A newly discovered member of the Atlastin family, BmAtlastin-n, has an antiviral effect against BmNPV in Bombyx mori

Atlastin is a member of the dynamin protein superfamily and it can mediate homotypic fusion of endoplasmic reticulum (ER) membranes, which is required for many biological processes. In this study, a new Atlastin homologous protein, BmAtlastin-n, was characterized in silkworms and was found to contain an N-terminal conserved GTPase domain and a coiled-coil middle domain. BmAtlastin-n is localized in the cytoplasm and enriched in silkworm midgut. Results also showed that overexpression of BmAtlastin-n in BmN-SWU1 cells could enhance resistance to BmNPV. To better confirm its antiviral effect, microRNA was used to knock down the expression of BmAtlastin-n in BmE-SWU1 cells with inducing the reproduction of BmNPV. A transgenic expression vector of BmAtlastin-n was constructed and introduced to silkworm embryos by microinjection. The transgenic silkworm also showed considerable antiviral capacity. In conclusion, these findings demonstrate that BmAtlastin-n plays an important role in BmNPV defense. More importantly, the current study may provide a new clue for Atlastin research.

nucleopolyhedrovirus (BmNPV) 22 , a member of the Baculoviridae family that causes widespread epidemics of nuclear polyhedrosis (also known as "grasserie" or "jaundice") and limits the production of silk and the development of sericulture 23,24 . BmN-SWU1 cells are vulnerable to BmNPV, while BmN-SWU2 cells lack the interaction between host and BmNPV during the viral entry process, which inhibits the reproduction of BmNPV 22 . Although there are no clear reports about the effects of Atlastin against various pathogens, it is here speculated that the different patterns of expression of this protein may be one of the targets for BmNPV resistance in the two cell lines.
In this study, an Atlastin homeotic gene was cloned from Dazao strain, identified and named BmAtlastin-n. To confirm speculation that BmAtlastin-n may have anti-BmNPV effects, proliferation of the virus via overexpression and interference of BmAtlastin-n in silkworm cells was here investigated. The results confirmed that BmAtlastin-n plays an important role in resistance to BmNPV infection in vitro. A silkworm strain overexpressing BmAtlastin-n was produced using transgenic technology. It showed significant resistance to BmNPV infection. Overall, these findings confirmed that BmAtlastin-n overexpression will inhibit the reproduction of BmNPV.

Results
Cloning and identification of BmAtlastin-n. BmAtlastin-n was cloned from larval cDNA from the Dazao strain, which has a 2313 bp coding sequence (CDS) encoding a 770 amino acid protein. BmAtlastin-n contained an N-terminal conserved GTPase domain and a coiled-coil middle domain but no transmembrane segments (Fig. S1). There were four conserved GTP-binding motifs, including the phosphates coordination of G1 (also known as P-loop), catalysis of G2, the formation of a hydrogen bond with the γ -phosphate of GTP of G3, and ribose coordination of G4 (Fig. 1A, Fig. S1). The two switch regions (switch I and switch II) were also observed in sequence analysis. These appeared as surface loops that would induce conformational changes upon GTP binding. Members of the dynamin superfamily were grouped using Bayesian inference method, including classical dynamins, dynamin-like proteins (DLPs), optic atrophy 1 (OPA1), Mitofusins, Mx proteins, guanylate binding proteins (GBPs), and Atlastins. Phylogenetic tree analysis showed that Atlastins and GBPs form one group and that BmAtlastin-n belongs to the Atlastin subfamily with closest relationship to silkworm Atlastin3 and silkworm Atlastin4 (Fig. 1B). These results suggested that BmAtlastin-n is a homologous gene of the Atlastin family.

