Dermanyssus gallinae and knockdown by RNA84 chemoreception (Barth, 2002; Willemart et al., 2007)....

Comparative morphological and transcriptomic analyses reveal novel chemosensory 1

genes in the poultry red mite, Dermanyssus gallinae and knockdown by RNA 2

interference 3

4

Biswajit Bhowmick1, Yu Tang1, Fang Lin1, Øivind Øines2, Jianguo Zhao1, 5

Chenghong Liao1, Rickard Ignell3, Bill S. Hansson4 and Qian Han1* 6

1Laboratory of Tropical Veterinary Medicine and Vector Biology, School of Life and 7

Pharmaceutical Sciences, Hainan University, Haikou, Hainan 570228, China. 8

2Norwegian Veterinary Institute, Ullevaalsveien 68 P.boks 750 Sentrum, 0106 Oslo, 9

Norway. 10

3Disease Vector Group, Unit of Chemical Ecology, Department of Plant Protection 11

Biology, Swedish University of Agricultural Sciences, Alnarp, Sweden 12

4Department of Evolutionary Neuroethology, Max Planck Institute for Chemical 13

Ecology, 07745, Jena, Germany. 14

*Correspondence: Qian Han Email: [email protected] 15

16

Abstract 17

18

Detection of chemical cues via chemosensory receptor proteins are essential for most 19

animals, and underlies critical behaviors, including location and discrimination of 20

food resources, identification of sexual partners and avoidance of predators. The 21

current knowledge of how chemical cues are detected is based primarily on data 22

acquired from studies on insects, while our understanding of the molecular basis for 23

chemoreception in acari, mites in particular, remains limited. The poultry red mite 24

(PRM), Dermanyssus gallinae, is one of the most important blood-feeding 25

ectoparasites of poultry. Unlike other ectoparasites on animals, PRM feeds mainly at 26

night. During daytime, these animals hide themselves in crevices around the poultry 27

house. The diversity in habitat usage, as well as the demonstrated host finding and 28

was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 11, 2020. . https://doi.org/10.1101/2020.04.09.034587doi: bioRxiv preprint

https://doi.org/10.1101/2020.04.09.034587

avoidance behaviors suggest that PRM relies on their sense of smell to orchestrate 29

complex behavioral decisions. Comparative transcriptome analyses revealed the 30

presence of candidate variant ionotropic receptors (IRs), odorant binding proteins 31

(OBPs), niemann-pick proteins type C2 (NPC2) and sensory neuron membrane 32

proteins (SNMPs). Some of these proteins were highly and differentially expressed in 33

the forelegs of PRM. Rhodopsin-like G protein-coupled receptors (GPCRs) were also 34

identified, while insect-specific odorant receptors (ORs) and odorant co-receptors 35

(ORcos) were not detected. Furthermore, using scanning electron microscopy (SEM), 36

the tarsomeres of all legs pairs were shown to be equipped with sensilla chaetica with 37

or without tip pores, while wall-pored olfactory sensilla chaetica were restricted to 38

the distal-most tarsomeres of the forelegs. Further, using the conserved odorant 39

binding protein (OBP) as a test case, the results showed that RNA interference 40

(RNAi) can be induced in D. gallinae chemosensory tissues. This study is the first to 41

describe the presence of chemosensory genes in any Dermanyssidae family. It is also 42

the first report of chemosensory gene knockdown by RNAi in any mite species and 43

demonstrate that their diminutive size, less than 1 mm, is not a major impediment 44

when applying gene knockdown approaches. Our findings make a significant step 45

forward in understanding the chemosensory abilities of D. gallinae. 46

47

Keywords: Poultry red mite, Dermanyssus gallinae, Sensory morphology, 48

Transcriptome, Chemosensory receptors, Differential gene expression, RNA 49

interference 50

51

1. Introduction 52

53

Mites are highly diverse arthropods of the subphylum chelicerata, one of the most 54

diverse group in animal radiations (Sharma, 2018). These eight legged arthropods 55

exhibit a tremendous variation in lifestyle, ranging from saprophagous to herbivorous 56

and from predatory to parasitic feeding behaviors (Van Leeuwen and Dermauw, 57

2016). Some of these species pose a threat to both animal and human health, such as 58

ticks, mites, spiders and scorpions, and may also vector pathogens of importance to 59


https://doi.org/10.1101/2020.04.09.034587

animals, insects, plants and humans (Javed et al., 2013). The poultry red mite (PRM), 60

Dermanyssus gallinae (de Geer, 1778), is an obligatory hematophagous ectoparasitic 61

mite that infests a wide range of hosts including wild birds, rodents and mammals 62

(George et al., 2015). The PRM is considered the major pest for the poultry industry, 63

with the annual cost in damage and control estimated at €231 million in European 64

countries alone (Flochlay et al., 2017; Bhowmick et al., 2019). Unlike other 65

ectoparasites in animals, PRM sucks blood mainly at night, while they hide 66

themselves in cracks and crevices during daytime so that chickens cannot peck and 67

eat them. Their complex behavior makes them difficult to control with conventional 68

acaricides and other treatment practices (Bhowmick et al., 2019). Hence, an increased 69

understanding of the peripheral olfactory system, including the identification of 70

chemosensory receptors and the ultrastructural characteristics of chemosensory 71

sensilla in PRM, is of great veterinary and economical interest. 72

73

Ticks and mites are two different species, but they are evolutionary related groups of 74

arthropods belonging to the class arachnida and the subclass acari. Unlike insects, 75

olfactory organs in mites and other arachnids are located exclusively on the legs 76

