Cloning and Functional Characterization of Two BTB Genes in the Predatory Mite Metaseiulus occidentalis

Proteins containing the BTB (Bric-à-brac, tramtrack, and Broad Complex) domain typically share low sequence similarities and are involved in a wide range of cellular functions. We previously identified two putative and closely related BTB genes, BTB1 and BTB2, in the genome of the predatory mite Metaseiulus occidentalis. In the current study, full-length BTB1 and BTB2 cDNAs were cloned and sequenced. BTB1 and BTB2 encode proteins of 380 and 401 amino acids, respectively. BTB1 and BTB2 proteins each contain an N-terminal BTB domain and no other identifiable domains. Thus, they belong to a large category of BTB-domain proteins that are widely distributed in eukaryotes, yet with largely unknown function(s). BTB1 and BTB2 gene knockdowns in M. occidentalis females using RNAi reduced their fecundity by approximately 40% and 73%, respectively, whereas knockdown had no impact on their survival or the development of their offspring. These findings suggest these two proteins may be involved in processes related to egg production in this predatory mite, expanding the list of functions attributed to these diverse proteins.


Introduction
The BTB domain (also known as the POZ domain) was originally identified as a motif of approximately 115 amino acids in the Drosophila melanogaster Bric-à-brac (Bab), tramtrack and Broad Complex transcription regulators and in many pox virus zinc finger proteins [1][2][3][4]. Proteins with the BTB domain have been found in plants, animals and viruses and typically share low levels of sequence similarity outside the domain [5].
In mammals, BTB-domain proteins participate in a variety of cellular processes including transcription regulation, cytoskeleton regulation, gating of ion channels and protein ubiquitination/degradation [6][7][8][9][10][11][12][13]. In arthropods, many BTB proteins, including D. melanogaster Fruitless (Fru) and Bab, have roles in various biological processes. The loss of the male-specific Fru protein causes impairment of mating behavior [14][15][16]. Paralogous Bab protein Bab1 and Bab2 are homeotic and morphogenetic regulators in the development of appendages and the abdomen in Drosophila [17]. Moreover, mutations in Bab1 and Bab2 result in defective ovaries, leading to impaired oogenesis [18,19]. Interestingly, both Bab1 and Bab2 are expressed in D. melanogaster adult organs such as the brain and digestive tract [20], suggesting that they participate in biological processes other than development.
The BTB domain is typically found as a single copy close to the N termini of BTB proteins. It is mainly involved in protein-protein interactions with itself (self-oligomerization) or with other proteins [21,22]. Two attributes of BTB proteins may explain why they are involved in so many functions. First, despite sharing similar secondary structures, the BTB domains from different BTB proteins show variability in their primary sequences, especially in peripheral sequences adjacent to the core BTB fold that consists of 95 amino acids [5]. This sequence variability between BTB domains allows BTB proteins to participate in a variety of protein-protein interactions [23][24][25][26][27][28]. Secondly, the presence of other protein motifs can facilitate further function differentiation in BTB proteins. For example, the kelch and BACK motifs are often found in BTB proteins involved in maintaining the stability and dynamics of actin filaments [5,8]. DNA-binding domains, such as zinc finger or helix-turn-helix motifs, are often found in BTB proteins involved in transcriptional regulation. These proteins include many mammalian BTB proteins as well as D. melanogaster Fru and Bab proteins [5,14,29].
Many eukaryotic species contain a large group of BTB proteins with extended regions that consist of no recognizable motifs [5]. In vertebrates such as Homo sapiens, Mus musculus and Danio rerio, approximately 20% of BTB proteins belong to this category. In invertebrates, this category often represents the largest class of BTB proteins in sequenced genomes, encompassing approximately 30%, 49% and 50% of the BTB proteins in D. melanogaster, Anopheles gambiae and Caenorhabditis elegans, respectively [5]. Despite the large number of BTB proteins belonging to this category in various species, little is known about their functional significance.
We recently identified two putative BTB genes, BTB1 and BTB2, in the genome of the predatory mite Metaseiulus occidentalis (= Typhlodromus or Galendromus) occidentalis (Nesbitt) (Arthropoda: Chelicerata: Acari: Phytoseiidae), based on homology searches using D. melanogaster fruitless as query [30]. Metaseiulus occidentalis is an agriculturally important biological control agent of plant-feeding pest mites such as Tetranychus urticae [31][32][33][34]. Previously, pesticide-resistant strains of M. occidentalis were developed through laboratory selection and were applied in biological control programs [35][36][37]. Further genetic improvement can benefit from studies on the molecular components such as BTB1 and BTB2 that may be involved in important biological processes.
Both M. occidentalis BTB1 and BTB2 are expressed in adult females and males, with BTB2 showing a male-biased expression [30]. Interestingly, the predicted BTB1 (380 aa) and BTB2 (401 aa) proteins are much shorter than D. melanogaster Fru (854 aa). It should be noted that neither the BTB1 nor the BTB2 Gnomon gene models, created by the NCBI's Eukaryotic Genome Annotation Pipeline, were validated by RNA seq. The NCBI's annotation pipeline sometimes produces erroneous (e.g. partial) gene models when annotating less-conserved genes [38,39]. Therefore, validation of software-annotated gene models is a prerequisite before commencing functional characterization of a putative gene. In the current study, we cloned and sequenced the full-length cDNAs of BTB1 and BTB2. We then evaluated the domain features and phylogenetic relationship of M. occidentalis BTB1 and BTB2 proteins, and their homologs from selected species. Finally, we investigated the functional roles of these BTB proteins in adult females using an RNAi approach that produces persistent and systemic gene knockdown in M. occidentalis [40].