Location and pattern of expression of BmAtlastin-n. The subcellular localization of endogenous
BmAtlastin-n was detected in BmN-SWU1 cells and BmN-SWU2 cells using immunofluorescence with specific BmAtlastin-n antibody (anti-BmAtlastin-n). The laser confocal images showed BmAtlastin-n was to be located in the cytoplasm in BmN-SWU2 cells and was not detected in BmN-SWU1 cells ( Fig. 2A). qRT-PCR and Western blot analysis were used to analyze the level of transcription and protein expression in order to assess the differential expression of BmAtlastin-n in the two ovarian cell lines. Results showed BmAtlastin-n exhibited high levels of expression in BmN-SWU2 cells and little or no expression in BmN-SWU1 cells (Fig. 2B,C). pIZ-BmAtlastin-n eukaryotic expression vector was constructed to determine the location of exogenous BmAtlastin-n in BmN-SWU1 cells and the exogenous BmAtlastin-n was also observed in the cytoplasm ( Fig. 2A).
The pattern of BmAtlastin-n expression was analyzed in different tissues of third-day of fifth-instar larvae of Dazao by RT-PCR, including the head, integument, midgut, silk gland, malpighian tubule, testis, ovary, fat body, and hemocyte. Results showed BmAtlastin-n was highest transcribed in midgut and minimum in silk gland (Fig. 2D). Meanwhile, the testis, ovary, fat body, and hemocyte had higher expression levels. When BmNPV infects silkworms, the virus must cross the midgut, first cell barrier, and were further blocked by the immune tissues of hemocyte and fat body. Overall, BmAtlastin-n was located in the cytoplasm and was found to have little or no expression in BmN-SWU1 cells. BmAtlastin-n showed different transcription levels in different tissues and different patterns of expression in the two silkworm ovarian cell lines. It is here speculated that these phenotypes are caused by the BmAtlastin-n function.

Inhibition of BmNPV reproduction via overexpression of BmAtlastin-n. The effects of
BmAtlastin-n overexpression on the replication of BmNPV in BmN-SWU1 cells were analyzed, treating the BmAtlastin-n expression level and the differential BmNPV susceptibilities of two ovarian cell lines. First, pIZ-BmAtlastin-n was transfected into BmN-SWU1 cells and the transcriptional level of BmAtlastin-n was significantly higher in transfected cells (Fig. 3A). Next, v39K prm -EGFP reassortant budded viruses (BVs) were used to incubate BmN-SWU1 cells of BmAtlastin-n overexpression and pIZ/V5-His control cells, and infection states were analyzed at 12 h, 24 h, and 48 h. Fluorescence images and statistical proportions of infected cells (GFP positive) showed that BmAtlastin-n overexpressed cells had fewer GFP positive cells than the control cells at 24 h and 48 h after infection (Fig. 3B, Fig. S2). Because the GFP was promoted by a 39 K viral late promoter, fluorescence was not observable in infected cells at 12 h. BmNPV infection rates of the two treatments were analyzed by flow cytometry to further confirm the inhibitory effects of virus reproduction by overexpression of BmAtlastin-n. The virus infected (GFP positive, GFP+ ) cells of BmAtlastin overexpressing (BmAtlastin-n-OE) cells was 5.30%, which was lower than 12.04% GFP+ cells of the control at 24 h post infection. At 48 h after virus infection, the 36.23% of BmAtlastin-n-OE cells were still significantly fewer than the 55.10% of control cells (Fig. 3C). To confirm the inhibitory effects of BmNPV, the level of expression of VP39, which is a late baculovirus capsid gene widely used as a marker gene of BmNPV, was investigated using qRT-PCR and Western blot analysis. The data showed VP39 expression levels to be significantly inhibited in BmAtlastin-n overexpressing cells (Fig. 3D,E). In conclusion, overexpression of BmAtlastin-n in BmN-SWU1 cells was found to inhibit the reproduction of BmNPV. BmNPV infection. A BmE-SWU1 cell line with high levels of BmAtlastin-n transcription was selected, and the expression of BmAtlastin-n was also inhibited by transfection with pIZ/DsRed-BmAtlastin-n-RNAi plasmids (Fig. 4A). v39K prm -EGFP reassortant BVs were used to incubated with BmAtlastin-n interfered BmE-SWU1 cells and control BmE-SWU1 cells. The data collected from flow cytometry analysis showed that the number of GFP+ cells increased in response to BmAtlastin-n interference in BmE-SWU1 cells. The proportions of BmAtlastin-n-RNAi cells at 24 h and 48 h post infection were 25.78% and 63.74%, which were more than the control values, 17.58% and 54.90% (Fig. 4B). Next, the transcription level of VP39 was determined for confirmation of the proliferation of the virus (Fig. 4C). The data showed VP39 transcript levels was significantly increased by interference of BmAtlastin-n at 12 h, 24 h, and 48 h after virus infection. The VP39 protein expression of the two groups was investigated further, and the results also indicated that level of VP39 expression have increased (Fig. 4D). These results suggested that BmAtlastin-n interference can induce reproduction of BmNPV in BmE-SWU1 cells.

