Identi fi cation and characterization of doublesex from the pumpkin fruit fl y , Bactrocera tau ( Diptera : Tephritidae )

The sex determination cascades in insects are diversifi ed at the top of the cascade, where different primary molecular signals are employed, while at the bottom of the cascades, particularly the doublesex genes, are highly conserved. Here, we identifi ed the doublesex ortholog (Btau-dsx) of Bactrocera tau, a pumpkin fruit fl y, and found that Btau-dsx is composed of six exons and fi ve introns with an additional short “m” exon located in the second intron. Btau-dsx is different from its orthologs in most dipteran insects: Its pre-mRNA is sex-specifi cally spliced to yield three (two male and one female) instead of two transcript variants. The two deduced proteins produced by the male-specifi c transcripts are a functional (Btau-DSXM1) and a truncated (Btau-DSXM2) protein, while the female-specifi c transcript produces the functional Btau-DSXF protein. These three proteins contain all conserved domains except Btau-DSXM2 which has no OD2 domain. The female-specifi c transcript is detected in both fertilized and unfertilized eggs and in both somatic and germ cells of the adult females, while the male-specifi c transcript is detected only in fertilized eggs and in the abdominal tissues and testes of adult males. The presence of the Btau-dsxM1 transcript in fertilized eggs at the early syncytium stage suggests that in XY embryos, the Y-linked M factor gene may function quite soon after fertilization to alter the splicing pattern of Btau-dsx pre-mRNA from the female-specifi c to the male-specifi c mode. Injection of Btau-dsxF dsRNA into recently emerging females can reduce the expression of vitellogenin (Btau-Vg) and causes some defects in the ovaries, indicating that Btau-dsxF works upstream of Btau-Vg. * Corresponding author; e-mail: lertlukngernsiri@gmail.com INTRODUCTION The most clearly known mechanism of sex determination among insects is obtained from studies of Drosophila melanogaster, whose sex determination cascade contains four key genes, sex lethal (Sxl), transformer (tra), transformer2 (tra2) and doublesex (dsx). These genes function in a hierarchical order, from the top one to the bottom gene (Cline, 1993; Bopp et al., 2014) . The uppermost gene, Sxl, is ON or OFF depending on the number of X chromosome. In XX fl ies, X-linked signaling elements (XSEs) act as a primary signal to activate Sxl and induce this gene to produce SXL protein. In XY fl ies, the concentration of XSEs is insuffi cient to activate Sxl expression and therefore the SXL protein is not produced. When Sxl protein is present, the downstream gene tra produces the female-specifi c proEur. J. Entomol. 115: 602–613, 2018 doi: 10.14411/eje.2018.057


INTRODUCTION
The most clearly known mechanism of sex determination among insects is obtained from studies of Drosophila melanogaster, whose sex determination cascade contains four key genes, sex lethal (Sxl), transformer (tra), trans-former2 (tra2) and doublesex (dsx).These genes function in a hierarchical order, from the top one to the bottom gene (Cline, 1993;Bopp et al., 2014) .The uppermost gene, Sxl, is ON or OFF depending on the number of X chromosome.In XX fl ies, X-linked signaling elements (XSEs) act as a primary signal to activate Sxl and induce this gene to produce SXL protein.In XY fl ies, the concentration of XSEs is insuffi cient to activate Sxl expression and therefore the SXL protein is not produced.When Sxl protein is present, the downstream gene tra produces the female-specifi c pro-The pumpkin fruit fl y, Bactrocera tau (Walker), is an important pest widely distributed throughout South East Asian countries including Thailand (Chen, 2001).The sex determination genes have been suggested as potential targets for creating genetic sexing strains or for RNAi technology used in the pest management programmes (Dafa'alla et al., 2010;Whyard et al., 2015;Leftwich el al., 2016).In this study, we identify and characterize the dsx homologue of B. tau and fi nd that the structure of Btau-dsx is the same as that of other Bactrocera dsx genes except for the existence of a short "m" exon located in the second intron.Interestingly, we detect the male-specifi c Btau-dsx M1  mRNA in fertilized embryos just after egg-laying.We also examine the function of Btau-dsx in regulating the downstream target gene, vitellogenin, by using adult abdominal double-strand RNA injection.We hope, therefore, that in the future the Btau-dsx gene may be utilized as a means to control the dispersion of this insect species.

Insect culturing
Pumpkin fruit fl ies, Bactrocera tau, were kindly provided by Prof. Dr. Sangvorn Kitthawee, Department of Biology, Faculty of Science, Mahidol University.Flies were cultured on artifi cial food, and kept at 25°C, 70% relative humidity, and with a photoperiod of 12L : 12D.Cucumbers were used for egg collection.

