Biological activity and genome composition of a Tunisian isolate of Spodoptera littoralis nucleopolyhedrovirus (SpliNPV-Tun2)

Background: The baculovirus Spodoptera littoralis nucleopolyhedrovirus (SpliNPV) is an entomopathogenic virus utilized as a biological control agent of the Egyptian cotton leaf worm, Spodoptera littoralis. Several studies have focused on the identification of different SpliNPV isolates from a biological and molecular point of view, but few of them conducted in-depth analyses of the genomic composition of these isolates. Results: Identification of a novel isolate of SpliNPV, termed Tun2, which was purified from infected S. littoralis larvae from Tunisia was reported. This isolate was propagated in vivo and its median lethal concentration (LC 50 ) was determined to be 1.5 × 10 4 occlusion bodies (OBs)/ml for third instar S. littoralis larvae at 7 days of post-infection. OB production in late fourth instar larvae was estimated to be at least 2.7 × 10 9 OBs/g larval weight. The completely sequenced genome of SpliNPV-Tun2 was 137,099 bp in length and contained 132 open reading frames (ORF). It showed a 98.2% nucleotide identity to the Egyptian isolate SpliMNPV-AN1956, with some striking differences; between both genomes, insertion and deletion mutations were noticed in 9 baculovirus core genes, and also in the highly conserved polyhedrin gene. The homologs of ORF 106 and ORF 107 of SpliNPV-AN1956 appeared to be fused to a single ORF 106 in SpliNPV-Tun2, similar to the homologous ORF 110 in SpltNPV-G2. Conclusion: SpliNPV-Tun2 is proposed as a new variant of SpliNPV and a potential candidate for further evaluation as a biocontrol agent for S. littoralis and probably other Spodoptera species .


Background
The Egyptian cotton leaf worm Spodoptera littoralis (Boisd.) (Lepidoptera: Noctuidae) is considered one of the major pests of cotton, tobacco, and corn in the Mediterranean Area and Asia. Larvae of S. littoralis are polyphagous, causing substantial economic losses in both greenhouse and open field crops on a broad range of ornamental, industrial, and vegetable crops (Martins et al. 2005). Due to the severe damage to various crops, controlling this pest is an important issue for integrated pest management. Up to now, S. littoralis management has mainly focused on chemical insecticides. However, numerous studies have been carried out on the possibility of biological control of the pest. Insect viruses and entomopathogenic bacteria, fungi, and nematodes have been investigated as biological control agents of S. littoralis (Hajek and Shapiro-Ilan, 2018). The Spodoptera littoralis nucleopolyhedrovirus (SpliNPV) is a baculovirus that has been evaluated, registered, and applied for control of S. littoralis, as well as the fall armyworm, Spodoptera frugiperda, and the tobacco cutworm Spodoptera litura in Africa, America and Japan (Abdel-Khalik Page 2 of 15 Ben Tiba et al. Egyptian Journal of Biological Pest Control (2022) 32:71 et al.2017;El-Sheikh 2015;Takatsuka et al. 2016). Baculoviruses comprise a large group of double-stranded, circular DNA viruses that infect insects from the orders Lepidoptera, Hymenoptera, and Diptera. Many of these viruses have been investigated because of their potential as biological control agents against agricultural and forest pests (Moscardi 1999). Based on phylogenetic analysis, the Baculoviridae family is separated into 4 genera: Alphabaculovirus (lepidopteran-specific nucleopolyhedroviruses, NPVs), Betabaculovirus (lepidopteranspecific granuloviruses, GVs), Gammabaculovirus (hymenopteran-specific NPVs) and Deltabaculovirus (dipteran-specific NPVs) (Jehle et al. 2006). SpliNPV belongs to the species Spodoptera littoralis nucleopolyhedrovirus of the genus Alphabaculovirus (Harrison et al. 2018). Different SpliNPV variants have been isolated from cotton leaf worm populations in different countries, and intra-specific variation between isolates was identified by restriction endonuclease or partial gene sequencing (Breitenbach et al. 2013;Cherry and Summers 1985;Kislev and Edelman 1982;Martins et al. 2005;). So far, only the Egyptian isolate SpliMNPV-AN1956 has been fully sequenced; its genome is 137,998 bp in length, harbours 132 ORFs, and 15 homologous repeat regions (hrs), and is closely related to the nucleopolyhedrovirus G2 (SpltNPV-G2). Comparisons of the genome sequence of SpliMNPV-AN1956 and SpltNPV-G2 revealed an average of 85% amino acid identity across all genes and high collinearity of the 2 genomes, despite the lack/gain of 16 ORFs (Pang et al. 2001). It was reported that NPVs isolated from Spodoptera spp. have a rather narrow host range (Jakubowska et al. 2010). For example, the Spodoptera exigua multiple nucleopolyhedrovirus (SeMNPV) infects only larvae of its host S. exigua (Jakubowska et al. 2010), whereas SpliNPV was shown to be infectious also to S. frugiperda, S. exigua, and S. litura (Takatsuka et al. 2016). Recently, a Tunisian isolate, named SpliMNPV-Tun, was detected in 2008 from infected cotton leaf worm caterpillars collected in Tunisian tomato greenhouses and identified as a SpliNPV variant based on the partial polyhedrin (polh) gene sequence (Laarif et al. 2011). Here, the identification of a further SpliNPV isolate, termed Tun2, which was obtained from a S. littoralis colony that was established from collected caterpillars from tomato fields in 2013 is reported. This isolate was tested for its activity towards third instar S. littoralis larvae, and its complete genome was determined to study its relationship to other SpliNPV variants.