Generation of BmAtlastin-n overexpression in transgenic silkworms. It was here determined that
BmAtlastin-n overexpression can suppress the reproduction of BmNPV in vitro. BmAtlastin-n was overexpressed in individuals to confirm its antiviral effects. PiggyBac [3 × P3-EGFP, IE1P-BmAtlastin-n] transgenic vector, BmAtlastin-n controlled by IE1 promotor, and Ser1 polyA termination signals of transcription were constructed (Fig. 5A). The transgenic vector piggyBac [3 × P3-EGFP, IE1P-BmAtlastin-n] and the helper vector pHA3PIG were co-injected into Dazao embryos to generate G0 offspring. G0 moths were inbred or backcrossed to product G1 progeny. Transgenic silkworms were produced via screening for GFP-positive expression from embryos and months of same brood of G1 (Fig. 5B). Inverse PCR was used to analyze the insertion site, and the PCR products were cloned and sequenced. Results indicated that fragments were inserted in the genome successfully, and there was only one insertion site. It was located on nscaf 2176 of chromosome 11 th (Fig. 5C).
The level of expression of BmAtlastin-n was detected in BmAtlastin-n-OE transgenic silkworm and Dazao control silkworm. qRT-PCR data suggested that BmAtlastin-n had been successfully overexpressed in transgenic lines (Fig. 5D). To confirm that the expression of nearest genes BGIBMGA001648 and BGIBMGA001649 to the left and right of BmAtlastin-n insertion site were affected in BmAtlastin-n-OE transgenic line, the level of transcription of the two genes was analyzed in two silkworm lines. Results showed there to be no significant difference (Fig. 5E). BGIBMGA001648 was predicted to encode ferrochelatase, a mitochondrial-like protein. No proteins showed similarity to BGIBMGA001649 under blast analysis. These results indicated that a BmAtlastin-n-OE transgenic line had been produced. This line could be used for next experiments.

Enhancement of anti-BmNPV capacity in BmAtlastin-n-OE transgenic line. To investigate
the anti-BmNPV capacity of the BmAtlastin-n-OE transgenic line, first-day fourth-instar larvae of the BmAtlastin-n-OE transgenic silkworm and Dazao control silkworm were infected with BmNPV via occlusion-derived viruses (ODVs) oral infection and BV infection. Each larva was fed 1 × 10 6 ODVs and injected with 1 × 10 6 BVs per larva. Then, mortality statistics were calculated each day until the 12 th day after infection. The oral ODV mortality of Dazao silkworm and BmAtlastin-n-OE transgenic silkworm was 40.3% and 26.3%, respectively, and the BV infection groups of Dazao silkworm and BmAtlastin-n-OE transgenic silkworm were 48.7% and 28% (Fig. 6A,B). The BmAtlastin-n-OE transgenic line had lower mortality in both oral infection and BVs injection groups than the Dazao line.
To further confirm the anti-BmNPV capacity of BmAtlastin-n-OE transgenic line, VP39 expression levels were analyzed 24 h and 48 h after oral ODV and BV injection. The qRT-PCR results suggested that overexpression of BmAtlastin-n inhibited the reproduction of BmNPV in vivo (Fig. 6C,D). The BmAtlastin-n-OE transgenic line was found to have significantly less reproduction of BmNPV.