Cloning and sequencing of Btau-dsx
Total RNA was individually extracted from the 15-day-old adult females and males separately using TRIzol reagent (Life Technologies, Carlsbad, CA, USA).One unit of DNase I per mg of RNA was added to the total RNA obtained.One microgram of total RNA was used for synthesis of the fi rst cDNA strand of Btau-dsx using the Revert Aid First Strand cDNA Synthesis Kit (Thermo Fisher Scientifi c, Waltham, MA, USA) and following the manufacture's protocol.Primers were designed based on Bactrocera dsx sequences available in the NCBI database.The fi rst cDNA strand was used as a template to amplify the Btau-dsx gene with primers BtdsxF and BtdsxR (Table 1) under the following conditions: preheating at 94°C for 3 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 1 min: with a fi nal extension at 72°C for 10 min.The PCR product was cloned in pGEM T Easy plasmid vector (Promega, Madison, WI, USA) and sent to Macrogen, Seoul, Korea, for sequencing.The BLAST program (NCBI) was used to check that the sequence obtained from Marcrogen was that of the dsx gene.
To obtain the full-length Btau-dsx cDNA sequence, both 5' and 3' Rapid Amplifi cations of cDNA Ends (RACE) were conducted using the smart cDNA Amplifi cation Kit (Clontech, Mountain View, CA, USA) according to manufacturer's instructions and the gene-specifi c primers, BtdsxR5'RACE for dsx 5' RACE and BtdsxF3'RACE for dsx 3' RACE (Table 1).The nested PCR for the dsx 5' RACE was performed using the same kit and a specifi c primer, BtdsxnestR5'RACE (Table 1).The 5' RACE PCR and 3' RACE PCR fragments were cloned and sequenced at Macrogen.All obtained PCR sequences were assembled to get the full length Btau-dsx cDNA sequence using the CAP3 Sequence Assembly Program (http://pbil.univ-lyon1.fr/cap3.php).

Identifi cation of the Btau-dsx genomic structure
The genomic DNA sequence of B. cucurbitae (NW_011863834.1)available in the NCBI database was used 2005).Instead, the maternal tra genes identifi ed in several tephritid insects are suggested to be the primary signals promoting embryonic sex fate.In XX embryos, functional TRA protein (TRA F ) and TRA2 protein are both maternally provided and form a maternal TRA/TRA2 complex, ensuring via an autoregulation feedback loop that all zygotic tra pre-mRNA is spliced into the female-specifi c mode (Pane et al., 2002;Verhulst et al., 2010).tra F mRNA produces the TRA F protein that, in turn, regulates female-specifi c splicing of dsx in order to produce the female specifi c protein, DSX F .In the presence of DSX F , XX embryos develop as females.In XY embryos, the presence of M factors on Y chromosomes was suggested to turn off the tra autoregulatory feedback loop and, therefore, in the absence of TRA, dsx produces the male-specifi c protein, DSX M , promoting male development of XY embryos (Hoshijima et al., 1991;Lagos et al., 2007;Concha & Scott, 2009;Salvemini et al., 2009;Sarno et al., 2010).
Besides dipteran insects, sex determination has been studied in seven insect orders (Price et al., 2015) including Lepidoptera (Ohbayashi et al., 2001;Shukla & Nagaraju, 2010), Hymenoptera (Cho et al., 2007;Oliveira et al., 2009) and Coleoptera (Shukla & Palli, 2012;Ito et al., 2013).The studies of sex determination in all insect species have thus far revealed that the primary signals initiating the sex determination cascade seem to vary, while genes, such as doublesex, functioning at the bottom of the cascades, are quite conserved.Doublesex (dsx), has been identifi ed and characterized in a number of different insects.In most dipteran insects, dsx genes are transcribed in both sexes but their pre-mRNAs are spliced in a sex-specifi c manner, producing one male-and one female-specifi c transcript that is translated into a DSX M protein in males and a DSX F protein in females (Baker & Wolfner, 1988;Burtis & Baker, 1989;Hediger et al., 2004;Lagos et al., 2005;Chen et al., 2008).DSX proteins are members of the doublesex/mab-3 related (Dmrt) family.Like other proteins belonging to the Dmrt family, DSX proteins have two conserved functional domains, an N-terminal DNA binding domain (DM or OD1 domain) and a C-terminal dimerization domain (DSX dimer or OD2 domain).OD1 is a common domain found in both DSX M and DSX F , while the C-terminal regions of the OD2 domain are different in the two proteins and direct an embryo containing DSX M towards the male sex fate or an embryo containing DSX F towards the female sex fate.
as a template for alignment with all Btau-dsx full-length cDNA sequences using ClustalW2 version 2.0.10 to get the number of exons and introns.