Insects and virus detection
Larvae of S. littoralis were collected in 2013 from tomato fields (Monastir, Central-East, Tunisia) to establish a laboratory colony at the laboratory of entomology at Regional Research Centre in Horticulture and Organic Agriculture (CRRHAB). For colony maintenance, larvae were fed on a semi-artificial diet (Shorey and Gaston 1965) and kept at a temperature of 28 ± 2 °C and 60 ± 5% relative humidity (RH). Adults were kept in cylinders, and egg deposits were collected on filter papers. The filter papers were transferred to Petri dishes until larval hatching. A piece of artificial diet was added to the Petri dishes where larvae were kept until pupation. Occasionally, larvae from the rearing showed symptoms of nucleopolyhedrovirus infection indicating activation of a covert infection of the S. littoralis population. Diseased larvae were removed from the rearing and stored individually at − 20 °C.

Occlusion body purification
Baculoviral occlusion bodies (OB) were purified from pooled infected cadavers according to El-Salamouny et al. (2003). Briefly, the cadavers were homogenized in sterile distilled water and the homogenate was filtered through a muslin cloth. The obtained crude OB suspension was washed twice with 0.1% sodium dodecyl sulphate (SDS) and pelleted by low centrifugation. The pellet was resuspended in 50 mM Tris/HCl (pH 8), transferred to a 2-ml Eppendorf reaction vial, and HCl (0.1 M) or Na 2 Co 3 (0.1 M) was added to adjust its pH to 7. Then, the OB suspension was centrifuged through a sucrose gradient and resuspended in H 2 O. OB concentration was counted using a Neubauer Cell Counting Chamber (0.1 mm depth) and phase contrast light microscopy (Leica DMRBE, Leica, Wetzlar, Germany) (Eberle et al. 2012).

Bioassays
For testing the biological activity of SpliNPV-Tun2, per os infection experiments were conducted with third instar larvae of S. littoralis. For this, 25 larvae were fed with 1.5-2.5 g artificial diet plugs prepared with final concentrations of 10 3 -10 8 OBs/ml (Shaurub et al. 2014). Untreated control groups consisted of 75 larvae. Each treatment consisted of 3 independent replicates. The mortality data were corrected with untreated control mortality using the formula of Abbott (1925). Calculation of the median lethal concentration (LC 50 ) at 7 days postinfection (dpi) was estimated by Probit analysis using linear regression implemented in the ToxRat 3.2.1 software package (ToxRat Solutions GmbH, Germany). From the same experiment, larval mortality was determined Page 3 of 15 Ben Tiba et al. Egyptian Journal of Biological Pest Control (2022) 32:71 for each concentration at least at 5 different time points within the time range of 1-14 dpi. Statistical analysis was done with R (version 4) and RStudio (version 1.1393). Survival analysis was conducted with R packages "Survival" (version 2.38) and "Survminer" (version 0.4.3). A test of significant variance between Kaplan-Meier curves was performed by a log-rank test (level of significance, P < 0.05).

