Functional characterization of BmOVOs in silkworm, Bombyx mori

Background In our previous study, we identified four isoforms of the Bmovo gene, Bmovo-1, Bmovo-2, Bmovo-3 and Bmovo-4 from the silkworm ovary and verified that ovarian development was regulated by the BmOVO proteins. Results: To understand the regulatory mechanisms of ovarian development, the regulation of four BmOVO isoforms on the B. mori ovarian tumor (Bmotu) promoter activity was investigated with luciferase reporter assays. The results showed the Bmotu promoter activity was positively regulated by BmOVO-1, BmOVO-2, BmOVO-3 and BmOVO-4 in a dose-dependent manner, of which BmOVO-2 had the highest transcriptional activation. However, the first (A1) and third acidic domains (A3) at the N-terminus of BmOVO-1 are transcriptional repression domains, while the fourth (A4) and fifth acidic domains (A5) are transcriptional activation domains. A recombinant BmOVO zinc-finger domain was found to bind to the GTACCGTTGTA sequence located at the Bmotu promoter. Furthermore, the Bmotu promoter activity was negatively regulated by ‘Tal-like’ peptide, which can trigger BmOVO-1 degradation at the N-terminus. Conclusions These results will help us to further understand the regulatory mechanisms of BmOVO isoforms on Bmotu promoter activity and ovarian development in the silkworm. Electronic supplementary material The online version of this article (10.1186/s12864-019-5697-y) contains supplementary material, which is available to authorized users.


Background
OVO proteins belong to members of the zinc finger protein family and serve as transcription factors that regulate gene expression in various differentiation processes. OVOs not only participate in the development of the neural tube and germ cells, but also play an important role in eye regeneration and the maintenance of the eye in adult planarians [1][2][3][4][5][6]. In Drosophila, the ovo locus encodes three isoforms, Svb, OVO-A and OVO-B [7]. They share four identical C 2 H 2 zinc fingers at their carboxyl (C) termini, which bind to specific DNA sequences [8], but the amino (N) terminus of these isoforms varies. Svb has a role in both larval and adult trichome development and has been detected in the embryonic ventral epidermis [9,10]. OVO-A and OVO-B are only expressed in germ cells and contribute to ovarian development. OVO-A and OVO-B have opposite regulatory activities that are required for female germline development and oogenesis. OVO-A is considered to be a transcriptional repressor, while OVO-B has been identified as an activator [7]. Ovarian tumor (otu), a target gene of OVO proteins, is essential for the viability and differentiation of the female germline in Drosophila [11]. OVO-B positively regulates otu gene expression while OVO-A represses it [8,12]. Moreover, OVO-B also plays an important role in epidermis development when it is expressed ectopic in yeast. [13].
The silkworm (Bombyx mori), a model species of Lepidoptera, has many genes which are homologous to those of Drosophila. In our previous studies, four alternatively spliced isoforms of the B. mori ovo (Bmovo) gene, which have been designated as Bmovo-1, Bmovo-2, Bmovo-3 and Bmovo-4 according to their deduced molecular weights from large to small, were identified in the ovary. Sequence comparisons showed that 203 amino acid residues were conserved at the C-terminus among BmOVO-1, BmOVO-2, BmOVO-3 and BmOVO-4, and four common C 2 H 2 zinc fingers were found at the C-termini of BmOVO-1, BmOVO-2 and BmOVO-3, while only one was found in BmOVO-4, suggesting that BmOVOs may be transcription factors [14]. Various effectors domains at the N-termini of BmOVOs may lead to differences in their functions, and our previous study showed that Bmovo-1 overexpression in silkworm ovaries might promote anabolism for ovarian development [15]. However, the roles of the effector domains of BmOVOs in the regulation of gene expression are still unknown.
