Investigations into the presence of nidoviruses in pythons

Background Pneumonia and stomatitis represent severe and often fatal diseases in different captive snakes. Apart from bacterial infections, paramyxo-, adeno-, reo- and arenaviruses cause these diseases. In 2014, new viruses emerged as the cause of pneumonia in pythons. In a few publications, nidoviruses have been reported in association with pneumonia in ball pythons and a tiger python. The viruses were found using new sequencing methods from the organ tissue of dead animals. Methods Severe pneumonia and stomatitis resulted in a high mortality rate in a captive breeding collection of green tree pythons. Unbiased deep sequencing lead to the detection of nidoviral sequences. A developed RT-qPCR was used to confirm the metagenome results and to determine the importance of this virus. A total of 1554 different boid snakes, including animals suffering from respiratory diseases as well as healthy controls, were screened for nidoviruses. Furthermore, in addition to two full-length sequences, partial sequences were generated from different snake species. Results The assembled full-length snake nidovirus genomes share only an overall genome sequence identity of less than 66.9% to other published snake nidoviruses and new partial sequences vary between 99.89 and 79.4%. Highest viral loads were detected in lung samples. The snake nidovirus was not only present in diseased animals, but also in snakes showing no typical clinical signs. Conclusions Our findings further highlight the possible importance of snake nidoviruses in respiratory diseases and proof multiple circulating strains with varying disease potential. Nidovirus detection in clinical healthy individuals might represent testing during the incubation period or reconvalescence. Our investigations show new aspects of nidovirus infections in pythons. Nidoviruses should be included in routine diagnostic workup of diseased reptiles.


Background
Snakes have become an increasingly popular exotic pet in the last decades. Especially snakes of the family Boidae, which includes the so-called boas and pythons, are kept in high numbers in captivity. These snakes are non-venomous and most of them are easy to handle. Owners appreciate their fascinating colors, the considerable size of some the species and interesting physiology. Most of these snakes are bred in captivity but new species, new morphs and fresh bloodlines for breeding stocks are imported from the wild. This offers the opportunity for an exchange of different pathogens between animals from different ecologically and geographically regions. Therefore, more and more reptile diseases have been detected and for some of them pathogens were identified. Pneumonia and stomatitis are common problems in captive snakes and can be caused by different infectious agents. Mostly, they are linked to bacteria, but often a viral disease represents the basic infection. Viruses infecting reptiles and especially snakes causing respiratory signs include paramyxoviruses, reo-, adeno-and arenaviruses. These findings are mostly based on clinical, pathological and histological observations [1][2][3].
In 2014, three working groups described contemporaneous using unbiased deep sequencing a new nidovirus in captive pythons suffering from pneumonia. The first report from the USA investigated material from eight deceased ball pythons (Python regius) with a pneumonia.
By deep sequencing, they were able to obtain a complete coding genome of a novel nidovirus. As a control, organ materials of 57 snakes, mostly colubrid, with clinical signs other than respiratory diseases were tested [4]. The other study from the USA focused as well on ball pythons. There, 12 snakes suffering from pneumonia, tracheitis and esophagitis were examined and were subjected to a metagenomic workflow and partial sequences from a novel nidovirus were derived [5]. The third study described the results from a diagnostic workup of an Indian python (Python molurus) diagnosed with severe necrotizing pneumonia from a zoological garden in northern Germany. Again, unbiased sequencing resulted in the complete coding genome sequence of a novel nidovirus, which could not be detected in ten healthy boas [6]. After that, a working group from Switzerland detected a nidovirus in different captive breeding colonies of green tree pythons (Morelia viridis). Twelve deceased animals with pneumonia were investigated and this virus shared only < 85% sequence similarity to the previous published sequences [7]. Marschang and Kolesnik detected the virus not only in deceased animals, but also in swab samples from living animals. Out of the 201 examined animals, 30 were RNA positive. The animals originated from different parts of Europe and positive animals included Python regius, Python molurus, Morelia viridis, Morelia spilota and Boa constrictor. Pythons (27.4%) were more affected than boas (2.4%) [8]. In Italy the virus was detected in six tracheobronchial lavages from diseased ball pythons [9]. A working group in the USA fulfilled Koch's postulates in 2018, which showed a causal relationship between the novel snake nidovirus and pneumonia in ball pythons by experimentally infection of three animals. The observed clinical signs covered oral mucosal redding, mucous secretions, open-mouth breathing and anorexia. Histologic examinations showed rhinitis, stomatitis, tracheitis, esophageitis and interstitial pneumonia. The virus was present in oral secretions as well as in feces, suggesting either transmission by aerosolization or by the fecal-oral route [10]. Besides these reports from snakes, nidoviruses were also found in other reptiles like turtles and lizards in the last years. In lizards, they have been proven in animals with and without clinical signs. In turtles, the virus was found only once in diseased animals [11,12].
The order nidovirales compromises divergent virus families with notable human and animal pathogens like for example severe acute respiratory syndrome (SARS) coronavirus and porcine reproductive and respiratory syndrome virus (PRRSV) [13]. The snake nidoviruses were assigned to the family coronaviridae in the subfamily torovirinae that covers two genera: torovirus and bafinivirus infecting mammals and fish, respectively [14][15][16][17]. The reptile nidoviruses cluster together in a monophyletic clade and a new genus for these unassigned viruses was proposed: barnivirus (bacilliform reptile nidovirus) [4]. Last year the ICTV changed the taxonomy of the order nidovirales from four to seven suborders. The subfamily torovirinae, based on molecular properties, is now no longer part of the coronaviridae, but belongs instead to a suborder called tornidovirineae in a new family called tobaniviridae. Within this family, the subfamily Serpentovirinae compromises in the genus Roypretovirus the ball python nidovirus (ICTV, 2018).
In this study, we report the detection of divergent snake nidoviruses after metagenomics analysis and their distribution in different snake species, which is not always correlating with disease signs.

