically Infected Pigs

Viral RNA was detected in the nasal swabs of two of the four calves during the 48 h that they were housed in contact with chronically infected pigs. The virus was detected in the oral swab of another calf on day two and the nasal swab of this same calf on day eight. This calf was the only one of the four to develop antibodies against Bungowannah virus. Neutralising antibodies were first detected 16 days post-exposure. The antibody titre in this calf peaked at 2048 between 63 and 84 days post-inoculation. Bungowannah viral RNA was not detected in any other samples collected. Calves remained healthy throughout the study.

#### *3.3. Infection of Pregnant Cows*

Each of the seven cows inoculated with Bungowannah virus were found to shed Bungowannah virus RNA in nasal secretions. Shedding began in most cows at day three, with peak virus shedding between days seven and 10. One cow was still shedding viral RNA in nasal secretions on day 14 (Table 5). Bungowannah virus RNA was not detected in conjunctival swabs but was detected in the serum of four of the seven cows for a period of one to three days. No signs of respiratory disease or pyrexia were observed in the cows.


**Table 5.** qRT-PCR results for pregnant cows infected with Bungowannah virus.

A dash indicates that Bungowannah virus RNA was not detected.

All cows developed neutralising antibodies against Bungowannah virus (Table 6). One cow had developed antibodies by day 11, and the remainder by day 14. Titres peaked on day 44 post infection. Neutralising antibody was detected for the duration of the pregnancy for all cows. A biphasic antibody titre developed in Cow 3 with a peak of 2048 on days 37 and 44, a decline to 512 on days 66 and 100, and then a rise to 1024 on day 128 and 2048 on day 163.


**Table 6.** Bungowannah VN titres from pregnant cows intranasally inoculated.

A dash indicates that neutralising anti-Bungowannah virus antibodies were not detected.

Pregnancies in the cows were unremarkable with no abnormalities were detected by ultrasound. At birth and before suckling three of the seven calves were found to have antibodies against Bungowannah virus. Bungowannah virus RNA was also detected in ear skin biopsy samples from these same three calves, with Ct values of 26.0, 32.7 and 33.6. The calf with the highest RNA concentration (lowest Ct value) was the progeny of Cow 3. Bungowannah virus RNA was also detected in the serum of this calf. Viral RNA was not detected in conjunctival, oral, nasal or rectal swabs. All calves appeared normal at birth and displayed normal behaviour. The virus was detected in the vaginal swabs from dams of the three positive calves, but not from the other cows. Cow 3 had a Ct value of 30.4, while the other two were 35.8 and 36.6.

Post-mortem examination of all calves was unremarkable, with no gross abnormalities detected. The three calves found to have Bungowannah virus RNA in the skin at birth were euthanased and

necropsied at four weeks of age. Viral RNA was again detected in the skin of these calves, and also detected in the testicle of the single male calf among these three (Table 7). Viral RNA was not detected in any other tissues (oesophagus, stomach, duodenum, jejunum, ileum, caecum, colon, liver, bile, mesenteric lymph node, tracheobronchial lymph node, prescapular lymph node, tonsil, thymus, spleen, trachea, lung, thyroid, adrenal gland, kidney, urine, epididymis, ovary, uterus, skeletal muscle, bone marrow, myocardium, cortex, cerebellum or medulla). Virus isolation was attempted on all qRT-PCR positive samples but was not successful.


**Table 7.** Summary of laboratory results for progeny of cows infected with Bungowannah virus.

A dash indicates that a result was undetected. Calf numbers correspond to cow numbers.

#### **4. Discussion**

In this study we have demonstrated that Bungowannah virus can infect a proportion of both cattle and sheep after intranasal or subcutaneous inoculation, although infection was less efficient by the intra-nasal route when compared to pigs [5]. Additionally, viral RNA was detected sporadically at low levels in the serum of a proportion of infected sheep and cattle. These results suggest that following intra-nasal or subcutaneous exposure, the infection is predominantly localized. Systemic infection, as determined by the detection of viraemia, only occurred in a proportion of infections.

No significant disease was observed in the infected sheep and cattle, despite all the intranasally infected cattle shedding Bungowannah virus RNA in nasal swabs for up to two weeks after infection. The qRT-PCR results indicated that only low levels of virus were shed. The pattern of detection of RNA was consistent with virus replication rather than residual inoculum as there was a short period after inoculation when no RNA was detected. Similarly, viral RNA was detected in the nasal swabs of five of the six sheep infected intranasally, yet no disease was observed. It is interesting to note that the development of the humoral immune response in the two subcutaneously inoculated sheep occurred over a similar period as the intranasally inoculated sheep. We speculate the infection was localised at the injection site in these two animals and, in the absence of a viraemia, shedding via the upper respiratory tract did not occur. The low level of viral RNA detection in nasal secretions of sheep and cattle would suggest that ruminants are unlikely to play a significant role in transmission of the virus.