Sex determination and differentiation of Bombyx mori has always been one of the important research directions in silkworm industry. In the long-term practice of sericulture, it has been found that the cocoon silk production capacity of male silkworm is generally higher than that of female silkworm, which is generally believed to be due to the fact that male silkworm is strong, does not need to consume extra nutrients for egg making and has high leaf silk conversion rate. B. mori ovarian tumor (Bmotu) is homologous with Drosophila otu, which is essential for female germline differentiation in Drosophila [16]. The Bmotu expression level was up-regulated when Bmovo-1 was overexpressed in silkworm ovaries and down-regulated when Bmovo was silenced [14], suggesting that Bmotu is a target gene, but to date it is still unknown whether BmOVOs can directly bind to the Bmotu promoter to regulate its expression. Therefore, to elucidate the molecular mechanism of Bmovo gene splicing in regulating the expression of otu gene can help us to further understand the regulatory mechanisms of BmOVO on Bmotu promoter activity and ovarian development in the silkworm, thus providing the theoretical basis and molecular target for regulating ovarian development and increasing silkworm silk production through genetic manipulation.
Moreover, in Drosophila, the small peptide Tal can induce selective hydrolysis of the N-terminal transcriptional repression domain of Ovo/Svb, which leads to a change from transcriptional repressor to a transcriptional activator [17,18]. A tal-like gene (NM_0010998471.1) was also found in B. mori, but whether BmOVOs are hydrolyzed by the small peptide encoded by this tal-like gene is unknown. In this research, the regulation of the four BmOVO isoforms on Bmotu promoter activity was investigated with luciferase reporter assays, the binding site of BmOVO to the Bmotu promoter was identified with an electrophoretic mobility shift assay (EMSA) and the BmOVO-1 degradation triggered by the Tal-like small peptide was studied. We found that the BmOVOs are transcriptional activators that directly bind to the Bmotu promoter, and degradation of BmOVO-1 at its N-terminus is mediated by the Tal-like small peptide. These results will help us to further understand the regulatory mechanisms of BmOVO on Bmotu promoter activity and ovarian development in the silkworm.
Transcriptional regulatory activity of BmOVO was regulated by Dpp, Daw, Ror2, STAT and BBx-B8 Previous studies in our laboratory have found that the genes (Dpp, Daw, BBX-B8, Ror2, STAT) are involved in important signaling pathway and associated with silkworm growth and development processes [19][20][21]. To further investigate the transcriptional regulatory function of BmOVO on the Bmotu promoter, pFast-potu-Luc-ie1-Bmovo2 (1 × 10 11 copies) was respectively co-transfected with the pIZT/V5-His, pIZT/ V5-His-STAT, pIZT/V5-His-DPP, pIZT/V5-His-Daw, pIZT/ V5-His-Ror2 and pIZT/V5-His-BBX-B8 plasmids (1 × 10 11 copies). The dual-luciferase reporter assay showed that Bmotu promoter activity clearly decreased in the cells that were co-transfected with pIZT/V5-His-STAT, pIZT/ V5-His-DPP, pIZT/V5-His-Daw, pIZT/V5-His-Ror2 or pIZT/V5-His-BBX-B8 compared with the activity of the cells co-transfected with pIZT/V5-His (Additional file 1: Figure  S2), suggesting that Dpp, Daw, Ror2, STAT and BBx-B8 had antagonistic effects on the transcriptional activation of BmOVO-2. These results provide a clue towards understanding the precise regulation of the target gene expression of BmOVO.  (Fig. 5a). The activity of different truncated Bmotu promoters was investigated with dual-luciferase reporter assays, and the results showed that deletion of the CE1 and CE2 elements didn't have a significant effect on the Bmotu promoter activity regulated by BmOVO-1 (Fig. 5b). When CE1 was deleted, Bmotu promoter activity mediated by BmOVO-2 plummeted, while Bmotu promoter activity did not change significantly with the deletion of both CE1 and CE2 (Fig. 5c); when CE2 was deleted, Bmotu promoter activity regulated by BmOVO-3 decreased, but a change in Bmotu promoter activity was not significant when CE1 was deleted (Fig. 5d). The results of EMSA showed that the C-terminal recombinant protein of BmOVO-2 (400-577 aa) bound to the Bmotu promoter at the CE1 site (probe otu-A) (Fig. 6a), suggesting that BmOVO directly regulates Bmotu gene expression by binding with its promoter. Furthermore, EMSA was carried out with the extracted nucleoproteins from BmN cells, and the  (10 5 ) with plasmids pFast-potu5-Luc (1 × 10 11 copies), pRL-TK (1 × 10 10 copies) and tal-like expression vectors (1 × 10 11 copies) were determined at 60 h post-transfection, 100 μg protein from the lysed cells was used for luciferase assay. The co-transfected BmN cells with pFast-potu5-Luc (1 × 10 11 copies) and pRL-TK (1 × 10 10 copies) plasmids was used as a control. b Plasmids pFast-potu5-Luc-ie1-Bmovo1(1 × 10 11 copies), pRL-TK (1 × 10 10 copies) and tal-like expression vectors (1 × 10 11 copies) were co-transfected into BmN cells (10 5 ), and the co-transfected BmN cells with pFast-potu5 (1 × 10 11 copies), pRL-TK (1 × 10 10 copies) and tal-like expression vectors (1 × 10 11 copies) plasmids was used as a control. Luciferase activities were determined at 60 h post-transfection, 100 μg protein from the lysed cells was used for luciferase assay. The ratio of relative luciferase activity means relative luciferase activity in the experimental group than that in the control group. Tal, a full-length cDNA sequence of tal-like; 5A1-4 + B, a sequence containing 5′-non-coding sequence and it's downstream 1A -4A with B; A1-4 + B, a sequence containing A1-A4 with B; B, a sequence containing B and its downstream non-coding sequence. (**p < 0.01, ***p < 0.001) results showed that the probes otu-A, otu-B and otu-C could also bind with nucleoproteins (Additional file 1: Figure S3a and b), suggesting that Bmotu gene expression could be regulated by other transcription factors. EMSA was carried out with the extracted nucleoproteins and the mutated probe otu-A (otuA-muts), and we found that the second base T mutated to C (T 2 -> C; otuA-mut1) at probe otu-A, causing its ability to bind with the nucleoprotein to be weakened. Additionally, the eleventh base A mutated to G (A 11 -> G; otuA-mut3), leading to enhanced binding capacity, while a clear change was not found when the third base A was mutated to G (A 3 -> G; otuA-mut2) (Additional file 1: Figure S3c). These results indicate that Bmotu gene expression is not only regulated by cis-regulatory elements, but also by the BmOVO transcription factors and other trans-acting regulatory factors. Sequence alignment of the predicted target binding sites (CE1, CE2 and CE3) also showed that three bases (T 2 , A 3 and A 11 ) of the CE1 element might play an important role in BmOVO zinc finger binding to CE1 (Fig. 6b). To confirm this prediction, a dual-luciferase reporter assay was further performed with serial luciferase reporter vectors in which the luc gene was controlled by Bmotu promoters with different mutated CE1 elements. The results showed that when T 2 was mutated to G or C, Bmotu promoter activity regulated by BmOVO-2 was enhanced. However, when T 2 was mutated to A, a significant change in Bmotu promoter activity was not found. When A 3 was mutated to T, G or C, the Bmotu For detection of DsRed, the mouse antibody against RFP was used as a primary antibody, antibody goat anti-mouse HRP used as a secondary antibody. The tubulin was used as internal reference. 30 μg protein from the lysed cells was used for western blotting promoter activity was weakened, while when A 11 was mutated to G, the otu promoter activity increased. Additionally, when T 2 , A 3 and A 11 were respectively deleted, there was no significant change in Bmotu promoter activity; interestingly, Bmotu promoter activity was enhanced with the deletion of all of these bases (T 2 , A 3 and A 11 ) (Fig. 6c). These results indicate that mutation of the CE1 element leads to a change in the advanced structure of the Bmotu promoter, which alters its ability to bind to BmOVO and results in a change in the Bmotu gene expression level.