Sample collection
In 2014, several snakes in a collection of Green Tree pythons suffered from severe stomatitis and pneumonia (up to 25 animals during several weeks). Bacteriological investigations of oral swabs yielded different results in the diseased snakes. Mycological and virological investigations (paramyxovirus, arenavirus, reovirus, adenovirus) were negative. Direct microscopic evaluation of mucus showed no parasites; parasitological investigations of feces were also negative. The stomatitis was treated locally (cleansing, antibiotic ointments or fluids). For the pneumonia, a systemic treatment with antibiotics was carried out according to an antibiogram following bacteriological investigations of oral swabs. As a supportive treatment, infusions (Ringer's solution, 10 ml/kg) and ZylexisR were applied (twice in a week interval). MetacamR was used to calm down the inflammation. Despite this therapy, most of the infected snakes died or had to be euthanized (20 out of a collection of 60 snakes).
A total of 1554 captive boid snakes were screened for snake nidoviruses. The available samples included organ tissue samples (mostly lung, in some cases liver, kidney, small intestine, brain and pancreas) from deceased animals (230) as well as oral swabs or tracheal washes from living snakes (1324). Some animals were tested at several time points.
The oral swabs and tracheal washes were sent to the institute between 2015 and 2018 from different parts of Germany (all 16 federal states) and neighboring countries. This includes oral swabs from Denmark (78), 15 samples from a collection in Italy, 20 samples from five different collections in France and 117 samples from seven collections in Austria. Some of the animals showed typical symptoms (stomatitis and / or pneumonia), others without any symptoms were investigated to obtain an overview of the infection status of the collection.

Pathological examinations
Nine green tree snakes were sent for necropsy. Diagnostic workup in the chemical and veterinary investigation office included a gross pathology, histopathology, virology and bacteriology. Necropsy and following investigations have been carried out using standard techniques. In short, after pathological examination of the carcasses tissues were fixed in formalin and stained with Haematoxylin-Eosin. The tissues were evaluated microscopically. All snakes were investigated virologically. Suspensions of liver, lung, kidney and intestine were inoculated bacteria-free into Viper Heart Cells at 29°C (ATCC CCL 140) and incubated for 1 week. Supernatant was transferred to a new cell culture and incubated for another week. Cells were inspected daily for the presence of cytopathic effects. PCRs for paramyxovirus [18], for reovirus [19], for reptarenavirus [20], ranavirus [21] and for adenovirus [22] were performed with organ tissues and cell culture supernatant. Bacteriological investigations from liver, kidney and lung were carried out in five snakes onto blood agar, incubated at 30°C. Isolated bacteria were typed using MALDI TOF.