It was also possible to infect a calf through co-habitation with a chronically infected pig for 48 h. Interestingly, this calf was the most curious of the four calves in the trial and interacted the most with the pigs. It is possible that if given longer exposure to the infected pigs the other calves may have become infected.

Although all cows became infected, transplacental infection of the calf only occurred in less than half of the cases. Viraemia was associated with in utero infection in two of the four cows in which it was detected. It is noteworthy that Cow 3 developed a rising anti-Bungowannah virus antibody titre during the last trimester of pregnancy. The progeny of this cow was the calf with the highest levels of Bungowannah RNA at birth. We speculate that the dam's immune system may have been stimulated by transplacental transfer of Bungowannah virus antigen during the last trimester of pregnancy.

Each of the three calves that became infected had seroconverted by the time of birth. Persistent infections were not established, despite the cow being infected at a stage of gestation where immunotolerance and persistent infections would be expected, as usually occurs with BVDV infection. These three calves had significant levels of Bungowannah virus RNA present in skin samples at birth, over 210 days after their dams were infected. We hypothesise that these calves did develop a generalised infection in utero, but subsequently cleared the infection some time prior to birth. The skin and testes were the only sites where the viral RNA was not cleared. The testes are an immunoprivileged site, and so it may be that the infection continued at this site for a longer period of time, perhaps much closer to birth. It has also been suggested that skin is also immunoprivileged [14]. It has been shown that a live BVDV vaccine is able to cross the bovine placenta, cause a prolonged transient infection in the foetus and result in detection of viral RNA in the skin for many weeks after birth [15]. It appears a similar mechanism of infection has occurred in these calves with prolonged detection in the skin. The detection of viral RNA in the post-partum vaginal swabs of the three dams provides further evidence for a prolonged infection in these calves and is comparable to what was observed following infection in pregnant pigs [9].

The timing of infection of pregnant cows was designed to maximise the likelihood of producing persistently infected calves. Infection of cattle with BVDV during the first 90 days of gestation will generally produce calves that are immunotolerant to BVDV [16,17]. The results of this trial indicate that Bungowannah virus and BVDV infections in cattle behave differently, with Bungowannah failing to establish immunotolerance and persistence in the bovine foetus.

Understanding why Bungowannah virus resulted in a humoral response in these calves may lead to a deeper understanding of the functional differences between specific proteins in BVDV and Bungowannah virus and how they do or do not affect evasion of the host innate immune response in the bovine or porcine foetus [18].

In conclusion, while cattle and sheep can be infected with Bungowannah virus, they appear to be less susceptible to infection compared to pigs when challenged with a similar dose of virus. When compared to the infection of pigs [9], ruminant species appear to shed less virus and for a shorter period of time. As a result, transmission of Bungowannah virus is likely to be inefficient in ruminants and, without further host adaptation, it will probably not be able to be sustained in these species.

**Author Contributions:** Conceptualization, A.J.R., D.S.F. and P.D.K.; methodology, A.J.R., D.S.F. and P.D.K.; investigation, A.J.R.; data curation, A.J.R.; writing—original draft preparation, A.J.R.; writing—review and editing, A.J.R., D.S.F. and P.D.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** We wish to acknowledge Bob Rheinberger for his generosity and expertise while conducting the ultrasound examinations of the cattle fetuses. We are indebted to the staff of the Virology Laboratory at EMAI (in particular Katherine King and Melinda Frost) for their invaluable assistance during the testing of the samples described in this study, and to Xingnian Gu for his veterinary assistance during sample collection.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Brief Report* **Single-Round Infectious Particle Production by DNA-Launched Infectious Clones of Bungowannah Pestivirus**

#### **Anja Dalmann 1, Kerstin Wernike 1, Eric J. Snijder 2, Nadia Oreshkova 2, Ilona Reimann <sup>1</sup> and Martin Beer 1,\***


Received: 15 July 2020; Accepted: 31 July 2020; Published: 4 August 2020

**Abstract:** Reverse genetics systems are powerful tools for functional studies of viral genes or for vaccine development. Here, we established DNA-launched reverse genetics for the pestivirus Bungowannah virus (BuPV), where cDNA flanked by a hammerhead ribozyme sequence at the 5 end and the hepatitis delta ribozyme at the 3 end was placed under the control of the CMV RNA polymerase II promoter. Infectious recombinant BuPV could be rescued from pBuPV-DNA-transfected SK-6 cells and it had very similar growth characteristics to BuPV generated by conventional RNA-based reverse genetics and wild type BuPV. Subsequently, DNA-based ERNS deleted BuPV split genomes (pBuPVΔERNS/ERNS)—co-expressing the ERNS protein from a separate synthetic CAG promoter—were constructed and characterized in vitro. Overall, DNA-launched BuPV genomes enable a rapid and cost-effective generation of recombinant BuPV and virus mutants, however, the protein expression efficiency of the DNA-launched systems after transfection is very low and needs further optimization in the future to allow the use e.g., as vaccine platform.