Multiple alignment and phylogenetic analysis
Full length Btau-dsx cDNA sequences were translated into the deduced protein sequences using the Expasy-Translation tool (web.expasy.org/translate/)and were then aligned with a selection of insect Dsx protein sequences (Table S1) available in the NCBI database using clustalW2 version 2.0.10 to highlight sequence identities.Phylogenetic trees were constructed using the neighbour-joining method in MEGA version 7.0.21(Kumar et al., 2016).The reliability of clusters within the tree was evaluated based upon 1000 bootstrap replications.

Expression analysis using RT-PCR
Total RNA was extracted from both fertilized and unfertilized eggs of B. tau at several development stages (0.5-24 h), from whole bodies of adult males and females 2-10 days old, and from some tissues (heads, thoraxes, abdomens, ovaries and testes) of adult males and females at 15, 20 and 25 days post-eclosion, using GenUP TM Total RNA Kit (Biotechrabbit, Hennigsdorf, Germany) following the manufacture's instruction.Three micrograms of total RNA was reverse-transcribed (Reverse transcription System Kit; Promega).Multiplex RT-PCRs were performed using these cDNAs as templates and three primers, BtdsxFF, R-dsxF and R-dsxM (Table 1), for amplifying dsx F and dsx M transcripts.The β-actin gene was used as a reference gene and amplifi ed with primers β-actinF and β-actinR (Table 1).Amplifi cation was carried out in a total volume of 50 μl, including 1 μg of cDNA, 2 μ l of each primer (10 mM), 1 μl of dNTP (10 mM), 0.5 μl of Taq DNA polymerase, and 33.5 μl of distilled water.The PCR reaction was performed at 95°C for 2 min, followed by 30 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 1 min, with a fi nal step of 5 min at 72°C.The PCR products were analyzed on 1% agarose gels.

Preparation of double-stranded RNAs and microinjection procedures
The region common to all dsx genes containing the OD1 domain sequence (752 bp long) was amplifi ed with primers dsx-dsRNAi2-F and dsx-dsRNAi2-R (Table 1) and cloned into a pGEM T Easy plasmid vector.This clone was used as a template for preparing Btau-dsx dsRNAs using a MEGA script T7 kit (Ambion, Foster City, CA, USA), according to the manufacturer's instructions.As a control, 562 bp of the DsRed gene was amplifi ed using primers dsDsRedF and dsDsRedR with the amplifi cation product being used as a template for synthesis the DsRed dsRNA using the same protocol (Table 1).One microgram per microliter of either Btau-dsx dsRNA or DsRed dsRNA was injected into the third abdominal segment of each adult female within 1 h of eclosion using a 10 μl Hamilton microsyringe.Two females were collected on each of the 2 nd , 4 th , 6 th , 8 th and 10 th day post-injection and kept at -80°C for analyzing the expression of Btau-dsx and Btau-Vitellogenin genes.Some remaining females were dissected on the 10 th day post-injection, their ovaries were observed under an Olympus stereomicroscope and photographs were taken with a DP72 Olympus digital camera.

Real-time PCR experiment
Total RNA was separately extracted from whole bodies of all collected Btau-dsx or DsRed dsRNAs injected B. tau females using TRIzol reagent and following the manufacturer's protocol.Three micrograms of total RNA were reverse-transcribed to cDNAs using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientifi c).The expression of Btau-dsx and Btau-Vg genes were examined by qPCR experiments using the Mastercycler® ep realplex (Eppendorf Co., Ltd, Hamburg, Germany).qPCR reactions were performed using RBC ThermOne™ Real-Time Premix (with SYBR Green).The qPCR program for amplifying the cDNAs was one cycle of 94°C for 3 min and 40 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 30 s, followed by melting curve analysis to detect a specific amplification product.The primers used for Real-time PCR were BTdsxF-qRT, BTdsxR-qRT (Table 1) for assaying the transcription level of Btau-dsx F , BTvgF-qRT, BTvgF-qRT and BTvgR-qRT (Table 1) for detecting the Btau-vitellogenin (Btau-Vg) gene and BTactF-qRT and BTactR-qRT (Table 1) for detecting the expression of the β-actin gene.Each sample was analyzed in triplicate and normalized to the β-actin expression of B. tau as an internal control.Real-time data were calculated using the relative expression level by the 2 -∆∆CT method (Livak & Schmittgen, 2001).
Primers for amplifying the Btau-Vg gene were designed using the sequences from a selection of insect vitellogenin genes avail-

Primer name
Sequence primer 3'-5' able on the NCBI database.RT-PCR was performed using the B.tau female cDNA as a template and amplifi ed with the designed primers.The obtained PCR fragment was cloned into a pGEM T easy vector and sent to Macrogen for sequencing.The obtained nucleotide sequence was blasted using the BLAST program (NCBI) to confi rm that it was the sequence of the vitellogenin gene.