Discussion
BmAtlastin-n was identified as a highly expressed protein in BmN-SWU2 cells using iTRAQ analysis of two cell lines. BmAtlastin-n formed a group with Atlastins and contained a GTPase domain, a coiled-coil middle domain and a C-terminal domain with no known function (Fig. S1). This is the fifth Atlastin homologous gene to be found in silkworms. However, it lacked the typical transmembrane segments of Atlastins, silkworm Atlastin3 and Atlastin4 also lacked. These three Atlastins could form the same branch. The other two silkworm Atlastins, Atlastin1 and Atlastin2, possessed characteristics typical of the Atlastin family, containing the typical transmembrane segments. They showed high homology with Drosophila melanogaster Atlastin and other entomic Atlastins. Atlastins and GBPs had a very high homology, sharing an N-terminal conserved GTPase domain, and they may have arisen from a common ancestor 5 . In the present study, no GBPs were found in the genomes of silkworms, Ciona, sea urchin, Drosophila, or C. elegans 25 . For these reasons, BmAtlastin-n was placed in the Atlastin subfamily.
The pattern of expression of BmAtlastin-n was investigated in two ovarian cell lines that the results were consistent with iTRAQ analysis. The levels of expression in tissues showed BmAtlastin-n to be enriched in the midgut. BmAtlastin-n was located in the cytoplasm in two ovarian cell lines ( Fig. 2A). Whether it is also located in intracellular membranes like the Atlastins of other species needs further research. Atlastins are a subfamily of the dynamin superfamily, and recent work has focused mainly on why Atlastin mutations cause HSP and other effects in the endomembrane. However, there were no clear reports that Atlastins could take part in host innate immunity. It was here confirmed that BmAtlastin-n overexpression could inhibit the reproduction of BmNPV both in vitro and vivo, and the interference of BmAtlastin-n was found to promote viral replication in BmE-SWU1 cells (Figs 3 and 6). However, there were no effects on viral infections via interfering BmAtlastin-n in BmN-SWU2 cells (Fig. S3). It was here speculated that BmN-SWU2 cells lack the entry mechanisms of BmNPV, causing the interference of BmAtlastin-n to have no effect on virus infection. In a previous study, results showed the entrance of BmNPV to be triggered by overexpressing BmREEPa in BmN-SWU2 cells 20 . However, the BmNPV that entered the BmN-SWU2 cells was still inhibited. This phenomenon suggested that there were other factors involved in resisting the proliferation of BmNPV inside the cells and BmAtlastin-n may be one of those factors. To explain why BmAtlastin-n exhibited an anti-BmNPV function, the fact that Atlastins and GBPs showed high homology and that GBPs plays an important role in resistance to many pathogens were considered. The functions of GBPs in the immune defense against Chlamydia trachomatis 26 32 have been characterized. Results also showed Atlastins to be similar to GBPs from the results of phylogenetic tree analysis (Fig. 1B). Taken together, these results suggested that BmAtlastin-n has functions similar to GBPs in the evolutionary process. However, the anti-BmNPV mechanism of BmAtlastin-n remains unknown.
Overexpression of an endogenous or exogenous antiviral gene is an effective method of improving the antiviral capacity of silkworms using transgenic technology 33 . However, most of the studies of antiviral genes are performed in vitro level, and any antiviral effects must be validated in the silkworms. In the current study, a BmAtlastin-n-OE transgenic line was generated via embryo microinjection. The transgenic silkworms reduced 21% mortality than controls after BVs injection and 14% mortality after oral infection (Fig. 6A,B). These results demonstrated that overexpression of BmAtlastin-n could enhance silkworm resistance to BmNPV. Previous studies have shown that the transgenic silkworm LI-A of Bmlipipase overexpression could improve the 33% survival rate after oral infection with BmNPV and the transgenic silkworm HEKG-B by overexpression of hycu-ep32, which was controlled by hr3 combined with 39 KP. The mortality of HEKG-B was then ≈30% lower than in controls 34,35 . The antiviral capacity of BmAtlastin-n-OE transgenic line was slightly lower than in these transgenic strains. This may have been caused by the differences in the promoter or strain. More importantly, the current study provides material for anti-BmNPV cultivation of silkworms that may facilitate functional research of silkworm Atlastins. This is the first study to show that the homeotic gene of Atlastins plays a role in innate immunity of silkworms and may provide new insights into the determination of Atlastin function. Silkworm Atlastin3 and Atlastin4 showed more homology with BmAtlastin-n, which lacked the typical transmembrane segments. More studies will be required to confirm whether they also have anti-BmNPV capacity. Vector construction. BmAtlastin-n was amplified with primers (forward 5′ CCATATTCAACGGCGAAGTC 3′ and reverse 5′ ACATTAAGGATAGGCGAGCA 3′ ) and cloned into pMD19-T vector (Takara, Dalian, China). The correct fragments were obtained by PCR using the primers (forward 5′ cgggatccATGAGCAGTCTCGGT-GAAGC 3′ and reverse 5′ ccctcgagCGTATACCGAACTCACGAGTG 3′ ) from pMD19-BmAtlastin-n. The second products and insect expression vector pIZ/V5-His (Invitrogen) were ligated using BamH I and Xho I sticky ends and constructed final vector pIZ-BmAtlastin-n-OE.
Cell transient transfection. Before transfection, the highly pure plasmids were prepared using Plasmid Mini Kits (Qiagen, Germany). Cells in log phase were cultured in 6-well plates or 24-well plates for approximately 24 h before transfection. Each well was transfected with plasmids and X-treme GENE HP DNA Transfection Reagent (Roche) mixture, which was mixed in 200 μ l antibiotic-free and serum-free medium according to manufacturer's instruction. Transfection medium was removed and medium containing antibiotic and serum were added after 6 h. The mixture was then cultured for 72 h.
Quantitative real time-PCR (qRT-PCR). Total RNA was purified from each samples using Total RNA Kit II (OMEGA) and reverse transcribed into cDNA. The specific primers of BmAtlastin-n (forward 5′ TCTATGGTCGGAACCGATTGT 3′ and reverse 5′ GAACTTGGAACGCTGCCTCA 3′ ), VP39 (forward 5′ CTAATGCCCGTGGGTATGG 3′ and reverse 5′ TTGATGAGGTGGCTGTTGC 3′ ) were used for Q-PCR. The primers of house-keeping gene, ribosomal protein gene (rpl3) (forward 5′ CGGTGTTGTTGGATACATTGAG 3′ and reverse 5′ GCTCATCCTGCCATTTCTTACT 3′ ), Action3 (forward 5′ AACACCCCGTCCTGCTCACTG 3′ and reverse 5′ GGGCGAGACGTGTGATTTCCT 3′ ), were used as internal gene. Q-PCR was carried out in a 15 μ l reaction mixture (1 μ l of cDNA, 0.5 mM of each primer and 2× iTaq TM Universal SYBR Green Supermix (Bio-Rad) in each well of 96-well plate. The reaction conditions were 94 °C for 30 s, followed by 40 cycles at 95 °C for 5 s and 60 °C for 15 s. Then the melt curve was analyzed from 65 °C to 95 °C at 0.5 °C increments of 5 s each. The data were analyzed as described in the statistical analysis section.