**Keywords:** Bungowannah virus; flavivirus; reverse genetics; single round infectious particle

#### **1. Introduction**

Bungowannah virus (BuPV) is an atypical pestivirus (species *Pestivirus F*) within the genus *Pestivirus* of the *Flavivirdae* family [1]. The virus was isolated for the first time in 2003 from a large Australian integrated pig farm during an outbreak of sudden death in young pigs, followed by an increase in stillbirth [2,3]. Although BuPV represents a potential threat to commercial pig farming, it has not yet been reported from any other region or country [4,5].

BuPV has a positive sense RNA genome that is approximately 12.6 kb in length. A single open reading frame, flanked by 5 and 3 non-translated regions, encodes a polyprotein, which is co- and post-translationally processed into structural proteins (C, ERNS, E1, E2) and non-structural proteins (NPRO, p7, NS2/NS3 (NS2, NS3), NS4A, NS4B, NS5A, NS5B) [2]. The envelope protein ERNS as well as the non-structural protein NPRO are unique to pestiviruses.

Genomic and antigenic properties of BuPV, as well as its broad in vitro host cell tropism, indicate remarkable distance to previously described pestiviruses [6–9]. In general, pestiviruses infect host-specific cells of ruminant, porcine, or sheep origin. Classical swine fever virus (CSFV) can only infect porcine cells efficiently, while bovine viral diarrhea virus (BVDV) and border disease virus (BDV) have the potential to infect broader host spectra [10–14]. However, only BuPV could infect cell lines of African green monkey, bat, human, and mouse origin [9].

To study the special characteristics of BuPV in detail, a robust reverse genetics system (RGS) is essential. For other pestiviruses, such as CSFV or BVDV, many RGSs based on cDNA copies of the viral genome cloned into plasmid or bacterial artificial chromosome (BAC) vectors have already been reported [15–20]. Furthermore, recombinant pestiviruses were generated by full-length genome RT-PCR-based amplification and direct RNA generation from the amplicons without cloning steps [21]. All these techniques have in common the use of a bacteriophage T7 or SP6 RNA polymerase promoter for in vitro transcription to synthesize infectious positive strand RNA, which is subsequently transfected into cells to produce infectious virus progeny.

Here, we report a first dual promoter DNA-launched BuPV RGS, which is based on a cDNA plasmid (pBuPV), with a cytomegalovirus (CMV) immediate-early promoter as well as the bacteriophage RNA polymerase T7 (T7) promoter upstream of the BuPV genome in a mammalian expression vector. For the generation of correct 5 and 3 ends, self-cleaving ribozyme sequences were inserted. This construct enables the transcription of the BuPV DNA by the CMV promoter in the nucleus and by the T7 promoter in the cytoplasm of polymerase expressing BSR cells (BSR-T7/5), where the latter serves as proof of principle to demonstrate that the modifications of the genome do not affect the transcription. This plasmid allows the rescue of infectious virus without in vitro RNA synthesis, and the virus can be passaged efficiently. We also established a split genome construct (pBuPVΔERNS/ERNS) with a large deletion in the *ERNS* gene, preventing the efficient generation of virus progeny, and a synthetic CAG promoter followed by the genomic region encoding the BuPV-ERNS protein downstream the T7 termination signal (T7term) for the expression of BuPV-ERNS and the production of single-round infectious particles (SRIPs) via *trans*-complementation. The particles generated in this way should be able to pass through an additional replication cycle to infect surrounding cells, but are not expected to be capable of further propagation.

#### **2. Materials and Methods**

#### *2.1. Cells and Viruses*

SK-6 cells (RIE262, Collection of Cell Lines in Veterinary Medicine (CCLV), Friedrich-Loeffler-Institut, Insel Riems, Germany) and BSR-T7/5 cells, constitutively expressing the T7 polymerase (RIE583, CCLV) [22], were grown in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal calf serum (FCS) at 37 ◦C and 5% CO2. rBuPVRNA was generated after RNA transfection of SK-6 cells with in vitro transcribed RNA from the previously described synthetic cDNA clone pA/BV [9]. The virus was propagated in SK-6 cells and virus stocks were generated after three cell culture passages.

#### *2.2. Plasmid Construction*

All plasmids were prepared by standard molecular biological methods and plasmid DNA was purified by the Qiagen Plasmid Midi kit (Qiagen, Hilden, Germany). The identity of the constructs was confirmed by Sanger sequencing using the Big Dye® Terminator v1.1 Cycle sequencing kit (Applied Biosystems, Foster City, CA, USA) and appropriate primers. Nucleotide sequences were read with an automatic sequencer (3130 Genetic Analyzer; Applied Biosystems, Foster City, CA, USA) and analyzed using Geneious software (version 10.2.3.). Primers used for cloning procedures are shown in Table 1.