Identifi cation of the Btau-dsx gene structure and its transcript variants
The structure of the Btau-dsx gene was identifi ed by aligning the cDNA sequences of all Btau-dsx transcripts with the genomic sequence of the Bactrocera cucurbitae dsx gene (NW_011863834.1).The result showed that Btau-dsx consisted of six exons and fi ve introns, the same genomic structure as that of D. melanogaster and of other Bactrocera insects reported earlier (Fig. 1).The Btau-dsx gene was transcribed in both sexes but its pre-mRNA was spliced in sex-specifi c manners, resulting in three transcripts: one female (Btau-dsx F ) and two male (Btau-dsx M1 and Btau-dsx M2 ).The female transcript consisted of four exons 1, 2, 3 and 4, while the male transcripts harbored fi ve exons, 1, 2, 3, 5 and 6 (Fig. 1).However, only the Btau-dsxM2 transcript contained an additional small exon (127 bp long, called the "m" exon) located within the second intron.Alignment of these two male-specifi c transcripts with the genomic sequence of the B. cucurbitae dsx gene revealed that this gene contained the "m" exon in the second intron.The full length cDNA sequence of Btau-dsx F (accession number: KP844899) was 3477 bp long and contained a 969 bp ORF encoding a deduced protein of 322 amino acid residues.While, the complete cDNA sequences of Btau-dsx M1 (accession number: KP844901) and Btau-dsx M2 (accession number: KP844900) were 3575 and 3436 bp long, respectively.In the Btau-dsx M2 transcript, the 127 bp "m" exon was located between exons 2 and 3, and contained a stop codon (TAA).Thus, the 786 bp long ORF of Btau-dsx M2 was quite short in comparison to Btau-dsx M1  (1206 bp) and encoded a short protein of 261 amino acid residues, while Btau-dsx M1 produced a deduced 401 amino acid functional protein.
The 3' UTR of the Btau-dsx F transcript contained fi ve conserved domains, four putative 13 nt dsx repeat elements (dsxRE) and a purine rich element (PRE), followed by a poly (A) signal.Comparison of the Btau-dsx F mRNA sequence and other Bactrocera dsx F mRNA sequences available in the NCBI database indicated that the dsx F mRNA of Bactrocera species commonly contained four dsxRE elements, with an identical eight nucleotide sequence (CAAT-CAAC) present in all Bacterocera dsxRE elements (Fig. 2).Moreover, the PRE element of Btau-dsx F was composed of 15 nucleotides and 13 of the 15 nucleotides were purines, while other Bactrocera PRE elements contained 12 purine nucleotides (Fig. 2).

