Poxviruses are large enveloped viruses containing a double-stranded DNA genome of 130-360 kb. They are important pathogens of high public health and economic impact and are able to infect a wide range of host species, ranging from insects to mammals. Within the family Poxviridae, two subfamilies (Chordopoxvirinae and Entomopoxvirinae) have been defined based on their hosts, which are either vertebrates or insects [1, 2]. The subfamily Chordopoxvirinae is divided into 18 genera, which include a total of 52 species [3].

Poxviruses infecting non-mammalian species belong to the genera Avipoxvirus and Crocodylidpoxvirus. As also observed in mammalian hosts, poxviruses of avian and crocodilian species are primarily associated with skin lesions but can also affect the upper respiratory and gastrointestinal tract [4, 5]. Poxvirus infections of reptiles other than crocodilians have been described in the literature, but no sequence information or further characterization of these viruses is available [6,7,8,9,10,11].

The species affected in this study, Crocodilurus amazonicus (“crocodile tegu”), is part of the family Teiidae. These lizards are native to the Amazon and Orinoco basins in South America [12] and belong to one of only two genera of living semi-aquatic teiids, and as such, they inhabit seasonally flooded forested areas near riverbanks or other watercourses. Their mean body temperature, 31.2°C, is higher than the environmental temperature but relatively low compared to that of other teiids, which could be due to their association with water habitats. Their variable diet consists of insects, other small reptiles, and frogs [13, 14].

In 2019, five C. amazonicus, owned by a private collector in Austria, were presented with skin lesions and weight loss. Due to the severity of the clinical signs, one animal had to be euthanized and was sent for further diagnostic evaluation to the pathologists of the Research Institute of Wildlife Ecology, University of Veterinary Medicine, Vienna, Austria. The animal was underweight and had multiple elevated, partially ulcerated skin lesions with a diameter of up to 4 mm on the back and the dorsal areas of head, neck, and tail that did not extend into the underlying muscle (Fig. 1A). Lesions were subsequently analyzed histologically, revealing multifocally and focally extensive epidermal hyperplasia of up to 10-15 times the normal thickness with a severely thickened stratum spinosum forming rete ridges. Keratinocytes showed ballooning degeneration and contained large eosinophilic, intracytoplasmic viral inclusion bodies (Bollinger bodies) up to 20 µm in size (Fig. 1B). Large numbers of poxvirus-like particles were detected within these lesions by transmission electron microscopy (Fig. 1C).

Fig. 1
figure 1

Pathologic and histologic examination. (A) Macroscopic presentation of skin lesions, represented by multiple, up to lentil-sized, elevated, partially ulcerated foci. This scale bar is in centimeters. (B) Hematoxylin-eosin-stained tissue section of an affected area of skin, displaying severe ballooning of the cells and prominent, eosinophilic intracytoplasmatic inclusion bodies. (C) Multiple poxvirus-like particles in skin lesions detected by transmission electron microscopy of uranyl-acetate-stained tissue sections.

A sample taken from a skin lesion tested positive for the presence of poxvirus DNA when a pan-poxvirus PCR for high-GC-content poxviruses developed by Li et al. was used [15]. Comparison of this 630-bp amplicon, which corresponded to part of the gene encoding the DNA-dependent RNA polymerase subunit rpo147 ortholog, to sequences available in public databases revealed the closest relationship, with 80% sequence identity, to members of the genus Avipoxvirus. This isolate was tentatively named "teiidaepox virus 1" (TePV-1). Cultivation in freshly isolated chicken embryo fibroblasts was attempted at different temperatures (22, 28 and 37°C), but this did not result in detectable virus replication, either by PCR testing of the supernatant or the development of a cytopathic effect. In order to characterize the virus further, we determined the sequence of the virus genome, employing a combination of Illumina sequencing technology (150-bp paired-end reads) and nanopore sequencing (MinION, Oxford Nanopore Technologies [ONT]) using DNA extracted from skin. Briefly, an aseptically dissected piece of skin containing lesions of approximately 10 mg was mechanically homogenized in 60 µl of PBS in a TissueLyser II at 30 Hz for 3 min (QIAGEN). For nanopore sequencing, 120 µl of lysis buffer was added, and the rest of the DNA preparation was performed following the manufacturer’s instructions for preparation of DNA from tissues using a QIAamp DNA Mini Kit. For Illumina sequencing, DNA was extracted using an NEB genomic DNA extraction kit according to the manufacturer’s instructions. The quality of the DNA preparation was checked using Genomic DNA ScreenTape (Agilent) on a 4200 TapeStation (Agilent) at the VetCore genomics facility. The library for Illumina sequencing was prepared using an NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) according to the manufacturer’s protocol and quality controlled using a fragment analyzer at the Vienna BioCenter Core Facilities before being sequenced using Illumina MiSeq chemistry. The raw reads were quality controlled, and adapter sequences were removed before commencing data analysis. For nanopore sequencing, the DNA was processed according to the protocol for sequencing of genomic DNA by ligation (SQK-LSK109) provided by ONT, and the library was loaded onto a Nanpore Flongle Flowcell (ONT).