**Table 1.** Primer sequences for plasmid construction.

The infectious clone pBuPV, the basis for DNA-launched BuPV, was constructed by multiple cloning steps. In a first step, donor splice sites, that were detected in the BuPV genome with a confidence >0.65 using the NetGene2 Server (http://www.cbs.dtu.dk/services/NetGene2 [23,24]; Table S1), were mutated in the full-length clone "pA/BV." For this, fusion PCR was applied, which is a restriction-free cloning method [7], using the Phusion® High-Fidelity PCR kit (New England Biolabs, Ipswich, MA, USA) and the primers Ph\_Donor\_C-Erns\_F, Ph\_Donor\_Erns\_R, Ph\_Donor\_Erns\_F, Bungo\_2534R, Bungo\_Npro\_Mut\_Donor\_F, Bungo\_1079R, Bungo\_Donor\_IV\_F, Bungo\_Donor\_IV\_R\_new, Bungo\_Donor\_V\_F, Bungo\_6186\_R, Bungo\_Donor\_VI\_F, Bungo\_Donor\_VI\_R\_new, Bungo\_Donor\_VII\_F, Bungo\_Donor\_VII\_R, Bungo\_Donor\_VIII\_F, Bungo\_Donor\_VIII\_R, Bungo\_Donor\_IX\_F, and Bungo\_Donor\_IX\_R (biomers.net GmbH, Ulm Germany), resulting in plasmid "pBV\_opt."

For pBuPV generation, the well characterized plasmid pHaHd was used, which contains in addition to the CMV and T7 promoters, ribozyme sequences for the generation of correct genome sequences. In order to insert the optimized BuPV-specific cDNA into plasmid pHaHd [25] by linear-to-linear homologous recombination (LLHR) [26], plasmid pHaHd was linearized with *BglII* and used as template for PCR amplification of a linear vector fragment with the Phusion® High-Fidelity PCR kit and primers pHaHd\_F and pHaHd\_R. A full-length BuPV-specific PCR fragment was amplified by using plasmid DNA pBV\_opt as template and primers Bungo\_LLHR\_F and Bungo\_LLHR\_R. To allow homologous recombination of the virus- and vector-specific fragments, 50 nucleotide-long vector specific homology arms were included in the primer sequences. Furthermore, for correct cleavage by the synthetic hammerhead ribozyme (HHr), five nucleotides complementary to the 5 end of the BuPV-genome had to be inserted into primer Bungo\_LLHR\_F (Table 1). Thereafter, the PCR fragments were digested with *DpnI* to remove residual template DNA and gel purified (QIAquick Gel Extraction kit; Qiagen, Hilden, Germany). Subsequently, both PCR fragments were subjected to LLHR [26]. Recombination was performed by electroporation of both DNA fragments in the *E. coli* strain GB05-dir (Gene Bridges, Heidelberg, Germany). In brief, fresh overnight cultures in lysogeny broth (LB) medium were incubated at 37 ◦C for 1.5 h and RecE/RecT recombination was induced by L-Arabinose. After an additional incubation period at 37 ◦C for 30 min, the cells were washed two times with ice-cold water. A total of 100 ng of the amplified DNA fragments were added to the pelleted bacteria and electroporation was done at 1350 V, 50 μF, and 600 Ω by using the Gene pulser Xcell Electroporation System (Bio-Rad, Hercules, CA, USA).

Plasmid pBuPVΔERNS with a deletion of 448 bases within the ERNS protein (aa 328–483) was generated by fusion PCR using pBuPV as DNA template and primer pair Bungo\_dERNS\_F and Bungo\_2164R. For construction of the split genome plasmid pBuPVΔERNS/ERNS, plasmid pCAGGS\_BuPV-ERNS [27] was digested with *SmaI* and *NotI* and the ERNS comprising fragment was ligated into plasmid pBuPVΔERNS, digested with *PmeI* and *NotI*. Further details of the plasmid constructions are available on request.

#### *2.3. cDNA Stability*

The plasmid pBuPV was propagated for 10 passages in *E. coli*. Subsequently, the DNA was purified using the Qiagen Plasmid Mini kit (Qiagen, Hilden, Germany) and analyzed by *HindIII* digestion. DNA preparations of passages 5 and 10 were also used to transfect SK-6 or BSR-T7/5 cells and investigate for virus rescue.