Colony sources and maintenance
The F10A inbred line was derived from the COS (Carbaryl-OP-Sulfur-resistant) colony [41,42] by sibmating single pairs for 10 generations, as described previously [43]. The Apple Orchard (AO) colony was derived from mites collected from an organic apple orchard in Washington state where no pesticides were applied [44]. Both colonies were maintained and all experiments were performed at 22-23°C and a relative humidity (RH) of 45-55%, under a 16L:8D photoperiod. All stages of T. urticae were brushed on to paraffin-coated construction paper (75 mm × 75 mm) resting on water-soaked cotton to serve as prey.
Age-matched, mated females for the loss-of-function and quantitative reverse transcription-PCR (qRT-PCR) analyses were produced as described previously [40]. Briefly, 20 females of unknown age were collected from either the M. occidentalis F10A or AO colony and placed on pinto bean (Phaseolus vulgaris) leaf discs (40 mm x 60 mm) resting on water-soaked cotton that were infested with approximately 50 T. urticae females to provide prey. The M. occidentalis females, requiring feeding on T. urticae prey to produce eggs [40], were allowed to lay eggs for one day and were then removed. The eggs produced were allowed to hatch and develop. Seven days later, adult females and males emerged and were allowed to mate. Two days later, gravid (mated) females were collected individually and placed on pinto bean leaf discs (15 mm in diameter) that were infested with 4-5 T. urticae females. One day later, the M. occidentalis females produced 1-2 eggs/female, indicating that they had mated and had normally developed ovaries. They were then used for subsequent experiments.

Cloning of the BTB1 and BTB2 genes
To clone the M. occidentalis BTB1 and BTB2 genes, primers were designed to amplify the bulk of BTB1 and BTB2 genes based on their Gnomon gene models (S1 Table). RNA from M. occidental females was extracted using RNAqueous 1 -Micro kit (Part Number Am1931, Life Technologies, CA, USA) according to the manufacturer's instructions. cDNA for cloning was made using the cloned AMV first-strand cDNA synthesis kit (cat. no. 12328-032, Invitrogen, CA, USA) according to the manufacturer's instructions. Nine μl of RNA isolated from a pooled sample (20) of M. occidentalis females were used in a 20-μl reverse transcription reaction containing the manufacturer's recommended ingredients including Oligo(dT) 20 primers. The reaction was performed in a thin-walled tube using a thermocycler (GeneAmp PCR system 9700, Applied Biosystems). The reaction was incubated at 50°C for 60 min, followed by incubation at 85°C for 5 min. PCRs were performed using 1 μl of cDNA using procedures described previously [40].
To obtain the complete cDNA sequence of the BTB1 and BTB2 genes, a new set of gene-specific primers matching the primers in the 5 0 -and 3 0 -Firstchoice RLM-RACE kit (cat. no. AM1700M, Ambion, Grand Island, NY, USA) were designed (S1 Table). 5'-and 3'-RACE were performed according to the manufacturer's instructions. RNA was prepared as described above. cDNA containing appropriate adapters for 5'-and 3'-RACE was prepared as instructed by the Firstchoice RLM-RACE kit manual. Nested PCR amplifications for either 5'-or 3'-RACE were first performed with outer primer sets and then inner primer sets. All PCR reactions (in 50 μl) contained 2 μl of primers, 5 μl of 10X buffer, 4 μl of dNTPs, 0.25 μl of myTaq (cat. no. BIO-21105, Bioline USA, Taunton, MA) and 1 μl of cDNA template. The thermocycler program included the initial denaturation at 94°C for 3 min, followed by 35 cycles of denaturation at 94°C for 30 sec, annealing at 60°C for 30 sec and extension at 72°C for 30 sec. Reactions ended with a final extension of 7 min. All PCR fragments were checked by gel electrophoresis, and then cloned into pCR™2.1 vector (Life Technologies, Grand Island, NY, USA). Sequencing of cloned fragments was performed by the Interdisciplinary Center for Biotechnology Research (ICBR) core laboratories at the University of Florida.