Preparation of BmAtlastin-n antibody (anti-BmAtlastin-n).
The primers (forward 5′ cgggatc-cAGTCTCGGTGTGAAGCCAAAGG 3′ and reverse 5′ cggaattcCGCCTTGTTGTGGAGTGAATC 3′ ) were used to amplify fragments from vector pMD19-BmAtlastin-n, and then constructed with pET32a-c (+ ) (Novagen). Vector pET32a-BmAtlastin-n was transformed into BL21 and induced by 0.5 mM IPTG at 37 °C for 5 h. The recombinant protein was purified by Ni affinity chromatography and the products were used for the immunization of New Zealand white rabbits (0.4 mg/each). Immune treatments were performed once per week for a total of four times, and then the serum was collected to prepare antibody of BmAtlastin-n. Western blotting. Cells were lysed with lysis buffer (Beyotime, China) followed by washing twice with PBS and denatured at 100 °C with 5 × SDS-PAGE Loading Buffer (Beyotime). Proteins were subjected to 12% SDS-PAGE and the bands were transferred to PVDF membranes. They were then incubated with primary antibody anti-BmAtlastin-n/anti-VP39/anti-α -Tubulin (Beyotime) (1/5000, 1 h, 25 °C) and secondary antibody HRP conjugated anti-rabbit IgG/HRP conjugated anti-mouse IgG (Beyotime). The final results were analyzed with ECL Western Blotting Detection System (Bio-Rad).
Microinjection and screening. The transgenic vector piggyBac [3 × P3-EGFP, IE1P-BmAtlastin-n] and the helper vector pHA3PIG were co-injected into the silkworm eggs within 4 h of spawning using a microinjector (Eppendorf) 37,38 . Then the G0 silkworms were reared at 25 °C. The G1 silkworms were produced by inbreeding and backcrosses of G0. BmAtlastin-n-OE transgenic silkworms were selected in G1 broods using a fluorescent stereomicroscope (Olympus), which were found to be EGFP positive in their compound eyes.

Oral inoculation and injection of BmNPV.
Fresh mulberry leaves were cut into circles 1 cm in diameter, then each piece was smeared with 1 × 10 6 ODVs. First-day fourth-instar larvae of BmAtlastin-n-OE transgenic silkworm and Dazao control silkworm were fed mulberry leaves with ODVs (one piece/larva). The silkworms, which each ate a whole piece of mulberry leaf, were collected for assessment of the mortality rate. Then a capillary tube was used to inject 1 × 10 6 BV particles/larva in the two lines form valve. Each treatment was performed in triplicate, and each replicate included 100 larvae. Statistical analysis. The student's t-test was used to assess any statistically significant differences between treatments. P-value < 0.05 was considered significant, here indicated with "* " and P-value < 0.01 was considered very significant, here indicated with "* * ". Data from three independent experiments are here presented as means ± SEM.