#### *2.4. Transfection and Virus Rescue*

DNA transfections of plasmids pBuPV, pBuPVΔERNS and pBuPVΔERNS/ERNS (2 μg DNA each) into SK-6 (plasmid pBuPV) or BSR-T7/5 (all plasmids) cells were performed by using Lipofectamine™ 2000 Transfection Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The transfected cells were seeded and incubated for three days at 37 ◦C and 5% CO2. For recovery of infectious viruses, supernatants of the transfected cells were harvested three days post transfection

(p.t.) and passaged on SK-6 cells. At the day of collection, replication of BuPV was monitored by immunofluorescence (IF) staining using monoclonal antibodies. Virus stocks of pBuPV were prepared after four cell culture passages. The identity of the recombinant viruses was confirmed by RT-PCR and sequence analysis using appropriate primers. For RT-PCR, total RNA of virus-infected cells was extracted using the QIAamp Viral RNA Mini kit (Qiagen, Hilden, Germany) according to manufacturer's instructions, and the cDNA was amplified using the OneStep RT-PCR kit (Qiagen, Hilden, Germany).

#### *2.5. Immunofluorescence Assay*

Transfected or infected cells were fixed and permeabilized with 80% acetone on ice for 15 min. After 30 to 45 min incubation with monoclonal antibodies specific for BuPV-ERNS (682/43C3, diluted 1:20; M. Dauber, Friedrich-Loeffler-Institut, Insel Riems, Germany), for BuPV-E2 (682/45F12, diluted 1:20; M. Dauber, Friedrich-Loeffler-Institut, Insel Riems, Germany) or the pan-*Pestivirus* NS3 antibody WB112 (diluted 1:500; CVL, Weybridge, UK), the cells were washed twice with phosphate buffered saline (PBS). Thereafter, cell cultures were incubated with a goat anti-mouse Ig Alexa-488 conjugate (1:1000; Thermo Fischer scientific Inc., Waltham, MA, USA) for 30 min and analyzed by using a fluorescence microscope (Nikon Eclipse; Nikon GmbH, Düsseldorf, Germany).

#### *2.6. Virus Titration and Growth Kinetics*

SK-6 cells were infected with recombinant BuPV recovered from pBuPV (rBuPVDNA) and recombinant BuPV recovered from pA/BV (rBuPVRNA) at a multiplicity of infection (M.O.I.) of 1. Supernatants were collected at 0, 8, 24, 48, and 72 h post infection (p.i.), and virus titers were calculated as a 50% tissue culture infective dose per ml (TCID50/mL) after IF staining.

#### **3. Results and Discussion**

Reverse genetics systems are important tools that enable the investigation of viral genes, viral replication cycles, or pathogenesis, and allow for the development of safe and efficacious vaccines. Infectious virus production from DNA plasmid transfections into mammalian cells using RNA polymerase I (pol I) or RNA pol II systems had been described for several RNA viruses [28–32]. These systems allow a simple and stable virus rescue, faster and less costly than conventional RNA-based reverse genetics. RGS using a pol I-promoter [33] or a pol II-promoter [34] were also described previously for the pestivirus CSFV. Both systems allowed the rescue of infectious CSFV with high virus titers. In our study, we established a dual promoter BuPV infectious clone "pBuPV" with both the CMV pol II promoter and the bacteriophage T7 promoter, which allows virus generation via DNA through the nucleus (CMV pol II promoter) or cytoplasmatic T7-based generation. The T7-based generation served as a proof of principle to demonstrate that the modifications of the genome do not affect the transcription.

In order to prevent the viral RNA from being spliced in the nucleus, donor splice sites with confidence >0.65 detected in NPRO, C, ERNS, E2, p7, and the other non-structural proteins were mutated in pA/BV (Table S1). Subsequently, the optimized BuPV-specific cDNA was used for construction resulting in plasmid pBuPV. In this construct, downstream of the CMV and bacteriophage T7 promoters, the BuPV genome termini were flanked by sequences coding for a synthetic HHr and hepatitis delta virus ribozyme (HDVr) to generate precise 5 - and 3 -terminal sequences. Downstream of the HDVr sequence, a T7 termination (T7term) signal allows transcription termination (Figure 1A).

**Figure 1.** Plasmid pBuPV, genetic stability and rescue of recombinant rBuPVDNA. (**A**) Schematic representation of the RNA Polymerase II-based plasmid pBuPV encoding full-length BuPV cDNA. Indicated are cytomegalovirus immediate-early (CMV) RNA Polymerase II promoter (open arrow), bacteriophage RNA polymerase T7 (T7) promoter (shaded arrow), hammerhead ribozyme (HHr), 3 hepatitis delta virus ribozyme (HDVr), and T7 terminator sequence (T7term). (**B**) Stability of the full-length cDNA clone pBuPV. The primary plasmid (P0) was passaged 10 times in *E. coli* DH10B (P1–P10) and investigated by restriction analysis using *Hind*III. + indicates generation of infectious rBuPVDNA in rescue experiments. (**C**) Multi-step growth curves determined after infection of SK-6 cells with rBuPVDNA (rescued by DNA transfection of pBuPV) or rBuPVRNA (rescued by transfection of in vitro transcribed RNA of the infectious cDNA clone pA/BV) at an M.O.I. of 1 showed similar growth characteristics for both viruses.