Amino acid sequence alignment and phylogenetics
The amino acid sequences of these three Btau-DSX proteins and those of some other insect DSX proteins available in the NCBI database were aligned using the clustalw2 program.The identity of Btau-DSX F to the two male-specifi c DSX proteins, Btau-DSX M1 and Btau-DSX M2 , was 93.48 and 95.40%, respectively, while the identity between Btau-DSX M1 and Btau-DSX M2 was 95.40%.The amino acid sequences of the three Btau-DSX proteins (at amino acid position 1-246) were identical at the N terminal (containing the DNA binding domains, OD1, at position 38-100) until the beginning of the oligomerization (OD2) domains.The three Btau-DSX proteins were different at their C terminals, where the OD2 domains and sex-specifi c regions were located.However, the common region of the OD2 domains  (46 amino acids at position 247-292) of Btau-DSX F and Btau-DSX M1 were also identical, while Btau-DSX M2 lacked the OD2 domain including the sex-specifi c region, indicating that it was a truncated protein.Alignment of these three Btau-DSXs with all other tephritid DSX proteins available in the NCBI database showed a percentage identity ranging from 85.17 to 95.95%, while the identities to some DSX proteins of other insect orders were between 34.21 and 57.99%.Comparison of all Bactrocera DSX proteins available in the NCBI database, including the three Btau-DSX proteins, showed that the amino acid sequences at the N-terminal regions were highly conserved.Furthermore, at the C-terminal regions, if Btau-DSX M2 was excluded, the amino acid sequences of the OD2 domains including the male-specifi c regions were completely identical, while the sequences of female-specifi c regions were also highly conserved (Fig. 3).
Moreover, the OD1 and the OD2 domains of B. tau DSX proteins (Btau-DSX F and Btau-DSX M1 ) were aligned with those of another 16 insect species available in the NCBI database.The alignment showed that at OD1 domains, 20 of 63 amino acid residues were conserved.The 20 conserved amino acids included six essential amino acid residues (C, H, H, C, C and R) necessary for DNA-binding activity in D. melanogaster (Erdman & Burtis, 1993) and three amino acid residues (E, T, Q) specifi c to the insect OD1 domains (Oliveira et al., 2009) (Fig. 4).In contrast, OD2 domains showed high variability, with only two positions out of 47 amino acid residues conserved (Fig. 4).
A phylogenetic tree was constructed using the deduced amino acid sequences of Btau-DSXs (Btau-DSX F , Btau-DSX M1 and Btau-DSX M2 ) and the DSX proteins of other insects available in the NCBI database, by the neighborjoining method replicated 1000 times with bootstrap resampling.The tree topology showed that sequences of the DSX proteins could be used to separate insects at the order and genus levels.Tephritid insects were grouped in the same clade that was, using a classical taxonomy, separated into three subclades, namely Bactrocera, Anastrepha and Ceratitis (Fig. 5).

Analysis of Btau-dsx gene expression patterns using RT-PCR
Expression levels of the three sex-specifi c Btau-dsx transcripts at embryonic stages were assessed, this included both male and female embryos, since they could not be separated from each other.The results showed that fertilized eggs after 0.5-24 h contained both Btau-dsx F (295 bp band) and Btau-dsx M1 (487 bp band) transcripts but the band of Btau-dsx F was stronger than that of Btau-dsx M1 at all stages detected (Fig. 6A).In unfertilized eggs at 0.5, 3, 6 and 12 h after oviposition, only the Btau-dsx F transcript was detected, but this transcript disappeared at 24 h (Fig. 6B).In adult stages, the Btau-dsx F transcript was detected in adult females on each examination day (2, 4, 6, 8 and 10 days post-eclosion).In adult males, the Btaudsx M1 transcript was detected in all examination days (2, 4, 6, 8 and 10 days post-eclosion), while the Btau-dsx M2 transcript was detected only in adult males at 8 and 10 days post-eclosion (Fig. 6C).We found that Btau-dsx expression in some adult tissues of both sexes 15, 20 and 25 days after eclosion gradually increased and that the Btau-dsx F tran-script was detected in all tissues examined (heads, thoraxes, abdomens and ovaries).Expression levels were similar in tissues from fl ies examined on the same day.In male tissues (heads, thoraxes, abdomens and testes), the Btaudsx M1 transcript was strongly detected only in abdomens and testes, while the Btau-dsx M2 transcript was detected at comparatively very low levels (Fig. 6D).

Functional analysis of Btau-dsx F using RNA interference
The Btau-dsx F transcript was only present in female fl ies.In this study, we reduced the quantity of Btau-dsx F transcripts in female fl ies by injecting early emerged females with Btau-dsx dsRNA.Ten days after injection, females were dissected and their ovaries were examined.The results showed that the ovaries of some injected females were mildly deformed: for example, eggs were shorter and their number was reduced, while the ovaries of females injected with DsRed dsRNA were relatively normal and were similar to those of uninjected females (Fig. 7).Moreover, the effi cacy of Btau-dsx dsRNA in the reducing the levels of endogenous Btau-dsx F mRNA was examined by quantitative real-time PCR (qPCR).Samples of females injected with Btau-dsx dsRNA were examined at 2, 4, 6, 8 and 10 days post-eclosion; the result showed that the levels of endogenous Btau-dsx F mRNA were signifi cantly lower than those of females injected with DsRed dsRNA (P < 0.01, Student t-test) (Fig. 8) except at 2 and 10 day post-injection.Furthermore, we also examined the effect of Btau-dsx dsRNA on the expression of the Btau-vitellogenin (Btau-Vg) gene.In oviparous animals including insects, the Vg gene is expressed in females to produce a vitellogenin protein that is accumulated in the yolk of the eggs and is an essential nutrient for embryos during their development.We found that the level of Btau-Vg mRNA in females injected with Btau-dsx dsRNA was dramatically reduced compared to that of females injected with DsRed dsRNA (P < 0.01,  The OD1 domains.Amino acid residues labelled with red are essential for DNA-binding activity in D. melanogaster and the three amino acids labelled with green are specifi c to the insect OD1 domains.(B) The OD2 domains.Amino acid residues labelled with blue are conserved residues found in all insect DSX proteins present here.
Student t-test), indicating that Btau-dsx functions as an upstream gene regulating the expression of Btau-Vg (Fig. 9).