OB productivity of S. littoralis larvae
An OB dose of 10 4 OBs of SpliNPV-Tun2 was pipetted onto cubes of diet of 5 mm 3 each and individually offered to early fourth instar larvae of S. littoralis (Grzywacz et al. 1998). When the doses were completely ingested within 2 days, a non-contaminated diet was added every second day until 12 dpi. Larvae were harvested at 14 dpi. Three different methods for OB purification were compared; low-speed centrifugation (LSC) (Harrison 2008), sucrose gradient ultracentrifugation (SGU) (El-Salamouny et al. 2003), and sucrose cushion centrifugation (SCC) (Wennmann and Jehle 2014). Purified OBs were counted as described above. OB counting was performed 3 times for each treatment; the obtained concentrations were multiplied with the volume (5 ml), and then normalized with the larva weight. Results were expressed as OBs/g of larval tissue and were used to compute the arithmetic mean of OB/g of each experiment. Differences in the mean number of OB/g were statistically evaluated for a significance value of P ≤ 0.05 using analysis of variance (ANOVA) and the Tukey's Honestly Significant Difference test (Tukey-HSD) comparison of means with standard R code (R version 3.3.1 in RStudio 3.4.0).

Viral DNA extraction
Viral DNA was extracted according to Bernal et al. (2013) with some modification. Occlusion derived virions (ODVs) were released from OBs by mixing 100 µl of the OB suspension (containing about 10 9 OBs/ml) with 100 µl Na 2 CO 3 (0.5 M), 50 µl SDS (10%, w/v) in a final volume of 500 µl. After incubation at 60 °C for 10 min, the suspension was neutralized to pH 7 by adding 0.1 M HCl. Undissolved OBs and other debris were pelleted by centrifugation at 3800 g for 5 min. The supernatant containing the released ODVs was transferred to a fresh vial and treated with 25 µl Proteinase K (10 mg/ml) for 1 h at 50 °C. Viral DNA was extracted twice with Tris/ HCl-saturated phenol and once with chloroform by using Phase Lock gel tubes (all purchased from, Carl Roth GmbH + Co., KG, Karlsruhe, Germany), followed by standard ethanol precipitation (Eberle et al. 2012). The DNA pellet was dissolved in 100 µl distilled H 2 O.

Sequencing and raw data processing
About 50 ng purified DNA was subjected to commercial NexteraXT library preparation and Illumina NextSeq500 sequencing at StarSEQ GmbH company (Mainz, Germany). In total, more than 1.76 million reads of 151 nt in length were obtained. Raw reads were processed by adapter trimming and quality filtering excluding reads with an average phred quality score below 30 (Gueli Alletti et al. 2017). Quality filtered reads with a length shorter than 50 nt were excluded from the analysis for paired reads and 51 nt for unpaired reads. Paired and unpaired reads were kept after all steps of filtering and quality control.

Genome sequence assembly
The remaining set of reads was used for de novo sequence assembly as well as mapping against the whole genome sequence of SpliMNPV-AN1956 (GenBank accession number JX454574) (Breitenbach et al. 2013). CLC de novo assembly resulted in multiple contigs (> 1000 bp). Contigs were mapped and fit together to a single contig comprising the whole genome. This contig was considered as a first consensus (cons1). In a second approach, all reads were mapped against the SpliMNPV-AN1956 genome using BWA-MEM. From here, a second consensus (cons2) was extracted applying a majority rule (> 99%). Both consensus sequences were then aligned to each other and checked for differences, which mainly occurred in repeated as well as homologous repeat regions (hrs). The alignment was then checked manually for ambiguities and sequence discrepancies. The correction was based on the read coverage supporting one ambiguous region per contig generated by CLC. The cutoff of the adopted corrections was coverage of 20 reads per ambiguous region. One final genome sequence of SpliNPV-Tun2 was generated based on the majority of read coverage and submitted to GenBank (Accession number MG958660).

Phylogenetic reconstruction
The 38 core genes of SpliNPV-Tun2 were translated to amino acid sequence, then aligned with core gene amino acid sequences from 88 group II NPVs, 39 group I NPVs, and from CpGV-M and SpliGV-K1 as outgroups using MUSCLE alignment tool v3.8.425 as implemented in Geneious Prime ® v11 (Biomatters Ltd., Auckland, New Zealand) (Edgar 2004). The concatenated alignments of the amino acid sequences of the 38 baculovirus core genes (Wennmann et al. 2018) were then used to infer a phylogenetic tree using the Minimum Evolution method implemented in MEGA.7 (Kumar et al. 2016).