Discussion
Previous studies have shown that the expression profiles of the Bmovo genes (Bmovo-1, Bmovo-2, Bmovo-3 and Bmovo-4) vary in the gonads of the silkworm at different developmental stages [15]. Bmotu is a target gene of the BmOVO proteins [15] that is essential for female germline differentiation, suggesting that Bmotu gene expression in the ovary of the silkworm at different developmental stages is regulated by different BmOVO isoforms. In the present study, we found that the transcriptional regulatory activity of the four isoforms of BmOVO on Bmotu was different: BmOVO-2, BmOVO-3 and BmOVO-4 are transcriptional activators. However, the positive regulatory effects of BmOVO-1 were not significant in cells transfected with a low dose of the Bmovo-1 expression vector, while the positive regulatory effect was increased in the cells transfected with a high dose of this vector, suggesting that BmOVO-1 is also a transcriptional activator. However, in vitro system cannot fully represent the situation in vivo, only to regulating the expression of endogenous BmOVO can verify the regulatory roles of BmOVO isoforms, which needs to be studied at the individual level. It may be necessary to establish BmOVO transgenic silkworm to solve this problem in the future. In Drosophila, the ovo locus encodes three isoforms, Shaven baby (Svb), OVO-A and OVO-B [7]. OVO-B positively regulates otu gene expression, while OVO-A is a repressor [8,12]. However, we did not find which of the known BmOVO isoforms a transcriptional repressor was. We inferred that the Bmovo locus might also encode a novel BmOVO based on N-terminal domain analysis of the known BmOVO, which may be a transcriptional repressor. Our previous study predicted that BmOVO-2 may be a transcriptional activator [14], because the N-terminal domain of BmOVO-2 contains a serine-rich region (57%), an acidic domain (pI = 3.57) and a glycine-rich region, which also exists in the transcriptional activator OVO-B (56%) [4]. Our results showed that BmOVO-2 significantly increased the activity of the Bmotu promoter, indicating that BmOVO-2 has a positive regulatory effect, which was consistent with the predicted results. BmOVO-3 significantly increased the activity of the Bmotu promoter. The N-terminus of the BmOVO-3 protein has only one acidic domain (A6), so it can be considered that the A6 acidic domain of BmOVO-3 is a transcriptional activation domain. BmOVO belongs to the C 2 H 2 -type zinc finger family, and generally three or more of these zinc fingers are required for binding to target DNA. BmOVO-4 only has one zinc finger structure and no known regulatory domain at its N-terminus, but BmOVO-4 positively regulated Bmotu promoter activity; therefore, we hypothesized that BmOVO-4 may interact with other trans-acting factors to lead to the increased expression of target genes. Usually, transcription factors bind to target DNA sequences in homodimers or heterodimers. It was found that the transcriptional activity of the Bmotu promoter could be improved by the interaction of BmOVO-1 with BmOVO-2, BmOVO-3 and BmOVO-4, the interaction of BmOVO-2 with BmOVO-3 and the interaction of BmOVO-2 with BmOVO-4, suggesting that these four BmOVO isoforms compete for binding sites and cross-regulate the expression of target genes as homodimers and heterodimers.
Our previous study showed that BmOVO-1, BmOVO-2 and BmOVO-4 are mainly expressed in the ovaries, while BmOVO-3 is mainly expressed in the testes [14]; therefore, BmOVO-3 may perform transcriptional regulation as a homodimer. OVO-B and OVO-A are transcription factors with opposing regulatory activities due to different effector domains at their N-termini. In Drosophila OVOs, one glycine-rich region, two acidic regions and an extensive glutamine/histidine-rich region comprise the activation region; a charged basic region and a serine-rich domain comprise the repression region [13]. The distribution and number of acidic and basic domains at the N-termini of BmOVO sequences differed from Drosophila OVO, suggesting differences in function between BmOVO and Drosophila OVO. In this study, we found that the A1 and A3 acidic domains at the N-terminus of BmOVO-1 were transcriptional repression regions, and the A4 and A5 acidic domains were transcriptional activation regions, while the transcriptional regulation of the A2 and B1 domains was not significant. The four BmOVO isoforms have different effector domains at their N-termini, which result in differences in transcriptional regulation activities between the BmOVO isoforms; this result was confirmed by dual-luciferase reporter assays. It has been found that when Drosophila OVO-B was fused with the transcriptional repression domain of Drosophila OVO-A, Drosophila OVO-B was changed from a transcriptional activator into a transcriptional repressor [13]. A similar phenomenon was found in BmN cells, where the A1 acidic domain of BmOVO-1 also turned BmOVO-2 from a transcriptional activator into a transcriptional repressor. This result strongly suggests that A1 is a transcriptional repression domain. Domains containing a high proportion of acidic amino acids generally play a transcriptional activation role [19,22,23], but our study indicated that the A1 and A3 acidic domains in BmOVO-1 did not. Thus, we conjectured that the N-terminal domain of BmOVO-1 changes its transcriptional regulatory activity through its interaction with other proteins and/or a change of the binding capacity of BmOVO-1 to cis-acting factors.