Before contig assembly, the nanopore reads were aligned to the genome sequence of penguinpox virus using bowtie2-2.2.8 [16]. Nanopore reads aligning to this sequence were subsequently used for contig assembly of the Illumina reads with SPAdes 3.14.0 [17]. This straightforward approach, combining short reads with reads of up to 30 kb length, resulted in the ab initio assembly of two contigs with a length of 116,666 and 46,221 bp, respectively. These contigs had overlapping ends and could therefore easily be merged into one genome assembly with a final length of 166,425 bp and a GC content of 35.5%. The overlap between the two contigs was confirmed by site-specific PCR and Sanger sequencing. The coverage of the Illumina reads was 124 (138,058 aligned reads), and that of the nanopore sequencing reads was 125 (11,248 aligned reads out of 94,061; average length of aligned reads = 1848). The length distribution of the aligned reads and a histogram of the coverage distribution are shown Supplementary File 1.

Phylogenetic analysis of the genome sequence of TePV-1 and the amino acid sequences of the putative DNA polymerase (highest amino acid sequence identity, 74.3%, to flamingopox virus, MF678796.1) and DNA topoisomerase (highest amino acid sequence identity, 76.0%, to fowlpox virus, NC_002188.1) revealed that TePV-1 is most closely related to members of the genus Avipoxvirus (Fig. 2).

Fig. 2
figure 2

Phylogenetic relationship of TePV-1 to other members of the subfamily Chordopoxvirinae. (A) Neighbor-joining tree based on the DNA polymerase protein sequence. (B) Neighbor-joining tree based on the DNA topoisomerase protein sequence. (C) Neighbor-joining tree of the full genome sequences based on avipoxviruses. All branches had a bootstrap value of 100. All trees were generated after multiple sequence alignment in CLC Workbench with 1,000 replicates. The GenBank accession numbers for sequences employed in this analysis are as follows: canarypox virus, NC_005309.1; eptesipox virus, NC_035460.1; flamingopox virus, MF678796.1; fowlpox virus, NC_002188.1; hypsugopox virus, MK860688.1; lumpy skin disease virus, NC_003027.1; magpiepox virus, MK903864.1; myxoma virus, NC_001132.2; Nile crocodilepox virus, DQ356948.1; penguinpox virus, KJ859677.1; pigeonpox virus: KJ801920.1; pteropox virus, NC_030656.1; saltwater crocodile poxvirus 1, MG450915.1; saltwater crocodile poxvirus 2, MG450916.1; shearwaterpox virus 1, KX857216.1; shearwaterpox virus 2, KX857215.1; squirrelpox virus, NC_022563.1; tanapox virus, EF420157.1; turkeypox virus, KP728110.2; vaccinia virus, NC_006998.1; variola virus, NC_001611.1; Yaba-like disease virus, NC_002642.1.

Using CLC Workbench, 154 open reading frames coding for ≥100 amino acids were detected, and these sequences were compared to the proteome of other avipoxviruses, using BLASTp. The results of this analysis are presented in Table 1. Apart from the presence of ankyrin repeat proteins at the very beginning and the end of the coding region of the TePV-1 genome, the ORF arrangement was the same as in other avipoxviruses, even though the TePV-1 genome is significantly shorter. Interestingly, nine reading frames coding for proteins of 101-206 aa could not be related to any other sequences based on amino acid sequence identity or the presence of conserved domains. Seven ORFs were found to encode proteins that are not related to other poxvirus proteins but are related to proteins found in eukaryotes (see Supplementary Table 1) and could have been acquired by horizontal gene transfer [18]. These sequences, as well as the assembly site of the two contigs and the transitions to the inverted terminal repeats were confirmed by specific PCR and Sanger sequencing.

Table 1 Annotation of ORFs in the genome of TePV-1 encoding proteins larger than 100 aa. The start and end of ORFs located on the positive strand are shown in bold. The length of each ORF is given in base pairs. "Protein" indicates the highest-rated BLAST hit for this ORF, which is further specified by its accession number (GenBank no.), location in the genome (ref loc), the % amino acid sequence identity (% identity), the alignment length (ali), the e-value and the species. AC, Anolis carolinensis; CG, Cricetulus griseus; CNPV, canarypox virus; EA, Equus asinus; FePV2, pigeonpox virus; FGPV, flamingopox virus; FWPV, fowlpox virus; LA, Lingula anatina; MPPV, magpiepox virus; N., ORF number; PEPV, penguinpox virus; PM, Protobothrops mucrosquamatus; SWPV, shearwaterpox virus; TKPV, turkeypox virus

Poxvirus-like lesions and infections have been described in various reptiles, including crocodilians, tortoises, chameleons, and tegus [6,7,8,9,10,11, 19]. Despite their description in the literature, they have not yet been characterized at the genetic level, except for poxviruses in Nile and saltwater crocodiles (Nile crocodilepox virus [CRV] and saltwater crocodilepox virus subtypes 1 and 2 [SwCRV1/2], respectively) [10, 11]. This first report of the genome sequence of a poxvirus causing disease in a lizard reveals it to be most closely related to avipoxviruses. This is surprising, considering the phylogenetic distance between avian and reptilian species and the differences in homeostasis. The GC content of TePV-1 (35%) is also more similar to that of avipoxviruses than to the known crocodile-infecting poxviruses (62% for CRV, Sw-CRV-1 and -2) [10, 11]. Interestingly, the initial diagnostic PCR only was positive when a primer set for high-GC-content poxviruses was employed. This might indicate that poxviruses of Reptilia are quite variable. Therefore, additional research is warranted to examine the diversity of poxviruses of Reptilia and their species specificity and zoonotic potential.