Annotation of the BTB proteins and a phylogenetic analysis
The amino acid sequences of M. occidentalis BTB1 and BTB2 proteins were used as queries to perform BLASTp searches on the GenBank database to retrieve the top hits from Ixodes scapularis, Stegodyphus mimosarum, D. melanogaster, Nasonia vitripennis, Aedes aegypti, Homo sapiens and T. urticae (BOGAS database). The conserved domains of the BTB proteins were identified by searching the conserved domain database of the NCBI (http://www.ncbi.nlm.nih. gov/Structure/cdd/wrpsb.cgi) using default settings [45] with their amino acid sequences. Putative nuclear localization signals in the BTB amino acid sequences were determined using cNLS Mapper (http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi) [46]. An alignment of the deduced amino acid sequences of selected BTB genes was conducted using MAFFT 7.147 [47] with the E-INS-i alignment algorithm and the BLOSUM 62 matrix. Model selection was done with ProtTest 3.2 [48] and according to the Akaike information criterion, the LG+I +G+F model was optimum for phylogenetic analysis. Finally, a maximum likelihood analysis was performed using RAxML 7.3.2 [49], bootstrapping with 1,000 replicates.

dsRNA synthesis and ingestion
Primers with the T7 promoter sequence added at the 5' ends (S1 Table) were used to amplify fragments (~500 bp) of BTB1 and BTB2 genes that were later used for dsRNA synthesis. PCR products were amplified from 1 μl of cDNA prepared as described above. Bands of the expected size (~500 bp) were extracted and purified using a Gel Extraction kit (Qiagen, Valencia, California) according to the manufacturer's protocol. Direct sequencing of the purified PCR products was performed at the ICBR at the University of Florida using the primers used for PCR amplification. dsRNAs were synthesized from 1 μg of purified PCR products or control template (500 bp) provided by the MEGAscript RNAi kit (cat. No. AM1626, Life Technologies, CA, USA) according to the manufacturer's instructions. Sizes of purified dsRNAs were confirmed by gel electrophoresis in 1% agarose gel containing TBE buffer. Concentrations of purified dsRNAs were determined by a spectrophotometer (NanoDrop 1000, Thermo Scientific, USA). Purified dsRNAs were stored in elution buffer at -20°C until further use.
Ingestion of dsRNA was performed as described previously [40,50]. Briefly,~20 agematched, mated F10A females (that had deposited 1-2 eggs each) were starved for 24 h and then placed on a parafilm disc (22 mm in diameter) resting on water-soaked cotton. Ten μl of solution containing 350 ng dsRNA/μl in 20% sucrose (Sigma) and with 3% blue food dye (McCormick, MD, USA) was applied to the parafilm disc. Both sucrose solution and blue food dye were boiled for 10 min to eliminate potential contamination by nucleases before use. F10A females were allowed to feed on the dsRNA/sucrose solution for 48 h.
Ingestion of control dsRNA at 350 ng/μl caused a small, yet significant reduction in the longevity and fecundity in AO females. In contrast, AO females fed on control dsRNA at 100 ng/ μl displayed normal longevity and oviposition phenotypes when compared to TE buffer controls (S2 Table). These results suggest that, probably due to a lack of exposure to pesticides, AO females likely possess lower capacity to metabolize xenobiotics such as dsRNA than F10A females. Therefore, control and BTB1 or BTB2 dsRNAs were used at 100 ng/μl for feeding in subsequent experiments involving AO females. For both strains, only females with blue dye within the gastric caecae were used for further analyses. Fewer than 10% of the females were lost due to runoff and, as expected, no eggs were produced by females during starvation and dsRNA ingestion periods.