The stability of plasmid pBuPV in bacteria was investigated by 10 serial cloning and passaging cycles in *E. coli* DH10B cells. Restriction enzyme analysis using *Hin*dIII provided some indication about the genetic stability of the construct, and the same restriction pattern for P0 (primary construct) and passages P1 to P10 were observed (Figure 1B). In transfection experiments, the DNA-launched recombinant rBuPVDNA was analyzed for RNA replication, expression of BuPV proteins, and virus growth in both SK-6 and BSR-T7/5 cells. At 72 h p.t., CMV-driven expression of the BuPV proteins NS3 (anti-NS3), E2 (anti-E2), and ERNS (anti-ERNS) was detected by IF staining of the pBuPV-transfected SK-6 cells.

Bacteriophage T7 RNA polymerase (T7-RNA-Polymerase)-driven cytoplasmatic expression of all proteins could be observed as well after transfection in BSR-T7/5 cells; DNA-transfection of SK-6 cells and transfection in BSR-T7/5 cells resulted in single NS3, E2, and ERNS expressing cells (Figure 2). Cell culture supernatants collected from both transfected cell lines were inoculated into fresh SK-6 cells and the presence of recombinant BuPV (rBuPVDNA) particles in supernatants of both cell lines could be confirmed by IF analysis. The rescued rBuPVDNA could be efficiently passaged in SK-6 cells (Figure 2). Virus rescue was possible, regardless of the bacterial passage number of pBuPV in *E. coli*,

indicating once more the general stability of the plasmid (data not shown). Virus stocks produced from pBuPV-transfected SK-6 cells after four passages in SK-6 cells with a titer of 106.25 TCID50/mL were used for growth kinetics analyses in comparison to rBuPV generated by the previously established T7-RNA-Polymerase based RGS [9]. Analysis of the multi-step growth curves in SK-6 cells revealed similar growth characteristics for both viruses, and final virus titers of 106.6 TCID50/mL (rBuPVDNA) and 106.5 TCID50/mL (rBuPVRNA) could be determined at 72 h p.i. (Figure 1C).

**Figure 2.** Rescue of rBuPVDNA in BSR-T7/5 and SK-6 cells. Cells were transfected with plasmid pBuPV. At 72 h p.t, IF staining with pan-pesti NS3-specific mab WB112 (anti-NS3), and E2-specific and ERNS-specific mabs verified expression of NS3, E2 and ERNS in transfected cells. At this time, recombinant virus in the supernatants was transferred to SK-6 cells (1st passage), and later on transferred for a 2nd passage. E2, ERNS and NS3-positive cells indicated the generation of infectious progeny virus at 72 h p.i. in both transfected SK-6 and BSR-T7/5 cells. Scale bars indicate 100 μm.

In a next step, we were interested in the production of BuPV single-round infectious particles (SRIPs), since RNA-based SRIPs generating systems have already been described for several other flaviviruses [35–38]. SRIP production relies on the transfection of in vitro transcribed replicon RNA with deletions within the genomic region encoding for one of the structural proteins C, ERNS, E1, or E2 in cells stably expressing either the protein missing in the replicon or all structural proteins [39–43]. The packaged replicon particles are infectious, but progeny virus cannot spread from the infected cells, because the packaged replicon genome lacks the respective structural protein genes. In experimental animal studies, packaged replicon particles were proven to be appropriate for the development of non-transmissible, life attenuated pestivirus marker vaccine candidates [44–46]. However, the production of the replicon particles is time-consuming and needs the establishment of *trans*-complementing cell lines, which in many cases do not allow further passaging of the packaged pestivirus replicon particles.

Here, DNA-based SRIPs were produced as packaged BuPV replicon particles. Other DNA-based flavivirus SRIPs generating systems are mostly based on co-transfection of two expressing plasmids directly in eukaryotic cells, a subgenomic replicon plasmid, which lacks the structural protein-coding region, and a structural protein-expressing plasmid [35–38]. In addition, split genomes with two CMV promoters in back-to-back orientation, directing either the transcription of a capsid-deleted replicon RNA or the transcription of capsid-encoding mRNA had been described [47]. Since this strategy was described for the flavivirus West Nile virus, but had not been applied to pestiviruses up to now, we first constructed pBuPVΔERNS, which is a DNA-based replicon plasmid with a deletion of a large portion (codons 328–483) of the genome region encoding ERNS (Figure 3A, upper panel). Subsequently, this plasmid was used for the establishment of the split genome construct pBuPVΔERNS/ERNS (Figure 3A, lower panel). In transfection experiments using pBuPVΔERNS and BSR-T7/5 cells, transient expression of NS3 could be detected at 24 h p.t. by IF staining, while expression of BuPV-ERNS could not be observed (Figure 3C, panels c–d). No infectious recombinant BuPV could be recovered, even after serial passages in SK-6 cells (Figure 3C, panels k–l and s–t).