RNA extraction for NGS and RT-qPCR
Small pieces of tissue samples were homogenized in 1 mL PBS with a 5 mm steel bead in a TissueLyserII (Qiagen, Hilden, Germany). Dry swab samples taken from the trachea of the snakes were resuspended in 2 ml cell culture medium by shaking for half an hour at room temperature. For metagenomics analysis 250 μl of homogenized lung tissue was mixed with 750 μl Trizol Reagent (life technologies, Darmstadt, Germany). Afterwards chloroform was added and RNA from the aqueous phase was precipitated with 75% ethanol. RNA was further purified with RNeasy Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions and eluted in 50 μl RNase free water. For screening purposes nucleic acids from tissue or swab samples were extracted using the King Fisher 96 Flex (Thermo Fisher Scientific, Braunschweig, Germany) in combination with the NucleoMagVet kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions and eluted in 100 μl.

Next-generation sequencing
Sample material from three deceased snakes was analyzed by a metagenomics workflow as described elsewhere [23,24].

Conventional RT-PCR
In order to find specific primers for one published ball python nidovirus (KJ541759) [4] the primer designed tool Primer-BLAST was used [25]. Three different primer pairs were tested. The RT-PCR was done with the One step RT-PCR kit (Qiagen, Hilden, Germany) and uses a forward primer (5'CAA CTC TGC ACA AAC GCG AA 3′) and a reverse primer (5'CGG CGA TCT TGA TGT TGC TG 3′) amplifying a PCR product of approximately 300 bp. The temperature profile consists of a reverse transcription step at 50°C for 45 min, an activation step at 94°C for 15 min, followed by 40 cycles of 94°C for 30 s, 57°C for 45 s and 72°C for 45 s as well as a final extension step at 72°C for 5 min. The PCR products were visualized by electrophoreses on an agarose gel.

Snake nidovirus specific RT-qPCR
To confirm the results of the metagenomic workflow and to screen for further infected animals, a real-time RT-PCR for the detection of snake nidoviruses was developed. Primers and probes targeting the replicase open reading frame 1B (ORF1B) were selected based on an alignment of published sequence information (NCBI GenBank) together with the two newly generated complete coding sequences from this study. The PCR was performed with the AgPath-IDTM One-Step RT-PCR kit (Thermo Fisher Scientific, Braunschweig, Germany) and a snake nidovirus specific FAM-labelled primer-probe mix consisting of 800 nM Nido-Snake-20, 528-F (5′ ACA TCT CGA GAC SAT YAT CCA 3′), 800 nM Nido-Snake-20,616-R (5′ CTG TAC TWG AAC AGA AYT CGT G 3′) and 200 nM Nido-Snake-20,579-FAM-as (5′ FAM-TTC CCA MGC YTT GTT CTS GTC GAC-BHQ1 3′) was used for broad-range nidovirus detection producing a 89 bp long fragment. For the RT-qPCR reaction, 10 μl master mix and 2.5 μl RNA were combined in a total reaction volume of 12.5 μl. The PCR was carried out using a Bio-Rad CFX 96 Real-Time Detection System (Bio-Rad, Hercules, CA, USA) and the following temperature profile: 10 min at 45°C for reverse transcription, 10 min at 95°C for denaturation and 45 cycles of 15 s at 95°C, 30 s at 56°C and 30 s at 72°C. In addition, a HEX-labelled internal control assay was used to monitor for efficient nucleic acid extraction [26].

