A new approach for detecting adventitious viruses shows Sf-rhabdovirus-negative Sf-RVN cells are suitable for safe biologicals production

Abstract

Background: Adventitious viral contamination in cell substrates used for biologicals production is a major safety concern. A powerful new approach that can be used to identify adventitious viruses is a combination of bioinformatics tools with massively parallel sequencing technology. Typically, this involves mapping or BLASTN searching individual reads against viral nucleotide databases. Although extremely sensitive for known viruses, this approach can easily miss viruses that are too dissimilar to viruses in the database. Moreover, it is computationally intensive and requires reference cell genome databases. To avoid these drawbacks, we set out to develop an alternative approach. We reasoned that searching genome and transcriptome assemblies for adventitious viral contaminants using TBLASTN with a compact viral protein database covering extant viral diversity as the query could be fast and sensitive without a requirement for high performance computing hardware.

Results: We tested our approach on Spodoptera frugiperda Sf-RVN, a recently isolated insect cell line, to determine if it was contaminated with one or more adventitious viruses. We used Illumina reads to assemble the Sf-RVN genome and transcriptome and searched them for adventitious viral contaminants using TBLASTN with our viral protein database. We found no evidence of viral contamination, which was substantiated by the fact that our searches otherwise identified diverse sequences encoding virus-like proteins. These sequences included Maverick, R1 LINE, and errantivirus transposons, all of which are common in insect genomes. We also identified previously described as well as novel endogenous viral elements similar to ORFs encoded by diverse insect viruses.

Conclusions: Our results demonstrate TBLASTN searching massively parallel sequencing (MPS) assemblies with a compact, manually curated viral protein database is more sensitive for adventitious virus detection than BLASTN, as we identified various sequences that encoded virus-like proteins, but had no similarity to viral sequences at the nucleotide level. Moreover, searches were fast without requiring high performance computing hardware. Our study also documents the enhanced biosafety profile of Sf-RVN as compared to other Sf cell lines, and supports the notion that Sf-RVN is highly suitable for the production of safe biologicals.

Keywords: Adventitious virus; Endogenous viral element; Errantivirus; Massively parallel sequencing; Maverick/Polinton/Polintovirus; Spodoptera frugiperda Sf-RVN.

Fig. 1

Workflow of the approach presented in this paper

Fig. 2

Structure and relative sizes of Sf Maverick 1 and 2 elements. ORFs conserved among Sf Maverick 1 and 2 are indicated: ORFs encoding putative proteins with conserved domains are shaded in light gray (ATP, INT, CAP, PRO and POLB), and ORFs that do not contain identifiable domains are shaded in dark gray (ORF2, ORF6-8)

Fig. 3

Sf-RVN cell Maverick 1 and 2 POLB are more similar to homologous proteins encoded by other Mavericks than to related viral proteins. a Amino acid alignment of the Sf Maverick 1 and 2 POLB proteins and the B. mori densovirus 3 DNA polymerase. The predicted Recombination Endonuclease VII and DNA POLB2 domains are indicated. b Unrooted tree showing the phylogenetic relationships among Sf Maverick and other insect and vertebrate Maverick POLB proteins, and B. mori densovirus 3 DNA polymerase. Abbreviations: At, the navel orangeworm Amyelois transistella; Bm, the silkworm Bombyx mori; Dr, the zebrafish Danio rerio; Px, the diamondback moth Plutella xylostella; Sc, the stable fly Stomoxys calcitrans; Sf, the fall armyworm Spodoptera frugiperda (this study); Tc, the flour beetle Tribolium castaneum