DISCUSSION
In this study, we identifi ed the doublesex (Btau-dsx) gene from the pumpkin fruit fl y, Bactrocera tau (Walker), an invasive insect pest distributed worldwide.The structure and function of Btau-dsx was similar to most dipteran dsx genes, however Btau-dsx showed some different aspects from most dipteran and other Bactrocera dsx genes such as the number of transcript variants, the presence of the "m" exon, and the presence of the male-specifi c transcript at the early syncytial stages of fertilized eggs.

The structure and transcript variants of the B. tau doublesex gene
The structure of the Btau-dsx gene was found to be similar to that of D. melanogaster, M. domestica and Bactrocera insects.Btau-dsx consisted of six exons and fi ve introns and was alternatively spliced in a sex-specifi c manner.The fi rst three exons were common to both male and female transcript variants, while exon 4 was a female-specifi c exon, and exon 5 and 6 were male-specifi c (Burtis & Baker, 1989;Lagos et al., 2005;Saccone et al., 2008).Similar to the dsx genes of M. domestica and L. cuprina, Btau-dsx also contained the 127 bp "m" exon located in the second intron between exon 2 and 3, and after splicing, the "m" exon remained only in one of the two male-specifi c transcripts, Btau-dsx M2 .The "m" exons of M. domestica and L. cuprina dsx genes were found to be 142 and 141 bp long, respectively, and located in the intron between exon 3 and 4 (Hediger et al., 2004;Concha & Scott, 2010).In M. domestica, after processing the dsx pre-mRNA, the "m" exon was present only in the male-specifi c and minor common transcript (Hediger et al., 2004).In fact, in Bactrocera insects, the "m" exon was also present in the second intron of the dsx genomic region of Bactrocera cucurbitae, but was not found in that of Bactrocera oleae.
In D. melanogaster, the 3' UTR of the dsx female-specifi c exon (exon 4) harbored six 13 nt repeat elements (dsxRE) and a purine-rich element (PRE) that were the binding sites for the Tra/Tra2 protein complex.This protein complex is necessary for activation of the female-specifi c mode processing of dsx pre-mRNA (Hoshijima et al., 1991;Inoue et al., 1992).Currently, these conserved domains have been identifi ed in many dipteran dsx female-specifi c transcripts including Btau-dsx F .The Btau-dsx F transcript harbors four 13 nt repeat elements (dsxRE) and a purine-rich element (PRE) at its 3' UTR same as did the 3' UTR of M. domestica and other Bactrocera insects.This suggests that the Btau-dsx female-specifi c splicing may also be regulated by the combined protein products of the two upstream genes, Btau-tra and Btau-tra2, which bind to the dsxREs and the PRE.Moreover, the copy number of the dsxRE domain seems to vary to as much as eight in Lucilia capurina (Concha et al., 2010); six in D. melanogaster (Inoue et al., 1992), Megaselia scalaris (Kuhn et al., 2000) and Anopheles gambiae (Scali et al., 2005); fi ve in Aedes aegypti (Salvemini et al., 2011); four in M. domestica (Hediger et al., 2004), C. ceratitis (Saccone et al., 2008) and Bactrocera species (Shearman & Frommer, 1998;Chen et al., 2008;Permpoon et al., 2011), and three in Anastrepha obligua (Ruiz et al., 2005).However, all Bactrocera dsx female-specifi c transcripts including Btau-dsx F contain four dsxRE copies and one PRE domain.Therefore, the presence of dsxRE and PRE domains found in the Btaudsx F transcript also strongly suggests that the TRA/TRA2 protein complex is required for dsx female-specifi c splicing in B. tau as with most dipteran insects.