(Carr and Roe, 2016). The four pairs of walking legs are covered with specialized 77

hairs, called sensilla, which protect sensory neurons from the insults of the external 78

environment. Three major types of chemosensory sensilla have so far been 79

characterized in arachnids, including the foreleg tarsi of the mites Varroa destructor 80

(Dillier et al., 2006; Häußermann et al., 2015), D. prognephilus (Davis and Camin, 81

1976) and D. gallinae (Cruz et al., 2005). These unique structures may 82

advantageously allow for the coordination of mechano-, hygro-, thermo and 83

chemoreception (Barth, 2002; Willemart et al., 2007). Even though the identification 84

of structure and the internal arrangement of sensilla is an effective approach to judge 85

their functions (Altner and Prillinger, 1980), very few studies have been conducted in 86

the acarines as a whole, and especially in mites (Tichy and Barth, 1992). Thus, 87

detailed ultrastructural studies using scanning electron microscopy (SEM) would 88

improve our knowledge of poultry red mite sensory biology dramatically. 89

90


https://doi.org/10.1101/2020.04.09.034587

The vast majority of receptors that detect chemosensory stimuli and convert ligand 91

binding into neural activity belong to one of three membrane-bound receptor 92

families: gustatory receptors (GRs), odorant receptors (ORs) and ionotropic receptors 93

(IRs). While insects make use of all of these receptor families, non-insect arthropods 94

rely on GRs and IRs for their chemical communication (Vosshall and Stocker 2007; 95

Benton, R, 2009; Groh-Lunow et al., 2015). The GRs detect non-volatile compounds, 96

including bitter and sweet compounds, salts and some gustatory pheromones 97

(Vosshall and Stocker 2007; Thorne et al., 2004), but also carbon dioxide in various 98

insect species (Croset et al., 2010). The IRs are involved in olfaction and gustation, 99

but also have non-chemosensory functions, including detection of humidity and 100

temperature (Van and Garrity, 2017). Proteins from other gene families have also 101

been shown to contribute to taste and olfaction. Members of the odorant binding 102

protein (OBP) and chemosensory protein (CSPs) families in insects are highly 103

concentrated in the sensillum lymph, and have been shown to bind odorant molecules 104

(Vogt et al., 1991; Sánchez-Gracia et al., 2009). OBP-like proteins have been well 105

characterized in chelicerata, performing roles analogous to those of insect OBPs and 106

CSPs (Renthal et al., 2016; Iovinella et al., 2016). Besides OBPs and CSPs, another 107

potential chemosensory protein family, Niemann Pick type C2 (NPC2), has been 108

identified in both insects and non-insect arthropods (Ishida et al., 2014; Pelosi et al., 109

2014; Iovinella et al., 2016). Other proteins commonly expressed in the 110

chemosensory system include sensory neuron membrane proteins (SNMPs), which 111

are related to scavenger proteins of the CD36 family (Benton et al., 2007), and 112

several classes of G protein-coupled receptors (GPCRs) which allow neurons to sense 113

a variety of extracellular signals, including e.g. hormones and neurotransmitters 114

(Pándy-Szekeres et al., 2018). 115

116

Reverse genetic tools such as RNA interference (RNAi) has been widely used to 117

validate the function of chemosensory genes (Palevich et al., 2018), and will be 118

important to establish in the context of PRM. In PRM, RNAi silencing have 119

previously been used to target gut associated genes for vaccine development (Kamau 120

et al., 2013; Campbell, et al., 2012). However, there are no reports of RNAi targeting 121

chemosensory tissues in any mite species including in this species. In this study, we 122


https://doi.org/10.1101/2020.04.09.034587

make use of a non-invasive immersion-based methodology for achieving RNAi, and 123

to show proof of concept that it is feasible to knock down the expression of a 124

chemoreceptor transcript in D. gallinae olfactory neurons. 125

126

2. Material and methods 127

128

2.1 Mites 129

130

D. Gallinae, the poultry red mite, was a kind gift from Professor Baoliang Pan, China 131

Agricultural University, Beijing, China. Upon arrival, mites were initially maintained 132

at 24±2°C and 65%±4% relative humidity for 24 h, and then kept in vivo using a 133

rearing system by feeding on chickens, as previously described (Wang et al., 2018). 134

The care and use of chickens in this study was approved by Hainan University 135

Institutional Animal Care and Use Committee. 136

137

2.2 Scanning electron microscope (SEM) 138

139

Adult mites and all pairs of legs were used for scanning electron microscopic 140

examination, as previously described (Gainett et al., 2017). Briefly, specimens were 141

cleansed with ultrasonic cleaning solution (Liquinox®) using an ultrasonic cleaner 142

(Digital ultrasonic cleaner®). Samples were subjected to supercritical drying 143

(Quorum® K850) with CO2 after being dehydrated in a graded series of ethanol (from 144