Loss-of-function analyses
To measure the effects of BTB1 or BTB2 gene knockdown on viability, oviposition, embryogenesis and offspring development, mated females that had ingested control dsRNA in 20% sucrose or BTB1 or BTB2 dsRNA in 20% sucrose were collected individually and placed, with a male, on a bean leaf disc (22 mm in diameter) on water-soaked cotton that had been infested for 24 h with 10 T. urticae females (in order to produce prey eggs). Ingestion of prey was indicated by an increase in female body size and gradual disappearance of blue color from the gastric caecae [data not shown]. Oviposition started~48 h after the spider mite diet was provided. The eggs produced by the M. occidentalis females were counted and collected every day with a fine sable-hair brush and transferred individually to a new bean leaf disc (22 mm in diameter) containing 4 T. urticae females per M. occidentalis egg. The deposited eggs were allowed to hatch and develop. The day of death for each female was recorded. The numbers and sex of developed offspring were recorded.
For the analyses of the levels of BTB1 or BTB2 gene knockdown, 2 days after being provided with spider mite prey, M. occidentalis females that were fed with control, BTB1 or BTB2 dsRNA were collected and RNA from individual females was extracted. qRT-PCR was performed subsequently to determine levels of BTB1 or BTB2 mRNA.

qRT-PCR analyses
qRT-PCR analysis was performed in a similar manner as described previously [40]. The sequences for the forward and reverse primers used for the qRT-PCR are in S1 Table. The primers were validated using standard curves based on serial dilutions of cDNA to determine the primer annealing efficiencies. One no-template control was included in each experiment to check for possible contamination. qRT-PCR (in technical triplicates) was performed using conditions as previously described [40]. Four biological replicates consisting of 3 females each were performed. All results corresponded to relative quantification using the M. occidentalis actin and GAPDH genes as internal controls using the 2 -ΔΔCt method [51]. Specifically, the Ct for each target (e.g. BTB1 gene) was subtracted by the geometric mean Ct of actin and GAPDH genes from each sample (e.g. control dsRNA or BTB1 dsRNA) to produce ΔCt. The ΔCt from the control was then averaged to produce a mean ΔCt of the control. Then the mean ΔCt of the control was subtracted from individual ΔCt values from control or BTB1 dsRNA treated mites, to yield the ΔΔCt. Then ΔΔCt was used to produce the 2 -ΔΔCt estimates. The specificity of qRT-PCR was confirmed by melting-curve analyses after each reaction. The actin and GAPDH genes were used as reference genes due to their demonstrated stability in M. occidentalis samples [30].

Statistical analyses
The means and standard errors of means (SEM) were analyzed by analysis of variance (ANOVA) (JMP 8; SAS Institute, Cary, NC), and means were separated using Tukey's HSD test (P < 0.05). For one-to-one comparisons, means and SEM were analyzed by ANOVA and means were separated by Student's t test.