Expression of Btau-dsx
In D. melanogaster, the dsx gene is transcribed in both sexes but its pre-mRNA was spliced in sex specifi c manners, producing one dsx male-specifi c mode in males and one female-specifi c mode in females.The dsx male-specific transcript is the default state while the female-specifi c mode is a regulated state since the dsx pre-mRNA spliced into the female-specifi c mode requires the activity of the TRA/TRA2 protein complex (Burtis & Baker, 1989;Hedley & Maniatis, 1991;Hoshijima et al., 1991;Inoue et al., 1992).In this study, both Btau-dsx M1 and Btau-dsx F transcripts were detected in early fertilized eggs, but only Btau-dsx F was detected in unfertilized eggs.The presence in unfertilized eggs of only the Btau-dsx F transcript sug-gests that Btau-dsx pre-mRNA may be a maternal substrate transferred to the eggs during oogenesis.Since we also detected the maternal transformer female-specifi c (Btautra F ) transcript (our unpublished data) and the maternal transformer2 (Btau-tra2) transcript (Thongsaiklaing et al., 2018) in early laid unfertilized and fertilized eggs, it is likely that in B. tau female embryos and in unfertilized eggs, the maternal Btau-dsx pre-mRNAs may be spliced into the female mode upon activation of the maternal Btau-TRA/Btau-TRA2 protein complex.In early fertilized eggs, the Btau-dsx M1 transcript was also detected, albeit at low levels, indicating that the Btau-dsx M1 transcript may exist only in male embryos.In XY fertilized eggs of C. ceratitis, presence of the Ccdsx male-specifi c transcript suggested that the male determining M factor on the Y chromosome could prevent maternal Cctra activity, thus leading to an absence of the CcTRA/CcTRA2 protein complex required for zygotic Cctra pre-mRNA splicing into the female mode via the positive feedback loop.As in D. melanogaster, in the absence of the female-specifi c Cctra F transcript, the CcTRA/CcTRA2 protein complex could not form and thus Ccdsx was spliced by default to produce the CcDSX malespecifi c protein, which in turn induced male differentiation of XY embryos (Hoshijima et al., 1991;Pane et al., 2002;Salvemini et al., 2009).Moreover, the Btau-dsx M1 transcript appeared in early laid eggs (0-0.5 h), much earlier than in embryos of other insects such as D. melanogaster, C. ceratitis and B. jarvisi.In the former two insect species, the dsx male-specifi c transcripts were fi rst detected in male embryos 10 h after oviposition (Hempel & Oliver, 2007;Gabrieli et al., 2010), while in the latter, the dsx M transcript was fi rst detected in male embryos at 6 h after oviposition (Morrow et al., 2014).However, the expression of the dsx male-specifi c transcript (Aadsx M ) in A. aegypti was similar to that of Btau-dsx M1 since it was detected in early laid eggs (0.5 h after oviposition) and in the ovaries of adult females.The explanation for the presence of the Aadsx M transcript in the adult ovaries was that the Aadsx M was not present in female tissues but presented in the fertilized eggs located within the ovaries of fertilized females (Salvemini et al., 2011).The presence of the Btau-dsx M1 transcript in early fertilized eggs suggests that the male determining M factor on the Y chromosome may function at the early syncytial stage  of XY fertilized eggs and that the M factor produced may function as a repressor to prevent the activity of maternal Btau-tra, resulting in the maternal Btau-TRA/Btau-TRA2 protein complex not being formed.Without this protein complex, the splicing pattern of the Btau-dsx pre-mRNA may be switched from the female (alternative) mode to the male (default) mode.In contrast, the Btau-dsx M2 transcript was fi rst detected in adult males eighth days post-eclosion.Since the Btau-dsx M2 transcript encoded a truncated protein without the OD2 domain and sex-specifi c region, the Btau-DSX M2 protein likely has no function, at least not in sex determination.Male dsx transcripts without the OD2 domain were reported in S. coprophila and S. ocellaris but in these two insects, male and female dsx transcripts were present in both sexes (Ruiz et al., 2015).However, Btau-DSX M2 contained the OD1 domain that functions as a DNA binding domain so the presence of the Btau-dsx M2 transcript in the adult stage could mean that this protein might be required for some other unknown function.