45% to 100%). Dehydrated mites were fixed on stubs with double sided conductive 145

carbon adhesive tabs and then sputter-coated with gold using Anhui Beq SBC-12 146

coating apparatus. Morphological analyses and photographs were done with a ZEISS 147

GeminiSEM (field emission scanning electron microscope) at the 148

Chinese academy of tropical agricultural sciences, Hainan, China. 149

150


https://doi.org/10.1101/2020.04.09.034587

2.3 Total RNA isolation, cDNA library preparation, and high-throughput 151

sequencing 152

153

Total RNA extracted from the forelegs and hindlegs (forelegs consider as a treated 154

group and hindlegs as a control group) of D. gallinae mites were used for RNA 155

sequencing. A total of 150 adult mites (300 forelegs and 300 hindlegs) were dissected 156

by using a scalpel blade (size no. 10) under a dissecting microscope, and three 157

biological replicates per group were generated. After dissections, all samples were 158

rapidly frozen using liquid nitrogen and kept at -80°C until use. RNA was extracted 159

from each group using Trizol® reagent (Thermo-Scientific, Carlsbad, CA, United 160

States), according to the manufacturer’s protocol. The RNA quality was analyzed 161

using an Agilent 2100 Bioanalyzer (Agilent tech., CA, United States), and 162

high-quality samples were subsequently sequenced at the Beijing Genome Institute 163

(BGI-Shenzhen, China), using a BGISEQ-500 platform. For generating first-strand 164

cDNA libraries, RNA was purified using DNA polymerase I and deoxyribonucleotide 165

triphosphate (dNTPs), following RNase H treatment. Then, the cDNA was generated 166

using superscript III reverse transcriptase with random hexamer primer sets. 167

Following sequencing, clean data (clean reads) were obtained by removing those 168

containing adapter or unknown nucleotides at greater than 5% and of low-quality 169

reads from the dataset. All clean reads were converted to the FASTQ file format. 170

Simultaneously, the quality score (Q20 and Q30), GC contents and sequencing 171

coverages were calculated. All of the downstream analyses were based on clean data 172

with high-quality reads. 173

174

2.4 Bioinformatics and differentially expressed genes (DEGs) analysis 175

176

The reference genome of D. gallinae was obtained from NCBI (GAIF00000000) 177

(Schicht et al, 2014) and used for building the index by HISAT2 (v2.0.4). 178

Hierarchical indexing for spliced alignment of transcripts (HISAT) was based on the 179

burrows-wheeler transform and the Ferragina-Manzini (FM) indices, using both 180

genome-wide and local genome index forms (Kim et al., 2015). The generated index 181


https://doi.org/10.1101/2020.04.09.034587

files were used to align the clean reads of the six RNA-seq samples to the reference 182

genome. Bowtie2 (v2.2.5) was used to compare clean reads to reference sequences, 183

and RSEM (v1.2.8) was used to calculate gene expression levels for each sample (Li 184

and Dewey, 2011; Langmead et al., 2012). Gene expression levels were estimated 185

based on the FPKM value (fragments per kilobase of exon per million fragments 186

mapped) using RSEM (RNA-Seq by expectation-maximization) method with default 187

parameters. Differential expression analysis of chemosensory genes in forelegs 188

versus hindlegs were performed by the DEseq2 R package (Wang et al., 2010). The 189

p-values were adjusted using q-values, with q-value ≤ 0.001 and log2 (fold change) 190

≥2 set as the threshold for significantly differential expression using the Benjamini 191

and Hochberg method (Guo et al., 2008; Wang et al., 2010). Differentially expressed 192

genes (DEGs) were cut-off by a false discovery rate (FDR) at 0.05, and then a gene 193

ontology (GO) term enrichments were performed using Blast2GO software (Conesa 194

et al., 2005). To clearly describe the DEGs of each olfactory-related family, we 195

performed cluster analysis using pheatmap package in R software based on 196

FPKM-values, represented as a heatmap. 197

198

2.5 Identification and annotation of chemosensory-related genes 199

200

Chemosensory-related genes from the transcriptomes of D. gallinae were identified 201

by the functional annotation results based on its molecular function, cellular 202

component and biological process, allowing for meta-analyses of gene populations. 203

Standalone tblastn searches (BioEdit software 5.0.6) were performed using 204

previously identified olfactory sequences of IRs, GRs, GPCRs, OBPs, NPC2 and 205

SNMPs from other mite and tick species as queries in order to identify additional 206

candidate chemosensory gene families in the available transcriptomes of D. gallinae 207

(Hall, 1999; Ashburner et al., 2000). All candidate gene families were manually 208

verified using the BLASTx (translated nucleotide to protein) program against 209

non-redundant protein databases available at the NCBI with a cut-off e-value < 10−6. 210

The open reading frames (ORFs) of all candidate genes were determined by using the 211

open reading finder at NCBI. Thereafter, all transcripts with BLAST hits were 212


https://doi.org/10.1101/2020.04.09.034587

included in an InterProScan tool search available at the EMBL-EBI with the 213

following integrated database: ProfileScan, SignalPHMM, Phobius, HAMAP, 214

FPrintScan, PatternScan, HMMPIR, Pfam, HMMPanther, HMMSmart, SuperFamily, 215

patternScan, SFLD, TMHMM, Coils and CDD (Götz et al., 2008). The presence of 216

putative N-terminal signal peptides were detected using the default parameter of 217