In Drosophila, the small peptide Tal induces selective hydrolysis of the N-terminal transcriptional repression domain of Ovo/Svb mediated by the proteasome, which leads to a switch of Ovo/Svb from a transcriptional repressor to a transcriptional activator [17,18]. The 31 residues of Ovo/Svb at its N-terminus act as a Tal-dependent degradation signal, while three lysines (K3, K8, K23) within this region play a key role in Ovo/ Svb processing [18]. It was found that these 31 residues were highly conserved between BmOVO-1 and Ovo/Svb and the small peptide Tal-like encoded by the silkworm could trigger the hydrolysis of DsRed fused with the N-terminal 28 residues of BmOVO-1. BmOVO-1 can partially rescue Bmotu promoter activity that is repressed by the small peptide Tal-like, suggesting that the transcriptional activation of BmOVO-1 is increased by the hydrolysis of the N-terminal transcriptional repression domain of BmOVO-1, triggered by Tal-like. These results indicate that BmOVO protein levels are not only determined by the transcriptional level of the different isoforms, but also by post-translational degradation.
OVO belongs to a large transcription factor family with conservation of function over evolutionary distance.
A previous study found that OVO transcription factors are present in Drosophila, B. mori, planarians, nematodes, zebrafish and mouse, and they interact with multiple signaling pathways and regulate growth and development [1,3,6]. In this study, we found that both Dpp and Daw, members of the TGF-β family, inhibited the transcriptional activation of BmOVO-2 on the Bmotu promoter. We speculated that over-expression of Dpp and Daw activated the TGF-β signaling pathway [20], and then down-regulated the activity of the Bmotu promoter together with BmOVO. Additionally, overexpression of BBX-B8, Ror2 and STAT also reduced Bmotu promoter activity. The expression of genes related to growth and development could be regulated by over-expression of the insulin-like peptide BBX-B8, an extracellular peptide [21]. Additionally, BmOVO was localized in the nucleus; therefore, the decline in Bmotu promoter activity was not caused by the direct interaction of BBX-B8 with BmOVO. Previous studies have shown that the classical Wnt signaling pathway is mediated by Ror2 in human lung cancer cells [24]. Bmotu promoter activity regulated by BmOVO may have been down-regulated by activating the Wnt signaling pathway mediated by Ror2 in the silkworm. STAT also regulates gene expression by binding to the promoter of target genes [25], which may hamper the binding of BmOVO to the promoter, and results in the decline of Bmotu promoter activity. These results provided new clues to later research.
There are 3 CE elements (CE1, CE2, CE3) on Bmotu promoter predicted by bioinformatics which were the binding site of zinc finger protein. Dual-luciferase reporter assay showed that the activity of Bmotu promoters with diverse deletion of CE elements was different for BmOVO-1, BmOVO-2 and BmOVO-3, suggesting that the relationship of epistatic properties of the CE elements in the Bmotu promoter and the activity of Bmotu promoters is complex. Moreover, we found that BmOVO binds to the CE1 element (GTACCGTTGTA) located on the Bmotu promoter, while the nuclear proteins extracted from BmN cells could bind to the Bmotu promoter at three sites (CE1, GTACCGTTGTA; CE2, AGGCCGTTAAG; CE3, CCTGAACTACA) by EMSA. These results suggest that Bmotu gene expression can be regulated by other transcription factors in addition to BmOVO, but these transcription factors need to be further screened and verified. It was also found that T 2 , A 3 and A 11 in the CE1 sequence might play an important role in binding of BmOVO to the Bmotu promoter. Bmotu promoter activity was altered with the mutation of the CE1 sequence and the deletion of the CE1 and CE2 elements. This may be related to changes in the Bmotu promoter structure. A study investigating the crystal structure of the zinc finger protein-DNA complex showed that the zinc finger was inserted into the large groove of the DNA double helix through its α-helix to combine with DNA [26]. Mutations in the CE1 sequence may change the structure of the DNA double helix, affecting binding of BmOVO to the DNA double helix and resulting in a change of Bmotu promoter activity. These results are helpful to further understand the regulatory mechanisms of BmOVO on Bmotu promoter activity and ovarian development in the silkworm.