Results and Discussion
Cloning of the BTB1 and BTB2 genes A 1,708-bp piece of the BTB1 gene was amplified and cloned to obtain a cDNA fragment. This fragment contains the putative start and stop codons predicted by the BTB1 Gnomon gene model. Additional 5'-UTR and 3'-UTR sequences were obtained by 5'-RACE and 3'-RACE methods, resulting in the identification of a 1,756-bp full-length cDNA (GenBank accession no. XM_003739471.1) of the BTB1 gene. This cDNA molecule contains a 92-bp 5 0 -UTR, 1,143-bp ORF, and 521-bp 3 0 -UTR, and it encodes 380 amino acids. This result is largely in agreement with the BTB1 Gnomon gene model, with two minor discrepancies in the 5'-UTR and 3'-UTR regions in which the Gnomon model predicts an extra 22 and 8 bases at the 5' and 3' ends, respectively.
Similarly, a 1,209-bp piece of the BTB2 gene was amplified and cloned to obtain a cDNA fragment that contains the putative start codon, but not the stop codon predicted by the BTB2 Gnomon gene model. Additional 5'-UTR sequence was obtained by 5'-RACE and the stop codon and 3'-UTR sequence were identified by 3'-RACE, resulting in the identification of a 1,800-bp full-length cDNA (GenBank accession no. XM_003746525.1) of the BTB2 gene. This cDNA molecule contains an 84-bp 5 0 -UTR, 1,206-bp ORF, and 510-bp 3 0 -UTR, and encodes 401 amino acids. This result validates the BTB2 Gnomon gene model in regard to the ORF region. However, there are two discrepancies between the cDNA sequencing result and the Gnomon model prediction. The BTB2 Gnomon model predicted an extra 25 bases at the 5'-UTR region when compared to the cDNA sequence. In addition, no 3'-UTR downstream of the stop codon was predicted by the Gnomon model. In summary, cloning and sequencing of the full-length cDNAs of BTB1 and BTB2 genes validated and refined their Gnomon gene models. Notably, no alternative transcripts for either gene were found.
The transcript structures of the BTB1 and BTB2 genes are shown in Fig 1. BTB1 (on genome scaffold scf7180000076543) and BTB2 (on genome scaffold scf7180000077497) span 2,113 and 2,908 bases of genomic DNA, respectively. Each gene has two introns with intron 1 of both genes located at a homologous position in the 5'-UTR. A BTB domain is located close to the Ntermini of both genes. No other conserved domains were identified.
A comparative study of the BTB1 and BTB2 homologs from selected species To perform a comparative study of the putative homologs of M. occidentalis BTB1 and BTB2 genes, the deduced amino acid sequences of the BTB1 and BTB2 genes were used as queries to perform BLASTp searches against databases of D. melanogaster, A. aegypti, N. vitripennis, T. urticae, I. scapularis, S. mimosarum, and H. sapiens. Notably, the same top hit from each species was retrieved when either BTB1 or BTB2 amino acid sequences were used as BLASTp queries (S3 Table).
Interestingly, bab2, but not fruitless, was the top hit retrieved from D. melanogaster (S3 Table), suggesting that BTB1/BTB2 genes of M. occidentalis are not the bona fide orthologs of fruitless as previously thought [30]. The BTB genes from different species encode amino acid sequences of varying lengths (352-1,066 aa). The deduced M. occidentalis BTB1/BTB2 proteins are similar in length to their putative homologs from T. urticae, I. scapularis, S. mimosarum, N. vitripennis and A. aegypti. Fig 2A shows a comparison of the conserved domains in the BTB proteins. The BTB domains are located in a similar position (near N-termini) in all proteins. As with M. occidentalis BTB1/BTB2 proteins, no other conserved domains were found in other BTB proteins with the exception of D. melanogaster Bab2 and Fru, which possess a helix-turnhelix Psq domain and two zinc fingers, respectively (Fig 2A).
An alignment of the amino acid sequences of the BTB1/BTB2 and their putative homologs from selected species shows that the homologous regions (even between M. occidentalis BTB1 and BTB2) are limited to the BTB domains, with regions beyond the BTB domains showing only low levels of sequence similarity (S1 Fig). All BTB proteins, with the exception of H. sapiens BTB18, contain a putative nuclear localization signal (S1 Fig). Fig 2B shows the result of a phylogenetic analysis of BTB genes. As expected, the BTBs from insects (Arthropoda: Mandibulata) cluster together. Metaseiulus occidentalis BTB1/BTB2 cluster with the BTBs from other chelicerates such as T. urticae, I. scapularis and S. mimosarum.
The significant differences in the domain composition and amino acid sequence length between M. occidentalis BTB1/BTB2 and D. melanogaster bab2 suggest that they may not be true orthologs either. In fact, orthologies among selected insect bab2 homologs are weak, indicated by the varying lengths of their amino acid sequences and the lack of DNA-binding domains in all but Drosophila Bab2 (Fig 2 and S1 Fig). The BTBs in selected chelicerate arthropods are closely clustered and share similarities in both domain composition and protein length. These results suggest that these BTBs likely represent a lineage-specific expansion and that they may encode proteins with similar functions.