SignalP4.0, while transmembrane domains were detected using the default parameter 218

of TMHMM2.0. In addition, olfactory soluble proteins, such as OBP and NPC2 219

protein sequences from the arthropod that include insects, mites, ticks were used for 220

motif-pattern analysis. The parameters setting in this study for motif predictions 221

were: minimum width of motif = 6, maximum width = 10 and maximum number of 222

motif to find = 8. Motif-based sequence analyses were predicted by using web-based 223

version 5.1.0 of the MEME server (Machanick and Bailey, 2011). All candidate 224

chemosensory unigene names were represented as follows: first letter of the genus 225

and species (e.g. Dermanyssus gallinae-Dg) followed by a member of one of these 226

protein families (e.g. IR, NPC2, OBP). 227

228

2.6 Phylogenetic analysis of chemosensory genes 229

230

For the qualitative report of gene family transcripts, the translated ORFs of D. 231

gallinae chemosensory transcripts that contain the relevant conserved domains were 232

retrieved and used to construct phylogenetic analysis. Amino acid sequences for each 233

gene family were aligned using MAFFT in “Auto” strategy method with 234

BLOSUM62 scoring matrix and other default parameters (Katoh et al., 2002). The 235

maximum-likelihood (ML) trees were constructed with PhyML using the best-fit 236

substitution model WAG+G+I as determined by ProtTest 2.4. Branch support was 237

estimated using a fast and accurate maximum likelihood-ratio test (aLRT) and 238

subsequently edited in iTOL webserver (Anisimova and Gascuel, 2006; Letunic and 239

Bork 2007; Guindon et al., 2010). The species maximum-likelihood phylogenetic tree 240

was conducted based on the alignment result of mitochondrial cytochrome c oxidase 241

genes (CO1) from different species using the software in MEGA7.0 (Kumar et al., 242


https://doi.org/10.1101/2020.04.09.034587

2016). The tree image was further viewed and graphically edited using the iTOL 243

online tool with the number of identified chemosensory genes. 244

245

2.7 Quantitative real-time PCR validation for DEGs 246

247

A total of 4 chemosensory-related transcripts identified by RNA-seq to be expressed 248

deferentially between forelegs and hindlegs were selected for real-time quantitative 249

PCR (qPCR) analysis using the LightCycler® (Roche Applied Sciences, Mannheim, 250

Germany). RNA extraction was performed as mentioned above. First-strand cDNA 251

was synthesized with a PrimeScript™ RT synthesis kit with one-step genomic gDNA 252

eraser kit according to instructions (TaKaRa, Tokyo, Japan). Gene-specific primer sets 253

were designed based on the ORF, and Fun14-like gene was used as housekeeping 254

gene (Bartley et al., 2012). Amplification was performed in a final reaction of 50 μl 255

containing 5 μl of sample cDNA, 3 μl of each specific primer (10 mM), 25 μl of 2× 256

SYBR® green master mix (TransGen Biotech, Beijing, China), and sterilized 257

ultra-pure grade ddH2O (14 μl). The thermal cycle conditions were as follows: 96 °C 258

for 8 min, followed by 40 cycles of 96 °C for 5 s, 56 °C for 15 s and 71 °C for 15 s. 259

Following the real-time PCR cycles, the specificity of the SYBR green PCR signal 260

was confirmed by melting curve analysis and agarose gel electrophoresis. Relative 261

quantification data were analyzed using the comparative 2–ΔΔCt method with three 262

biological repeats each containing three technical replicates (Livak and Schmittgen, 263

2001). 264

265

2.8 Preparation of double stranded RNA (dsRNA) 266

267

Total RNA was isolated from ten individual mites using Trizol® reagent and was then 268

treated with a PrimeScript™ RT synthesis kit with genomic gDNA eraser according to 269

instructions (TaKaRa, Beijing, China). A 414 base pair (bp) fragment of the PRM 270

OBP-like gene was amplified from the cDNA using the following the primer sets: 271

(F-5′-ATAAGAATGCGGCCGCATGAAGGCAATCGTGCTTGTG-3′) and 272


https://doi.org/10.1101/2020.04.09.034587

(R-5′-CCGCTCGAGGGCTTACTGTTCCACCTTGCACT-3′) containing EcoRI and 273

XhoI restriction enzyme sites (Thermo-Scientific, Carlsbad, CA, United States). Two 274

control groups were used: blank control (without any dsRNA) and negative control 275

(mites exposed to a non-related dsRNA). The β- glucuronidase (GUS) gene 276

(KY848224) was used as negative control group, as reported (Whyard et al., 2015). 277

After purification of PCR products, the amplified PCR products were cloned into a 278

T/A cloning vector PMD18-T (TaKaRa, Tokyo, Japan), and later excised from 279

PMD18-T plasmid using NotI and XhoI restriction sites, then ligated into a 280

similarly-digested plasmid pL4440, a vector possessing convergent T7 promoters. 281

Positive colonies were transformed into Escherichia coli HT115 (DE3) competent 282

cells, which are RNase-III-deficient strain, and then a single bacterial colony was 283

cultivated in liquid LB medium containing 120 μgml−1 ampicilin and 13.5 μgml−1 284

tetracyclin for 12-15 hours at 370 C at 110 rpm. The culture was then diluted 50× in 285