Function of the Btau-dsx gene
We examined the function of the Btau-dsx gene by injecting Btau-dsx dsRNA into the ventral side of the abdomens of just-emerged females, since this technique was quite suitable for examining the function and expression of genes, particularly genes expressed in the fat body such as vitellogenin (Vg) genes (Amdam et al., 2003).The effi ciency of this technique was confi rmed in our experiment by comparing the levels of endogenous Btau-dsx F mRNA in females injected with Btau-dsx dsRNA and females injected with DsRed dsRNA.We found that on all days posteclosion, Btau-dsx dsRNA injected females had lower levels of endogenous Btau-dsx mRNA than that of the DsRed dsRNA injected control females, particularly on the fourth day post-injection.
We also examined the level of Vitellogenin (Btau-Vg) mRNA in females injected with either Btau-dsx or DsRed dsRNAs since the expression of Vitellogenin was under the control of the female-specifi c DSX protein (Coschigano & Wensink, 1993;Chen et al., 2008;Shukla & Nagaraju, 2010).Vitellogenin (Vg) is a precursor of the major yolk protein, an essential nutrient for the embryonic development of oviparous animals including insects (Valle, 1993).Vg is synthesized in the fat body and then accumulates in the eggs during oogenesis (Raikhel & Dhadialla, 1992).In other insects, decreased expression of Vg genes affects the development of the oocytes, which in turn affects the morphology of ovaries (Chen et al., 2008;Shukla & Nagaraju, 2010;Veerana et al., 2014).In this study, the ovaries dissected from females injected with Btau-dsx dsRNA showed only mild defects.In B. dorsalis, the ovaries of females injected with 2 μg/μl dsRNA per female were abnormal (Chen et al., 2008) and in Antheraea assama, ovaries dissected from females injected with 70 μg of Aadsx dsRNA per larva were deformed (Shukla & Nagaraju, 2010).Therefore, it is likely that the concentration of Btaudsx dsRNA (1 μg/μl per female) used in this study may not be enough to effi ciently suppress expression of the Btaudsx gene, such insuffi cient suppression would not lower Btau-Vg expression levels under the threshold necessary to cause strong ovary defects.
In conclusion, the present study shows that the structure of the Btau-dsx gene is similar to that of most dipteran insects.In contrast to other tephritid orthologs, Btau-dsx produces three (one female and two male) instead of two (one female and one male) transcript variants.The novel malespecifi c protein (Btau-DSX M2 ) has no OD2 domain.The Btau-dsx M2 is fi rst detected in eight day-old adult males, indicating that Btau-dsx M2 might have a role and is required only during the development of mature adult males, after sex has already been determined by DSX M1 .Indeed, similar to A. aegypti dsx, Btau-dsx M1 is detected at the early syncytium stage, but differs from the dsx genes of most insects previously studied.The knowledge obtained from this study might be useful for developing advanced biotechnological methods to control the spread of B. tau and other insect pests, since sex determination genes can be used to develop transgenic sexing strains for male sterile insect techniques or pest management programmes using RNAi technology (Raphael et al., 2004;Dafa'alla et al., 2010;Fu et al., 2010;Whyard et al., 2015;Leftwich et al., 2016).

Fig. 2 .
Fig. 2. Comparison of the 3' UTR of Bactrocera dsx female-specifi c transcripts show fi ve conserved domains, four dsxRE elements (13 nucleotides) indicated by black lines above the domains and the purine-rich element (PRE) indicated by a dashed line above the corresponding domain, while white letters in black boxes show the identical eight nucleotide (CAATCAAC) sequences present in all Bactrocera dsxRE elements.

Fig. 3 .
Fig. 3. Comparison of the amino acid sequences of BtauDSXM1, BtauDSXM2 and DSXF protein with other Bactroceran DSX proteins available in the NCBI database.(A) The two common regions, DNA binding domain (OD1) and oligomerization domain (OD2) are present in both male and female DSX proteins.The OD1 domain is labeled in gray and indicated by a black line above the domain and the OD2 domain is shaded in gray and indicated by a black dotted line above the domain.Six amino acids labeled with red are conserved to those of D. melanogaster in which they are required for DNA-binding activity and three amino acids labeled with green are specifi cally conserved to all insect OD1 domains.(B) The femalespecifi c region is labeled in gray.(C) The male-specifi c region is labeled in gray.

Fig. 5 .
Fig. 5. Phylogenetic tree constructed using amino acid sequences of DSX proteins of a selection of insects available in the NCBI database using the neighbor-joining method in MEGA 7.0.21.

Fig. 4 .
Fig. 4. Alignment of the OD1 and the OD2 domains of Btau-DSXs and some insect DSX proteins available in the NCBI database.(A)The OD1 domains.Amino acid residues labelled with red are essential for DNA-binding activity in D. melanogaster and the three amino acids labelled with green are specifi c to the insect OD1 domains.(B) The OD2 domains.Amino acid residues labelled with blue are conserved residues found in all insect DSX proteins present here.

Table 1 .
Primers used in this study.

Table S1 .
Doublesex proteins of a selection of insect species available on the NCBI database used for alignment.Supplementary data Spreadsheet containing the raw data obtained from qPCR analysis for Btau-dsx and Btau-Vg expression levels, comparing adult females injected with the Btau-dsx dsRNA to adult females injected with DsRed dsRNA, examined at 2, 4, 6, 8 and 10 days post-injection.The results for detecting the expression of dsx and Vg are displayed in Figs 8 and 9, respectively.