By insertion of a synthetic CAG promoter and the genomic region encoding the N-terminal signal sequence and the BuPV-ERNS protein in plasmid pBuPVΔERNS downstream the T7term signal, the split genome plasmid pBuPVΔERNS/ERNS was generated (Figure 3A, lower panel). This construct is capable of transcribing two separate RNA species from two different promoters.

The CMV promoter directs the transcription of BuPV replicon RNA BuPVΔERNS, which expresses all non-structural protein genes and the structural protein genes *C*, *E1*, *E2*, and a truncated *ERNS* gene (ΔERNS), whereas a synthetic CAG promoter [48] downstream of the BuPV replicon genome directs transcription of mRNA encoding full-length ERNS for complementation. Together, the two promoters allow the expression of the complete BuPV genome including all structural proteins. The ERNS-deleted replicon genome is amplified by the BuPV non-structural proteins NS3, NS4A, NS4B, NS5A, and NS5B and can be packaged by the structural proteins C, ERNS, E1, and E2, essential for virus assembly, to generate SRIPs (Figure 3B). Secreted SRIPs are able to infect surrounding cells, where the replicon RNA can be replicated. As the structural proteins C, E1, and E2, but not BuPV-ERNS can be expressed from this RNA by the non-structural proteins, the RNA cannot be packaged again into new particles and no further spread of the SRIPs is possible (Figure 3B) resulting in a self-restricted system.

To examine the ability of the split genome plasmid pBuPVΔERNS/ERNS to produce SRIPs, BSR-T7/5 cells were transfected with pBuPVΔERNS/ERNS and compared with cells transfected with the replicon plasmid pBuPVΔERNS and the full-length plasmid pBuPV. Autonomous replication of the newly synthesized BuPV RNA was shown by the expression of BuPV-NS3 in cells transfected with plasmids pBuPVΔERNS and full-length pBuPV. However, unexpectedly no NS3 expression was observed for the SRIPs, which might be due to the low transfection efficiency, especially since also only single positive cells could be shown for the other constructs. Expression of ERNS could only be detected in cells transfected with pBuPVΔERNS/ERNS and the full-length pBuPV, but not in pBuPVΔERNS DNA transfected SK-6 cells (Figure 3C, panels a, c, and e). The observation that the SRIPs showed an increased ERNS expression (Figure 3C, panel a) could be related to the fact that the inserted sequence was optimized for the expression system applied in this study.

**Figure 3.** Schematic representation of the plasmids pBuPVΔERNS and pBuPVΔERNS/ERNS and production of rBuPV and BuPV-SRIPs. (**A**) The replicon construct pBuPVΔERNS was generated on the basis of the CMV immediate early promoter containing plasmid pBuPV by partial deletion of the ERNS encoding genomic region (aa 328–483); the split genome plasmid pBuPVΔERNS/ERNS contains two eukaryotic promoters. The CMV promoter (open arrow) controls transcription of BuPV replicon RNA, BuPVΔERNS, which expresses the non-structural protein genes and the structural protein genes C, E1, and E2. The CAG promoter (shaded arrow) downstream T7term directs the expression of BuPV-ERNS. Indicated is also the T7 promoter (black arrow), and the hammerhead ribozyme (HHr), and the 3 hepatitis delta virus ribozyme (HDVr), which are important for the generation of the correct termini of the transcribed replicon RNAs. (**B**) Generation and operation mode of BuPV single round infectious particles (SRIPs) [47]. When cDNA of pBuPVΔERNS/ERNS is transfected into susceptible cells, RNA-transcription starts in the nucleus under the control of the CMV promoter. The structural proteins C, E1, E2, and the non-structural proteins (NSP) are then expressed in the cytoplasm, while ERNS is expressed by the CAG promoter. The self-replicating, truncated RNAs can be packaged in SRIPs by the four essential structural proteins. The secreted SRIPs are able to infect new cells. The released RNA replicates autonomously in the cytoplasm and allows the expression of the structural proteins C, E1, and E2 but not of ERNS. Therefore, no further SRIPs can be produced and spread again (self-restriction). (**C**) IF analysis of BSR-T7/5 cells transfected with pBuPVΔERNS, pBuPVΔERNS/ERNS or pBuPV (a–f) or SK-6 cells infected with supernatants of DNA-transfected cells (1st passage, i–n) or infected with supernatants collected from cells after the first infection (2nd passage, q–v). IF staining using anti-NS3 or anti-ERNS monoclonal antibodies was performed at 72 h p.t. and 72 h p.i., respectively. Non-transfected or uninfected cells were used as control (g–h, o–p, and w–x). Scale bars indicate 100 μm.