Conclusion
In conclusion, the Bmotu promoter activity was positively regulated by four BmOVO isoforms and the BmOVO isoforms could bind to the Bmotu promoter. These results will help us to further understand the regulatory mechanisms of BmOVO on Bmotu promoter activity and ovarian development in the silkworm, thus providing the theoretical basis and molecular target for regulating ovarian development and increasing silkworm silk production through genetic manipulation. In addition, we detected the interaction between the key genes of silkworm important pathway and Bmovo gene, and hope to provide a new clue for the further study of the regulatory role of BmOVO.

Cell culture and transient transfection
The BmN cell line originating from the silkworm ovary was stored in our lab and cultured at 26°C in TC-100 medium (AppliChem, Darmstadt, Germany) containing 10% fetal bovine serum (FBS) (Gibco-BRL, Gaithersburg, Maryland, USA).
Cells were seeded at 1 × 10 5 /well in 24-well culture plates. Expression vectors were transfected into the BmN cells using Lips2000 (Roche, Basel, Switzerland). After 4 h incubation in TC-100 medium without FBS, the medium was replaced with new TC-100 medium with 10% (v/v) FBS, and the cells were incubated for 60 h.
In Drosophila, the small peptide Tal triggers the hydrolysis of the N-terminal transcriptional repression domain of Ovo/Svb, which leads to a change from a transcriptional repressor to a transcriptional activator [17,18]. To investigate whether the degradation of BmOVO-1 is mediated by the small peptide encoded by the tal-like gene in the silkworm, the dsRed reporter gene fused with the N-terminal domain sequence (28 aa) of BmOVO-1 was amplified with the primer pairs ovo1-dsred-F/ovo1-dsred-R (Additional file 1: Table S3), and the obtained fragment was inserted into the pIZT/ V5-His (Invitrogen) to generate the vector pIZT/ V5-His-ovodsRED.

Luciferase reporter assay
Reporter activity was determined using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) according to the manufacturer's instructions. Briefly, BmN cells were lysed for 15 min at room temperature using 1× passive lysis buffer which is made from the dilution of the 5 × passive lysis (Promega, Madison, WI, USA), and then the lysed cells were collected and centrifuged at 12,000 g for 10 min. The supernatant was used for the determination of luciferase activity. Luciferase activity was measured using GloMax Multi Jr. at 490 nm (Promega, Madison, WI, USA). T-test was performed for statistical analysis by the software Graphpad Prism5.

Expression of the BmOVO zinc-finger domain in Escherichia coli
The sequence encoding the last 178 aa containing the four zinc-finger domains of BmOVO was amplified with the primer pairs Bmovo2-3/Bmovo2-2 (Additional file 1: Table  S3) and was then cloned into the pET28a (+) vector (Invitrogen) for the expression of the recombinant BmOVO zinc-finger domain in E. coli. The recombinant protein BmOVO zinc-finger domain with 6-His-tagged was purified with Ni-NTA agarose (Jinyitai, Wuhan, China) according to the manufacturer's instructions. Then, the refolding Primers were showed as → of the BmOVO zinc-finger domain was carried out by dialysis in TGN buffer (50 mM Tris-base, 0.5 mM EDTA, 50 mM NaCl, 1% arginine, 10% glycerol, 5 mM GSSG and 2 mM DTT).

SDS-PAGE and western blotting
30 μg proteins lysed from the transfected cells were subjected to SDS-PAGE, and then transferred to a polyvinylidene fluoride membrane (Roche). The anti-RFP antibody (Abcam, Cambridge, UK) was used, and immunoblotting with anti-β-tubulin antibody (Proteintech, Chicago, IL, USA) was conducted as internal control. The signals were measured using an Enhanced Chemiluminescence (ECL) western blot detection kit (Sangon).