Snake retrovirus specific SYBR green PCR
The metagenomics workflow provided some reads of snake nidoviruses. To see how frequent these viruses occur, a SYBR green based screening RT-qPCR was developed. The RT-qPCR reaction was prepared using a SensiFAST SYBR No-Rox Kit (Bioline, London, United Kingdom) in a volume of 10 μl including 400 nM Snake_ RV_2624-F (5′ ACA GTG CCT GAC CCA TAC AC 3′), 400 nM Snake_RV_2716-R (5′ AAG ACC AAA ATG CAT CTT TCA GAT C 3′), 400 nM Snake_RV_ 2773-R (5′ TGT ATC TGG GTC AGT CCA TTC AA 3′) and 2 μl of extracted RNA. The reaction was performed for 10 min at 45°C for reverse transcription, 2 min at 95°C for activation of the polymerase, and 45 cycles of 5 s at 95°C and 20 s at 60°C followed by a melting curve analysis. The PCR was carried out using a Bio-Rad CFX 96 Real-Time Detection System (Bio-Rad, Hercules, CA, USA).

Dideoxy chain termination sequencing
For phylogenetic analysis, primers were designed to generate a partial sequence of the ORF1AB RNAdependet-RNA-polymerase gene by dideoxy chain-termination sequencing [27] from representative snake nidovirus RT-qPCR-positive samples. The amplification reaction was performed using the SuperScript III Onestep RT-PCR Kit (Thermo Fisher Scientific, Waltham, Massachusetts, USA) combined with the primers Nido-Snake-19,971-F (5′ ATC GGA GTC WCA AAA TTC CGA G 3′) and Nido-Snake-21,007-R (5′ CAC GTR TAG CAY TGC TGC TG 3′). For the amplification, 800 nM of each primer in 12.5 μl total reaction volume including 2.5 μl template RNA was used. The amplified

Phylogenetic analysis
The newly generated sequences of two full-length snake nidoviruses were aligned to each other as well as to further complete coding genome sequences from different members of the subfamily Tobaniviridae obtained from Gen-Bank using the method MAFFT [28] as implemented in the Geneious software (version 10.2.1). Furthermore, the 36 partial ORF1AB sequences were aligned to each other and to other full-length snake nidovirus genomes. Based on these alignments, maximum-likelihood trees (PhyML) were calculated using the HKY85 model [29] with 1000 bootstrap replicates by the Geneious software [30].

Pathological findings
Nine green tree pythons were sent for necropsy. Four snakes were investigated pathohistologically, bacteriologically and virologically in detail (Table 1). After the first results indicated a viral aetiology, five more were examined only pathohistologically and virologically. All snakes revealed pneumonia of varying degrees. Five of the snakes also showed a diphtheroid-necrotizing stomatitis, and four suffered from renal gout ( Table 1). The results of the bacteriological studies showed different germs. Paramyxo-, Reo-, Adeno-and Arenavirus were excluded by PCR and/or cell culture. Attempts to isolate viruses in cell culture were unsuccessful.

Metagenomic sequencing and comparative analysis
Organ material of one green tree pythons (Morelia viridis) was subjected to a metagenome analysis. Thereby, a few reads with the highest overall identity to a ball python snake nidovirus were detected. Besides, reads of a retrovirus could be classified. Afterwards two other snake samples with a higher viral load were chosen for whole genome sequencing. A primer walking approach could close the gaps of the snake nidovirus sequences and thus two complete coding genomes could be generated with a sequence length of 32.88 kb and 32.75 kb respectively. The overall genome organization with eight open reading frames is similar to the other published snake nidoviruses (see Additional file 1: Figure  S1). Due to their origin, they were named Morelia viridis Fig. 1 Phylogenetic tree of complete genome nucleotide sequences of the family Tobaniviridae, suborder Tornidoviridae. The two newly generated snake nidoviruses are highlighted in red snake nidovirus. These two sequences share a nucleotide identity of 99.7% to each other. The highest overall sequence similarity to published sequences appears to be 66.8 and 66.9%, respectively to a nidovirus sequence derived from a green tree python in Switzerland (MF351889 [7]). The overall sequence similarities to other published snake nidoviruses is between 64.8 and 65.5%. The ORF1B (part of the replicase gene) seems to be most conserved with sequence similarities on nucleotide level varying between 79.8 and 80.8% and roughly 85% on amino acid level. The most variable area is the ORF3 (glycoprotein) with nucleotide sequence similarities of 53.5-47.5% and amino acid similarities of 40.7-33.2% (Table 3). All reptile nidoviruses cluster together in the genera Serpentovirinae inside the family Tobaniviridae within the order Nidovirales (Fig. 1).