Fig. 4

Sf-RVN cell Maverick 1 and 2 ATP and CAP are more similar to homologous proteins encoded by other Mavericks than to related viral proteins. a Amino acid alignment of the AAA+ ATPase domain regions of Sf Maverick 1 and 2 ATP, vertebrate Maverick ATP, cowpox virus A32 ATPase, and A. honmai entomopoxvirus ATPase. Conserved amino acid residues that constitute the ATPase Walker A and B motifs are indicated with a filled circle. b Unrooted tree showing the phylogenetic relationships among Sf Maverick and other insect and vertebrate Maverick ATP proteins, cowpox virus A32 ATPAse and A. honmai entomopoxvirus ATPase. c Amino acid alignment of the N-terminal regions containing the PLA2 domain of Sf Maverick 1 and 2 CAP, human parvovirus B19 capsid, B. mori densovirus 1 capsid, and A. honmai entomopoxvirus capsid-like protein. Conserved amino acids involved in catalysis are indicated with a filled circle, those involved in Ca2+ cofactor binding with an asterisk, and those that are otherwise conserved with an plus sign. d Unrooted tree showing the phylogenetic relationships among Sf Maverick and other insect Maverick CAP proteins, human parvovirus B19 capsid, B. mori densovirus 1 capsid, and A. honmai entomopoxvirus capsid-like protein. Abbreviations: Ah, the summer fruit tortix Adoxophyes honmai; At, the navel orangeworm Amyelois transistella; Bm, the silkworm Bombyx mori; Dr, the zebrafish Danio rerio; Ln, the black garden ant Lasius niger; Px, the diamondback moth Plutella xylostella; Sf, the fall armyworm Spodoptera frugiperda (this study); Tc, the flour beetle Tribolium castaneum

Fig. 5

Sf-RVN cell Maverick 1 and 2 INT and PRO are more similar to homologous proteins encoded by other Mavericks than to related viral proteins. a Amino acid alignment of the conserved integrase domain containing regions of Sf Maverick 1 and 2 INT, vertebrate Maverick CAP, and HIV integrase. Conserved amino acids involved in catalysis are indicated with a filled circle. b Unrooted tree showing the phylogenetic relationships among Sf Maverick and other insect and vertebrate Maverick INT proteins, and HIV integrase. c Amino acid alignment of the conserved cysteine endoprotease domain contraining regions of Sf Maverick 1 and 2 PRO, vertebrate Maverick PRO, and adenovirus adenain. Conserved amino acids involved in catalysis are indicated with a filled circle. d Unrooted tree showing the phylogenetic relationships among Sf Maverick and other insect and vertebrate Maverick PRO proteins, and adenovirus adenain. Abbreviations: At, the navel orangeworm Amyelois transistella; Bm, the silkworm Bombyx mori; Dr, the zebrafish Danio rerio; Ln, the black garden ant Lasius niger; Px, the diamondback moth Plutella xylostella; Sc, the stable fly Stomoxys calcitrans; Sf, the fall armyworm Spodoptera frugiperda (this study); Tc, the flour beetle Tribolium castaneum

Fig. 6

Sf-RVN cell R1 LINE transposons encode proteins with SF1H domains that are more similar to previously identified insect R1 LINE ORF2 proteins with C-terminal SF1H helicase domains, than to proteins from (+)ssRNA viruses. a Domain structure and relative size of full-length Sf-RVN and Plutella xylostella R1 LINE with SF1H domain. The size and order of conserved domains in ORF2 are very similar. b Predicted amino acid sequence alignment of SF1H domains of Sf-RVN and Plutella xylostella R1 LINEs, and Negev virus (YP_009256205.1) and TMV replicase (NP_597746.1). The shading is proportional to the degree of amino acid conservation. c Phylogenetic analysis of newly identified Sf-RVN and Plutella xylostella R1 LINE SF1H domains, and Negev virus and TMV replicase. The LINEs cluster closely together in a group that is distant from ssRNA viruses

Fig. 7

Phylogenetic analysis of the Sf cell Taï virus N-like EVE product, homologous lepidopteran gene products, and related bunyavirus N proteins. The lepidopteran gene products cluster closely together in a group that is distant from the bunyavirus N proteins

Fig. 8

Sf-RVN cells contain diverse transcribed errantivirus (ErV) sequences. a Phylogenetic analysis of Sf-RVN errantivirus sequences containing intact ORFs 1, 2 and 3 based on an alignment of the predicted protein products. b Normalized read abundances for Sf-RVN errantivirus sequences containing intact ORFs 1, 2, and 3. c Phylogenetic analysis of all identified Sf-RVN errantivirus sequences based on an alignment of their DNA sequences. Parentheses indicate previously identified sequences [20] found in the larger supercontigs assembled in this study (See Additional file 1, Sf-Errantiviruses)