Comparison of the SpliNPV-Tun2 genome to other NPVs
All of the 132 SpliNPV-Tun2 ORFs were tested for sequence similarity using BlastX. A detailed comparison of the similarity with genomes of SpliNPV-AB1956 and SpltNPV-G2 was made. The genome characteristics were compared in terms of length, GC%, ORF number, presence of genes.

Results
In 2013, a laboratory colony of S. littoralis collected from tomato fields in Monastir (Tunisia) was established. In the reared colony, an occasional occurrence of moribund larvae with symptoms of a nucleopolyhedrosis infection was observed. The purification of viral OBs and DNA, PCR amplification using polh specific primers and sequence analysis (data not shown) indicated that the infective agent was a SpliNPV isolate, which was eventually termed SpliNPV-Tun2.

Virulence and OB yield of SpliNPV-Tun2
Concentration mortality bioassays with third instar larvae were performed to determine the virulence of SpliNPV-Tun2. The LC 50 value at 7 dpi was estimated to 1.5 × 10 4 OB/ml with a 95% confidence interval of 0.2-5.6 × 10 4 OB/ml (n = 525, slope probit line = 0.42, Chi 2 = 8.81). The survival rates determined at various time points after infection were inversely proportional to the applied OB concentration of 10 3 -10 8 OB/ ml ( Fig. 1). In the uninfected control, a slight decrease in the survival probability with 84% was observed at 14 dpi [95% Cl (76.1-92.7%)]. A concentration-dependent decrease in the survival probability was observed in the treatment groups starting from 4 dpi with 96.7% [95% Cl (96.0-97.4%)] and reached 7.81% [95% Cl (6.48-9.40%)] at 14 dpi. The median mortality was obtained between 7 dpi for applied concentrations of 10 7 and 10 8 OB/ml and 10 dpi for the lowest concentrations10 3 and 10 6 OB/ ml of SpliNPV-Tun2. To estimate the survival covariance by time and by treatment, the survival was presented by percentage and survival data were normalized with lambda = 0.57 (Table 1). The survival time was statistically different depending on the applied virus concentrations. By using the different concentrations of SpliNPV-Tun2 OBs, all treatments produced different survival percentages depending on the time (F = 7.78, P value < 0.01).  OB productivity of late fourth and fifth instar larvae was quantified. The mean weight of larvae with virus infection symptoms was 1548 mg with a standard deviation (s.d.) of 82.5 mg. The OBs were harvested at 14 dpi when infected larvae were seen as highly moribund. Three different standard methods for OB purification were compared, i.e. LSC, SCC, and SGU (Wennmann and Jehle 2014). OB yield was found to be significantly different among LSC, SCC and SGU purification methods (ANOVA, P ≤ 0.05) [F(2,6) = 88.11, p < 0.001]. LSC yielded 2.7 × 10 9 OB/g larvae weight, followed by SCC 1.3 × 10 9 OB/g larvae weight, whereas SGU yielded only 5 × 10 8 OB/g larvae weight. (Fig. 2).

Genome sequence of SpliNPV-Tun2
A total of 1,597,175 filtered reads amounting to (90.6%) of the total raw reads were used for the analysis. From the total of the filtered reads, 1,508,620 paired reads and 88,555 unpaired reads could be mapped to the reference genome of SpliNPV-AN1956, whereas about 13,500 reads did not map to SpliNPV-AN1956 but gave BLAST hits with insect or bacterial DNA sequences.
The obtained genome consensus sequence of SpliNPV-Tun2 (MG958660) was supported by an average of 720-fold sequencing depth (s.d. = 316). It had a length of 137,099 bp and a GC content of 44.7% (Table 2).
It contained 132 open reading frames (ORF) and 15 homologous repeat regions (hrs). Based on the nucleotide sequences, the genomes of SpliNPV-Tun2 and SpliMNPV-AN1956 were 98.2% identical and the hrs in both genomes were at the same location. The genome of SpliNPV-Tun2 was 899 bp shorter than that of SpliM-NPV-AN1956 through alignment of the 2 genomes revealed the same number of ORFs, but with some differences. Sixty-nine ORF had a 100% predicted amino acid identity to SpliMNPV-AN1956 ORFs. The rest of the ORFs' amino acid identities ranged between 90 and 99%, whereas ORF 37 had only 97% amino acid (aa) sequence identity ( Table 2). The sequences encoding putative proteins accounted for 88.4% for SpliNPV-Tun2, while the coding density was 87.9% in SpliMNPV-AN1956. The number of intergenic regions is 101 for SpliNPV-Tun2 and 102 in SpliMNPV-AN1956, with mean distances of 124 bp and 121 bp, respectively. Some of the intergenic regions consisted of palindromic sequences; both isolates contained 15 hrs at the same genome location (Table 2).