Molecular assays and tissue tropism
To confirm the results from the metagenomic analysis different RT-qPCRs and RT-PCRs for the snake nidoviruses as well as a RT-qPCR for the snake retrovirus were developed. The snake nidovirus RT-qPCRs were compared to other published RT-qPCRs [7] and thereby the results further validated (Additional file 1: Table S1). For this, 60 negative and 46 positive oral swabs were investigated using two different PCRs. Different tissue samples from the nine dissected green tree pythons were tested by RT-qPCR. The highest viral loads could be detected in lung and liver samples matching the histopathologic results whereas intestine and kidney samples were mostly negative or showed discontinuous results (Table 4). To confirm whether the identified snake nidovirus and retrovirus are related to the observed clinical signs, further samples of affected and apparently healthy or otherwise diseased animals were tested ( Table 5). The retrovirus could not be found in all affected animals, but instead in animals deceased from a lymphoma (data not shown), which did not exhibit typical symptoms like pneumonia and stomatitis. The snake nidovirus RT-qPCR was positive for most of the animals with pneumonia, but also for a few of asymptomatic snakes (Table 6). To exclude contamination of the RT-qPCR set up an attempt was made to obtain a partial sequence from the RT-qPCR positive samples. The thereby produced sequences (in total 36 sequences, see Additional file 1: Table S2) were not identical, but shared instead a similarity between 99.89 and 79.4% to each other. A phylogenetic analysis showed no specific clustering according to the stock, the animal species or the occurrence of disease in the snake (Fig. 2).

Screening for snake nidoviruses
Until the end of 2018, a total of 1554 boid snakes were tested by RT-qPCR for snake nidoviruses resulting in 439 positive animals ( Table 2 Table 2). Samples were evaluated according to their anamnesis. Anamnesis was classified in typical symptoms or pathology (stomatitis and/or pneumonia), other symptoms or pathology, no symptoms/routine investigations and the group of snakes for which no information was available. The three most common python species providing most samples were evaluated (Green Tree pythons, ball pythons, carpet pythons). From 377 snake nidovirus RNA positive animals 75 showed signs of a disease and 285 did not show any clinical disease (Table 6). In 913 negative animals, 131 snakes revealed clinical symptoms or pathology, whereas 703 were without any symptoms (Table 6). No correlation between symptoms and finding of nidovirus could be established.
Follow-up investigations of a few snakes show a snake nidovirus RNA positive oral swab over several months. In addition, other animals from the same breeding stock never scored positive.  We further validated the new developed real-time PCR by comparing it to the PCR used from Dervas et al. [7]. In total, 60 negative and 46 positive swabs were tested using both methods. Results of our PCR were in line with the PCR developed by Dervas et al. (Additional file 1: Table S1). All positive samples showed comparable results with both PCRs. In two samples, results were negative in the PCR according to Dervas et al., whereas low reactions was found in our PCR assays.