1× LB broth and grown until OD600 value of 0.4-0.6. Isopropyl 286

β-D-1-thiogalactopyranoside (IPTG) was added in a final concentration of 0.4 mM to 287

induce T7 polymerase activity. The expressed dsRNA was extracted using Trizol 288

reagent, and the quality of dsRNA were analyzed by electrophoresis on 2% agarose 289

gel. The RNA concentration was quantified using a spectrophotometer (Implen 290

NanoPhotometer®N50). dsRNA aliquots were stored at −80 °C until used. 291

292

2.9 Delivery of dsRNA by non-invasive immersion method 293

294

Ten D. gallinae adult mites were immersed in separate sterile Eppendorf tube in 295

35-40 μl dsRNA encoding either DgOBP-1 (target gene) or GUS (negative control), 296

diluted in saline water (0.9% Nacl) to a final concentration of 3.5 μg/μl, with 3 297

biological replicates per treatment and control group, as reported previously (Marr et 298

al., 2015). The blank control group was directly immersed with DEPC-treated water. 299

The treatment and control group were then incubated at 4 °C for 14 hours. 300

301

2.10 Validation of knockdown by real-time quantitative PCR 302


https://doi.org/10.1101/2020.04.09.034587

303

Real-time quantitative reverse transcription (qRT-PCR) assay was performed in 304

triplicates. After 14 h of immersion, RNA extraction and first-strand cDNA synthesis 305

were performed as described above. The cDNA from each replicate treatment was 306

then used to assess the extent of RNAi by measuring levels of gene expression using 307

qPCR. For relative quantification of expression, the Fun14-like gene was used as a 308

housekeeping gene to compare levels of RNAi as described previously. The primer 309

sequences were as follows: Fun14-like (F-5'-TAA CTG GCT CAC CCG AAC TC-3′ 310

and R-5'-CTT TCT CCA CAG CCT TCC AG-3′; product size 208 bp) (Bartley et al., 311

2012). Melt curve analysis was performed and confirmed that only a single product 312

was amplified with each primer pair in every sample. The 2−ΔΔCt method was used to 313

analyze gene expression profiles, comparing expression in specific dsRNA treated 314

samples to gus-dsRNA treated samples (Livak and Schmittgen, 2001) 315

316

2.11 Data analysis 317

318

The data were analyzed using prism 5.01 software (GraphPad Prism, California, 319

USA) and plotted as bar graphs, and the values are represented as means ± standard 320

errors of means (SEM) of three biological replicates. The differences between two 321

groups were analyzed using t-tests, and the differences between multiple groups were 322

analyzed using one-way ANOVA followed by a Tukey's multiple comparison test. 323

Differences were considered to be significant at (*) p values of <0.05. 324

325

326

3. Results 327

328

3.1 Identification of sensilla on the legs of D. gallinae 329

330


https://doi.org/10.1101/2020.04.09.034587

Using scanning electron microscope, the forelegs and hindlegs of D. gallinae were 331

analyzed, demonstrating a higher number of sensillum types on the anterior foreleg 332

tarsi than on the posterior ones (Fig. 1B and C). Sensilla chaetica are classified into 333

different morphological types based on location, size, shaft morphology, structure of 334

socket and the presence and location of pores: sensilla chaetica with or without a tip 335

pore (Sc-tp) and sensilla chaetica with wall pores (Sc-wp). The Sc-tp, which was the 336

longest sensillum type, displayed a regular distribution pattern on the tarsi (Fig. 2A). 337

This sensillum type has a prominent socket and an articulating membrane (Fig. 2B), 338

with the sensillum wall having continuous longitudinal ridges along the shaft, and the 339

shaft being gradually tapered toward the tip (Fig. 2D and E). Even though tip pores 340

on the Sc-tp of the distalmost tarsomeres (DT) I-II were indistinct in the SEM (Fig. 341

2E), tip pores on DT III-IV were delimited by a round margin (Fig. 2F). In contrast to 342

Sc-tp, all (Sc-wp) were restricted to tarsi I (Fig. 3A), and their shafts were inserted in 343

a shallow membranous cuticle with an incomplete articulation (non-socketed) (Fig. 344

3B). The shaft tapered toward the tip (Fig. 3C), and the thick wall contained a large 345

number of evenly distributed pores (Fig. 3E). 346

347

3.2 Mapping and transcriptome analysis 348

349

To identify olfactory genes, RNA extracts of PRM forelegs and hindlegs were 350

sequenced separately using a BGISEQ-500 HiSeq platform. On average, 22.5 and 351

21.7 million raw reads were generated for the foreleg and hindleg samples, 352

respectively. After quality control screening, 69.42% of the reads per sample mapped 353

to the D. gallinae reference genome (Table S1). While the resulting 25,604 sequences 354

were identified as unigenes, these might not necessarily represent unique genes. The 355

read sequences from the six transcriptomes were submitted to the sequence read 356

archive (SRA) at NCBI under the accession number of PRJNA602095. 357

358

3.3 Homology and functional annotation 359

360


https://doi.org/10.1101/2020.04.09.034587

Of the 25,604 unigenes, 14,203 (55.47%) showed significant similarity to genes 361

encoding for known proteins in the NCBI non-redundant protein database (NR), 362

whereas 10,285 (40.16%) and 5,464 (21.34%) were identified following KEGG and 363