Phylogenetic reconstruction and genetic distance
A minimum evolution phylogenetic tree based on the concatenated amino acid sequences of 38 baculovirus core genes of group I and group II NPVs was inferred (Fig. 3). It corroborated the close relationship between SpliNPV-Tun2 and -AN1956. The next neighbour to both isolates was SpltNPV-G2. The SpliNPV isolates and SpltNPV-G2 are only distantly related to other Spodoptera-specific NPVs, such as SeMNPV, SpltNPV-II and Spodoptera frugiperda multiple nucleopolyhedrovirus (SfMNPV) (Fig. 3).
For baculovirus species demarcation, the Kimura-2-Parameter (K2P) distance of the 38 baculovirus core genes can be used as criterion, according to which, 2 isolates are considered to belong to the same species if their K2P distance is smaller < 0.021 and to different species if the K2P distance is > 0.072 (Wennmann et al. 2018). With a K2P distance of 0.001, SpliNPV-Tun2 and -AN1956 are 2 isolates belonging to the same species Spodoptera littoralis nucleopolyhedrovirus. In contrast, SpltNPV-G2 is distant enough (0.099 from the two viruses to constitute a separate species Spodoptera litura nucleopolyhedrovirus (Wennmann et al. 2018). The other known Spodoptera-specific NPV isolates constitute several other discrete species (Fig. 3).

Main genome differences between SpliNPV-Tun2 and -AN1956
Compared to SpliNPV-AN1956, the new isolate SpliNPV-Tun2 showed insertion and deletion mutations in 62 ORFs, of which 37 ORFs are with predicted function. ORFs with significant changes caused by deletions and insertions are illustrated in (Fig. 4). These differences affect ORFs coding for predicted virus proteins related to virus structure, such as the structural protein PP78/81, the capsid-associated protein VP80 and VP1054, the OB matrix protein (Polyhedrin, POLH), the nucleotide metabolism (Ribonucleotide Reductase, RR1), proteins involved in viral DNA replication (Late Expression Factor 2 (LEF-2) and LEF-10, Protein kinase 1 (PK-1), LEF-5, and the group II Alphabaculovirus-specific HOAR and the BRO-a. Furthermore, a considerable number of amino acid changes were found but will not be further detailed here. A notable difference is the presence of a tyrosine residue close to the N-terminus fifth amino acid position of the predicted POLH of SpliNPV-Tun2, a residue which is missing in the POLH of SpliNPV-AN1956 (Fig. 4). Another difference between the genome sequences of SpliNPV-Tun2 and -AN1956 is related to ORFs 106 and Lymantria Sp. Baculovirus phylogeny based on Minimum Evolution (ME) method of amino acid sequences of 38 core genes of different NPV. Translated amino acid sequences of each core gene were separately aligned, and alignments were concatenated using Geneious 8. The consensus tree was obtained by a heuristic search with 500 bootstrap replicates. The Model is the JTT model (Jones et al. 1992) with a uniform rate among site and without invariant site. Bootstrap values (> 50%) are shown at each node. Only Alphabaculovirus group II is given in detail. Alphabaculovirus group I and Betabaculovirus were reduced. Betabaculoviruses represented with only CpGV-M and SpliGV-K1 for this analysis were used as outgroup Page 12 of 15 Ben Tiba et al. Egyptian Journal of Biological Pest Control (2022) 32:71 107. Whereas in SpliNPV-AN1956 two ORF 106 and ORF 107 were located from genome position 110,884 < 111,843 (319 aa) and 111,873 < 112,064 (63 aa), respectively, these two ORFs were identified as one single ORF in SpliNPV-Tun2 (ORF 106, genome position 109,998 < 111,161, (387 aa)). The split of the ORF 106 homolog of SpliNPV-Tun2 into two ORFs 106 and 107 in SpliNPV-AN1956 is caused by a missing thymidine residue at genome position 110,991 of SpliNPV-AN1956, causing a frameshift and separation into two ORFs (Fig. 5). Interestingly, a similar homologous ORF 110 (genome position