Discussion
Respiratory diseases are quite a common problem in many collections of boid snakes. Viral agents like paramyxoviruses, arenaviruses and others are able to produce respiratory symptoms. However, in many collections, respiratory disease with high morbidity and mortality was found which was not caused by one of the well-known viruses. In the last years, with the discovery of snake nidoviruses the knowledge about pneumonia in boid snakes improved. These viruses were detected after different pythons succumbed to disease after a few months [4][5][6]. In our case, the first nidovirus detection occurred in a breeding stock of green tree pythons in which several animals showed severe respiratory signs, purulent stomatitis, poor or non-existing appetite, and weight loss. Mortality rates were high despite supportive treatment and care. Unbiased deep sequencing showed reads of a nidovirus and from two deceased animals fulllength sequences could be assembled. These sequences are a little bit shorter than the other published fulllength sequences of snakes, but belong still to the longest RNA genomes. The sequence identity to the other published genomes is rather low (< 66.9% on nucleotide sequence) with the highest similarity to the virus described in green tree pythons from Switzerland (Table 3) [7], whereas the three sequences published in 2014 are more similar to each other. Nevertheless, all reptile nidoviruses cluster together within the genus Pregotovirus (Fig. 1) [4]. Besides the snake nidovirus, the metagenomics analysis showed reads of a snake retrovirus. This retrovirus could be found in control animals showing no signs of respiratory disease and it is probably an already known endogenous retrovirus without a link to pneumonia [1,3,31]. The bacterial findings were not consistent and were probably a matter of secondary infections. No evidence for other pathogens could be found. With a newly developed RT-qPCR different tissues from nine deceased green tree pythons were tested to further investigate the tissue tropism. Thereby, a connection between the degree of histological changes and viral RNA detection was indicated (Tables 1 & 4). The highest viral loads were detected in the lung, whereas the other tested organs showed inconsistent viral RNA amounts. This indicates the respiratory tract as primary location of virus replication, makes the transmission by respiratory secretions possible and further strengthens the usefulness of oral or tracheal swabs as in-vivo sampling method [8]. We used the RT-qPCR for an initial screening for further snake nidovirus infected animals, including some animals deceased from other diseases or even apparently healthy (Table 5). To exclude unspecific amplification and laboratory contaminations, we generated partial sequences of the highly conserved ORF1B. Through this approach, 36 partial nidovirus sequences were obtained. Samples with very low viral loads did not result in a suitable sequence. Sequence comparison showed an identity between 99.89 and 79.4% indicating multiple virus strains. No direct relationship between collections, species or severity of disease is visible (Fig. 2). The host range of these viruses is not known and further virus strains not detectable by the used primer pairs could be possible.
Further screening of a total of 1554 animals resulted in 439 nidovirus RNA positive animals ( Table 2). From 377 (for which information about the disease status was available) nidovirus RNA positive animals 285 showed no respiratory disease (Table 6). In addition to the species described in previous reports, we could prove the infection in Python brongersmai, Bothrochilus albertisii, Brothrochilus boa, Morelia boeleni, Aspidites melanocephalus as well as Papua pythons (Apodora papuana, data not shown, 2019) further expanding the viral host range. Approximately 31%    of all tested pythons were positive. In spite of this, only one boa out of 128 animals revealed the presence of nidovirus genome. This is in concordance to one published study [8].
Unfortunately, no material for sequencing was available from the infected boa. This roughly confirms the 27% positive pythons [8]. At least in our study, the detection of viral RNA correlates not always with clinical signs. Interestingly, in pythons originating from the asian continent, the prevalence of nidovirus was much higher than in other pythons, for example ball pythons (Africa). A total of 41% of the investigated Green Tree pythons were positive for the virus; in Carpet pythons 24% and in Ball pythons 22% were positive, respectively.
Hoon-Hanks et al. fulfilled the Koch's postulates by experimental infection of ball pythons [10]. Therefore, the detection of nidovirus RNA in apparently healthy individuals may reflect testing during the incubation period or a previous nidovirus infection, because some animals stayed positive in oral swab samples over several months (data not shown). Whether it is infectious virus, or rather a form of RNA persistence is unclear. Other animals from infected collections never turned positive, suggesting a non-airborne transmission. Co-infections or non-pathogenic causes like e.g. stress through newly purchased animals may play a crucial role in the development of clinical disease. No specific treatment is available, infected snakes should be isolated and the testing for nidovirus included in standard diagnostic workup.

Conclusion
Our results show a nationwide distribution of nidoviruses in Germany with possible many existing strains.
In total 439 of 1554 tested snakes were positive for nidovirus but only a few of them revealed clinical signs like stomatitis or severe respiratory disease. Therefore, no obvious correlation between virus and clinical disease could be established. Some of the positive results may be due to testing during the incubation period or samples may have been taken during reconvalescence of a nidovirus infection. Results indicate that a nidovirus infection in pythons may cause no to severe disease possibly depending on the snake species, immune status of the snake, pathogenic potential of the virus strain or other unknown factors. Our investigations show new aspects of a nidovirus infection in pythons and contribute to the understanding of the biology of snake nidoviruses.