GO annotation, respectively. According to the KEGG annotation results, signal 364

transduction (environmental information processing) was the most highly expressed 365

categories (Figure S1). GO annotation was used to categorize annotated genes into 366

functional groups based on three main categories: molecular functions, cellular 367

components and biological process. Among the biological process terms, the most 368

represented biological processes were cellular (1,234 unigenes) and metabolic 369

processes (9,54 unigenes). In the cellular component terms, the genes expressed were 370

predominantly cell (1,309 unigenes), membrane (1,285 unigenes), and membrane 371

part (1,221 unigenes). In the molecular function category, binding (1,477 unigenes) 372

and catalytic activity (1,390 unigenes) had a huge preponderance and were the most 373

highly expressed categories (Fig. 4). In general, GO term enrichment analysis 374

indicated that forelegs were mostly associated with binding, catalytic activity, signal 375

transducer and transporter activity in the molecular function category. 376

377

3.4 No gustatory receptors (GRs), olfactory receptors (ORs) found in D. gallinae 378

transcripts 379

380

BLASTx and BLASTn searches did not identify any transcripts putatively encoding 381

for GRs and Ors in the Dermanyssus transcriptome. To further validate these 382

findings, tBLASTn (e-value ≤ 1) and protein domain based searches did not identify 383

any matches. 384

385

3.5 Identification of candidate IR family 386

387

Eleven transcripts encoding IR/iGluR homologs were identified, five of which 388

contained the specific domain signature of the ionotropic glutamate receptors (Pfam 389

domain-PF00060). Out of the eleven transcripts, only one transcript (DgIr3943) 390


https://doi.org/10.1101/2020.04.09.034587

encoded all of the characteristic domains of the IR/iGluR proteins, namely, the 391

amino-terminal domain (Pfam domain-PF01094), the ligand-binding domain (Pfam 392

domain-PF10613) and the ion channel domain (Pfam domain-PF00060).The protein 393

domain organization of iGluRs/IRs in PRM is shown in histogram (Fig. 5a). Among 394

the IRs, all sequences showed complete ORFs, with >112 amino acids. However, 395

read counts for the identified IR/iGluR transcripts were low in both forelegs and 396

hindlegs: in the forelegs, IR transcript expression levels ranged from 0.06 to 0.13 397

FPKM, and in the hindlegs from 0.01 to 1.12 FPKM. None of the identified IR/iGluR 398

transcripts were significantly found in the forelegs (Fig. 5B). To further distinguish 399

putative IR gene families from the transcriptome of D. gallinae, all of the IRs in our 400

transcriptomes were aligned with IRs from Varroa destructor, Tetranychus urticae, 401

Galendromus occidentalis, Ixodes scapularis, Tropilaelaps mercedesae and 402

Drosophila melanogaster by MAFFT (Table S2). The phylogenetic analysis 403

demonstrated that eight genes were clustered with the iGluR sub-group, and one gene 404

was close to the NMDA sub-group. Apart from that, two Dermanyssus transcripts 405

formed a distinct sub-group together with a sequence of Varroa mites, suggesting that 406

a mite-specific IR sub-group is present. The IR phylogeny did not reveal any 407

potential orthologous relationships with insect IRs, for example there were no 408

relatives of the divergent class of insect IRs (Fig. 6). 409

410

411

3.6 Identification of putative OBP-like and NPC2 proteins in Dermanyssus 412

sequences 413

414

415

A total of five and six candidate OBP- and NPC2-encoding genes were identified, 416

respectively, in the transcriptomic analysis of Dermanyssus chemosensory tissues. 417

These genes were similar to those reported for other mite species (Eliash et al., 2017). 418

The complete ORFs were identified for five OBP candidate genes, revealing >137 419

amino acids encoding for proteins with high homology to known OBP-like proteins 420

from V. destructor (XP_022653281.1; XP_022653293.1; XP_022645714.1; 421


https://doi.org/10.1101/2020.04.09.034587

XP_022666940.1; XP_022672530.1), Amblyomma americanum (JZ172282.1) and I. 422

scapularis (XP_002433530.1). SignalP revealed an 18-amino-acid-long signal peptide 423

cleaved between position 18 and 19 (data not shown). Alignment of amino acid 424

sequences from these species revealed that all OBP-like sequences contain six 425

conserved cysteine (C) residues, with the exact same spacing pattern of all cysteines 426

“C1-X16-C2-X18-C3-X37-C4-X15-C5-X9-C6” (where X represents any amino acid; 427

prosite pattern notation, http://prosite.expasy.org/prosuser.html#conv_pa) (Fig. 7a). 428

429

Two OBP-like genes (DgOBP-1 and DgOBP-2) revealed higher levels of expression 430

in forelegs, whereas three genes (DgOBP-3, DgOBP-4 and DgOBP-5) had 431

expression in both forelegs and hindlegs. Gene expression levels of all OBP-like 432

proteins identified from the forelegs and hindlegs transcriptomes were compared 433