When in vitro-transcribed RNA produced from an infectious cDNA clone of BuPV was transfected, the deletion of ERNS still allowed the generation of infectious particles [27]. However, in this study there were no indications of cell-to-cell spread in pBuPVΔERNS DNA-transfected cells or production of SRIPs in transfection supernatants (Figure 3C, panels c, d, k, and l), while transfection with pBuPV resulted in single NS3 and ERNS positive cells. In cells transfected with pBuPVΔERNS/ERNS, amplified BuPVΔERNS replicon RNA was packaged into BuPV-SRIPs (Figure 3C, panel a). The infectivity of BuPV-SRIPs and rBuPV was also demonstrated after inoculation of the supernatants of the pBuPVΔERNS/ERNS DNA-transfected cells to fresh SK-6 cells (1st passage) as indicated by the IF-detection of NS3 at 72 h p.i. Here, replicon RNA amplified itself since no additional SRIPs were produced; only single cells infected with BuPV-SRIPs were observed by IF staining using the anti-NS3 mab and no positive signals could be detected by ERNS staining (Figure 3C, panels i and j).

In contrast, after one passage, the cells infected with rBuPV produced large foci of BuPV-NS3 and -ERNS protein-positive cells resulting from virus replication and spread (Figure 3C, panel m and n). The supernatants from all investigated clones obtained after infection were again transferred to fresh cells (2nd passage), but infectious progeny virus was only detected in rBuPV infected cells (Figure 3C, panels u, v).

While the newly established DNA-based system is well suited to produce infectious viruses, it unfortunately shows only a very low transfection efficiency, as demonstrated by the IF staining of relatively few positive cells at 72 h after transfection (Figures 2 and 3). A possible explanation for the reduced efficiency might be the still insufficient removal of donor splice sites. As BuPV does not naturally replicate in the cell nucleus, we have modified several donor splice sites, although additional splice sites may have an impact on efficiency [49,50]. In addition, further optimization attempts such as the modification of the CMV promoter or the insertion of a polyA tail were made following other flavivirus systems, but they did not lead to an increase in efficiency after transfection for the newly established BuPV system (data not shown).

However, also SRIP systems for the flavivirus West Nile virus showed reduced efficiency after transfection, but still were able to induce neutralizing antibodies and confer protection in immunized mice and horses [47]. Whether this is the case for the BuPV system as well, needs to be evaluated in vaccination experiments. Nevertheless, since a high transfection efficiency might be important for SRIP systems, further optimization of the BuPV system will be necessary and useful. Roby et al. could increase the SRIP production of their beta-galactosidase expression system by using an elongations factor EF1α promoter for the expression and optimized the codon of the capsid protein [51]. Thus, the replacement of the CMV promoter by another, more efficient promoter in SK-6 cells, e.g., the CAG promoter, might be an option to improve the BuPV system in the future. In addition, it could be explored whether other structures of the system, such as the ribozyme sequences, influence the efficiency as was shown previously [31,52,53].

#### **4. Conclusions**

In summary, we established a cDNA clone for the rescue of infectious BuPV using an RNA polymerase II-driven system. The full-length clone pBuPV enables further investigation of the atypical pestivirus BuPV by rapid and cost-effective generation of BuPV mutants. Even though the split-genome strategy has to be optimized due to its very low transfection efficiency in cell culture, it could allow the establishment of a platform for *trans*-complementation of viral proteins with a single plasmid. A more efficient version could be an attractive alternative to conventional complementation approaches and will therefore be a major focus for the future research in this field.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1999-4915/12/8/847/s1. Table S1: Mutated sequences for donor splicing site deletion.

**Author Contributions:** Conceptualization, E.J.S., N.O., I.R., and M.B.; methodology, A.D. and I.R.; formal analysis, I.R.; investigation, A.D., K.W., and I.R.; writing—original draft preparation, A.D. and I.R.; writing—review and editing, K.W., E.J.S., N.O., and M.B.; visualization, A.D. and K.W.; supervision, K.W., I.R., and M.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Zoonoses Anticipation and Preparedness Initiative (ZAPI, Grant Agreement No. 115760) within the Innovative Medicines Initiative (IMI Call 11—IMI-JU-11-2013-04).

**Acknowledgments:** We thank Doreen Schulz and Gabriela Adam for excellent technical assistance and Stefan Finke for kindly providing the plasmid pHaHd.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Viruses* Editorial Office E-mail: viruses@mdpi.com www.mdpi.com/journal/viruses

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18