Loss-of-function analyses
Oral delivery of BTB1 dsRNA resulted in approximately 77% and 75% reduction in the BTB1 mRNA levels in F10A and AO M. occidentalis females, respectively (Fig 3A). In contrast, BTB1 dsRNA delivery did not affect the BTB2 mRNA levels (Fig 4A), suggesting that the BTB1 gene knockdown was specific. Similar to control females, BTB1 dsRNA-treated females appeared gravid after feeding on spider mite prey [data not shown]. However, egg production was reduced. Each F10A and AO female treated with control dsRNA produced, on average, 24.0 and 22.2 eggs, respectively. In contrast, each F10A and AO female treated with BTB1 dsRNA produced, on average, 14.3 (a 42% reduction) and 13.8 (a 39% reduction) eggs, respectively (Fig 5). The reduction in egg laying occurred in a uniform manner over the life span of the females [data not shown]. The eggs produced by BTB1 dsRNA-treated and control females were morphologically indistinguishable and they all hatched [data not shown], suggesting that embryogenesis was not affected in these eggs. Furthermore, the BTB1 gene knockdown had no effect on the longevity of the females from either colony. Nor did it affect the development or sex ratios of the offspring (Table 1).  An approximately 65% and 60% BTB2 gene knockdown was achieved in F10A and AO females, respectively. In addition, BTB2 dsRNA delivery had no effects on the BTB1 mRNA levels (Fig 4B), suggesting that, similar to the BTB1 gene knockdown, the BTB2 gene knockdown was specific as well. In regard to egg production, each F10A and AO female treated with BTB2 dsRNA produced, on average, 6.5 and 5.6 eggs, respectively. This represents a reduction of 73% and 75% when compared to controls (Fig 5). Furthermore, the reductions in egg production in BTB2 dsRNA-treated F10A and AO females were significantly more severe than those in BTB1 dsRNA-treated females ( Fig 5). As with the BTB1 gene knockdown, the decrease in egg production occurred uniformly over time. The BTB2 gene knockdown had no effect on the morphology of the mothers and the few eggs they produced, or the hatch rates of eggs [data not shown]. Nor did it affect the survival of the females or the development and sex ratios of the few offspring produced (Table 1).
Egg production in phytoseiids is a complex process that involves oogenesis, fertilization, embryogenesis and deposition of eggs. In theory, disruption in any of the aforementioned steps may lead to the reduced egg production observed in BTB1-or BTB2-knockdown females. Notably, the egg production defect seen in BTB1 or BTB2 knockdown females is reminiscent of the phenotype observed in Drosophila strains containing mutant alleles of bab1 and bab2. However, it is unclear what, if any, roles Bab1 and Bab2 may play during egg production in adult Drosophila females because in these mutant strains the reduced oviposition, associated with decreased oogenesis, appeared to have resulted from defective ovary formation during development [17][18][19]. Our gene knockdown experiments, in contrast, used mated M. occidentalis adult females that had produced 1-2 eggs before the start of the experiments, suggesting that they had functioning ovaries capable of producing oocytes that were later fertilized. Thus, our findings indicate that BTB1 and BTB2 proteins likely play a role in some aspect of egg production in M. occidentalis females.
The decrease in the numbers of eggs produced by the females treated with BTB2-dsRNA correlates approximately to the extent of the BTB2-gene knockdown (Figs 3 and 5). In contrast, the BTB1 dsRNA treatment resulted in a less-severe reduction in egg production, despite achieving slightly higher levels of gene knockdown than the BTB2 dsRNA treatment (Figs 3  and 5). These results suggest that BTB1 and BTB2 may be involved in distinct steps of the egg production process in which BTB1 is less critical than BTB2. Alternatively, BTB1 and BTB2 may participate at the same step of the egg production process and the variance in fecundity may be caused by a difference in their protein stability. The presence of putative nuclear localization signals in BTB1 and BTB2 suggests that they could be nuclear proteins and possibly transcription factors similar to Drosophila Bab1 and Bab2. Obviously, further studies are needed to pinpoint the specific processes impaired in the BTB1 and BTB2 dsRNA-treated females, including their cellular and subcellular localization, and the presence of possible DNA-binding motifs in these proteins.
Interestingly, both BTB1 and BTB2 genes are expressed in M. occidentalis males [30]. Considering BTB1/BTB2's apparent significance in female reproduction, it would be interesting to examine any functional significance these two genes may have in males. Similarly, future studies may reveal whether these genes are involved in development like their homologs in Drosophila [17][18][19].
In summary, our study represents the first to illustrate the functional significance of two proteins belonging to a large category of seldom-investigated BTB-domain proteins. We speculate that these BTB proteins are likely involved in diverse and important biological processes, based on the large numbers of proteins in this category and the diversity in their sequences ( Fig   Fig 5. Functional effects of BTB1 and BTB2 dsRNA deliveries. BTB1 and BTB2 dsRNA deliveries significantly reduced egg production of F10A (A) or AO (B) females: One-way ANOVA, F 2,52 = 41.18 and F 2,51 = 48.17 for F10A and AO females, respectively; P < 0.0001 for both F10A and AO females, Tukey-Kramer HSD lettering for all comparisons. N represents the number of biological replicates.  2). Our study highlights the need, as well as the potential rewards, for further research in this class of BTB proteins.
Supporting Information S1 Data. Data used in the current study.
(XLSX) S1 Fig. A multiple sequence alignment of selected BTB genes.
(DOCX) S1 Table. PCR primer sequences used in the current study.