(Fig. 8b). The phylogenetic analysis showed that all of Dermanyssus OBPs were 434

clustered together on the same subclade with high bootstrap support. In addition, 435

Dermanyssus OBPs were clustered with a clade containing OBPs of the mites V. 436

destructor and T. mercedesae, with more than 80% branch support (Fig. 8a and Table 437

S3). Conserved OBP motifs were predicted to better understand the protein’s 438

evolution and function. Since a high number of OBP genes have been reported in 439

both insects and arachnids, we performed a motif-pattern analysis between these two 440

classes (Table S4). Only motif 1 (CMDYHJSQIC) and motif 2 (TCALKSEGWF) 441

encoded the OBP family were present in all the OBP orthologs except for 442

Trichomalopsis sarcophagae (Ts2), D. willistoni (Dw) and D. ficusphila (Df) (Fig. 9). 443

Since motif 1 and 2 were found in all OBP proteins including the five Dermanyssus 444

OBPs, the results provide confidence of their identification as bonafide OBP-like 445

encoding genes and infer a functional similarity with other OBPs. 446

447

Six NPC2 proteins were annotated from the D. gallinae transcriptomes. These 448

putative NPC2 genes consisted of complete ORFs with lengths ranging from 471 bp 449

to 591 bp nucleotides, and all sequences had an N-terminal signal peptide that are a 450

common feature of the secretory proteins. To compare sequence similarity with other 451

acari, a multiple sequence alignment showed that all six cysteines were conserved in 452


https://doi.org/10.1101/2020.04.09.034587

all NPC2 proteins (Fig. 7b). Phylogenetic analysis of the NPC2 genes revealed three 453

groups of D. gallinae NPC2 genes, with two groups being unique to arachnida (Fig. 454

10a and Table S5). The differential expression of the chemosensory-related proteins 455

between the forelegs and hindlegs comparisons was less prevalent. Only a single 456

NPC2 transcript, with a low abundance, demonstrated a foreleg-biased expression 457

(Fig. 10b). 458

459

The conserved motifs are important elements of functional domains. A total of 32 460

sequences from both insects and arachnids including six Dermanyssus NPC2 were 461

used to compare and search for shared motif patterns (Table S6). The results showed 462

that motif 1 (CPLKKGKDYT) and motif 2 (PFPGPKSDAC) were present in all the 463

NPC2 orthologs except for I. scapularis (Isc1), Tropilaelaps mercedesae (Tm3,1), D. 464

melanogaster (Dm2,4), D. gallinae (DgNPC2) and V. destructor (Vd4) (Fig.11). In 465

addition, the homologous NPC2 from different species had similar motif patterns; for 466

example, D. gallinae (DgNPC2-1, 2 and 3), V. destructor (vd2) and T. mercedesae 467

(Tm2) had same motif-pattern as 4-7-2-8-1-6-3. 468

469

470

3.7 Identification of candidate sensory neuron membrane proteins (SNMPs) 471

472

Five candidate SNMPs transcripts were identified based on similarity to known 473

SNMPs, within the CD36 superfamily (PF01130.21), of mites, ticks and insects. 474

None of the identified transcripts were significantly found in the forelegs (Fig. 12b). 475

To extend and specify this classification, the constructed phylogenetic tree revealed 476

that acarian SNMP sequences were clearly separated from those of insects with the 477

exception of three SNMPs (TrpWZC0, Var27012 and IscPD82) that showed 478

proximity to the D. melanogaster clusters (Fig. 12a and Table S7). 479

480

3.8 Identification of candidate G protein-coupled receptors (GPCRs) 481


https://doi.org/10.1101/2020.04.09.034587

482

One putative GPCR was identified exclusively in the transcriptome of forelegs and 483

contained a rhodopsin-like domain (PF00001.21). Protein domain and phylogenetic 484

analyses of the putative GPCR transcript revealed that the transcript belongs to the 485

putative clade A, rhodopsin-like GPCRs showing GPCR and photoreceptor activity. 486

In insects, GPCRs are divided into three major classes, rhodopsin (clade A), secretin 487

(clade B), metabotropic glutamate (clade C) and atypical (clade D) (Carr et al., 2017) 488

(Fig. 13 and Table S8). We also found foreleg-biased expression of a G protein, and 489

expression of this was verified by RT-PCR (qPCR) (Fig. 14). 490

491

3.9 Differentially expressed gene (DEG) validation using qRT-PCR 492

493

In order to verify the differentially expressed genes identified through the 494

transcriptome analysis, the expression pattern of identified gene were analyzed using 495

qRT-PCR, including 2 OBPs (DgOBP-1, DgOBP-2), 1 NPC2 (DgNPC2-5) and 1 496

GPCR (DgGPCR-1). The expression levels were largely consistent with the 497

transcriptome profile analyses concluding that the RNA-seq data were reliable (Fig. 498

14). The primers used for real-time PCR are listed in Table S9. 499

500

501

3.10 The efficiency of RNAi in PRM 502

503

DgOBP-1 was used as a test gene to assess the feasibility of using a dsRNAi gene 504

knockdown approach to knock down gene expression in D. gallinae. As shown in 505

Fig. 15, mites immersed in OBP dsRNA for 14 hours exhibited significant gene 506

knockdown when compared to controls. Mites that were immersed in blank or 507

negative control, exhibited no significant differences in OBP mRNA levels, as 508

determined by qRT-PCR. The primers used for RNAi and real-time PCR are listed in 509

Table S10. 510


https://doi.org/10.1101/2020.04.09.034587

511

4. Discussion 512

513

Compared to other arthropod in general and to ticks in particular, the molecular basis 514

of chemoreception in mites is largely unexplored. Most arachnids including PRM are 515

active at night and thus do not primarily depend on visual sense. Their most 516

important sensory input comes from mechanical stimuli such as air currents, touch 517

and vibrations. To better understand how mites perceive olfactory chemical cues, 518

chemosensory genes were identified using transcriptomic analysis of D. gallinae, 519

along with an ultrastructural study of tarsal sensilla. Our results suggest that the 520

forelegs are equipped with sensilla containing neurons with chemo- and 521

mechanosensory related functions. As a means to gain insight into the function of the 522

chemosensory system of the PRM, an RNAi silencing approach was developed. 523

There, we demonstrate a clear knock-down in the relative expression of the OBP-1 524

gene in dsRNA treated mites. In general, an improved understanding of the olfactory 525

system may provide a good start of controlling this problematic mite. 526

527

528

Three types of setiform sensilla were distinguished, which is typical for all acari: no 529

pore (np), wall-pored (wp) and tip-pored (tp) sensilla. Although each sensillum type 530

has a characteristic ultrastructure, functional properties can only be established with 531

certainty by data from physiological recordings, which we so far lack. Wall-pore 532

sensilla are, in general, considered to have an olfactory function (Altner et al., 1977; 533

Altner and Prillinger, 1980). The exclusive presence of wp sensilla on the distal-most 534

tarsomeres of the front legs indicate that these are important sensory (Fig. 3) 535

appendages also present in ticks (Haller’s organ) and other mite species (Eliash et al., 536

2017; Carr et al., 2017; Lei et al., 2019). Moreover, the presence of tp sensilla 537

suggests a contact-chemoreceptive (gustatory function) and 538

contact-mechanoreceptive function (Fig. 2), as previously described in other 539

arachnids. In contrast to wp and tp sensilla, np sensilla are considered to be 540

mechanosensory function to sense a mechanical distortion of the exoskeleton or seta 541


https://doi.org/10.1101/2020.04.09.034587

(Fig. 2C) (Altner and Prillinger, 1980).These results are consistent with ticks and 542

other types of arachnids (Gainett et al., 2017; Lei et al., 2019). 543

544

In forelegs, four molecular function categories might be connected to olfaction 545

(odorant binding), enzyme activity (catalytic) and signal transduction (signal 546

transducer activity) (Fig. 4). These results are likely to reflect that forelegs play an 547

important chemosensory role. Nevertheless, GO annotations are common to all 548

tissues types that do not provide any information on the expression levels of the 549

individual gene types. Our study did not identify any transcripts putatively encoding 550

gustatory receptors, despite well-documented morphological evidence of 551

gustatory-like tp sensilla (Fig. 2e). This result might reflect that the transcriptome 552

sequencing depth, method and number of samples were not sufficient, and therefore, 553

further research is needed to validate this result. Another striking difference was the 554

small number of chemosenory genes in PRM compared with other homologous 555

species (Fig. 16), which possibly indicates differences in their ability to discriminate 556

volatile chemical compounds and communicate with conspecifics. This difference 557

may have evolved as a result of very different nesting behaviours, feeding habits and 558

host-specificity of the poultry red mite. For example, most of endoparasitic mite 559

species do not actively seek out their hosts. Transmission occurs through close or direct 560

physical contact. The loss of active host-seeking behavior might thus have resulted in a 561

reduction both in number of genes involved in chemosensation and in sensillum 562

numbers. Both commonly observed in species that transition from ectoparasites to 563

endoparasites. 564

565

OBPs have recently been reported in some chelicerata, demonstrating that these 566

proteins are structurally reminiscent of insect odorant binding proteins, and therefore 567

named OBP-like. These proteins were first described in the tick Amblyomma 568

americanum and the honey-bee mite V. destructor, suggesting a likely involvement in 569

acari chemodetection (Renthal et al., 2017; Eliash et al., 2017). These secreted 570

proteins have six cysteine amino acid residues as is also observed in the classical 571

OBP family. The arrangement pattern of six conserved cysteines in the D. gallinae 572


https://doi.org/10.1101/2020.04.09.034587

OBP family is comparable to the patterns in other OBP-like proteins of chelicerata 573

(Fig. 7a). A major finding of our study is the identification of transcripts encoding 574

OBP-like soluble proteins, some of which were highly and differentially expressed in 575

the forelegs. This was also evident in I. scapularis ticks, where some of the OBP-like 576

transcripts were highly expressed in the forelegs (Josek et al., 2018). Phylogenetic 577

analysis revealed that all Dermanyssus OBPs were clustered within arachnida only, 578

indicating that this branch evolved after the splitting of insects and arachnids. The 579

continued annotation of OBP genes from chelicerate genomes will likely clarify this 580

issue. In addition, soluble proteins of a different class (NPC2), which was also 581

detected in the varroa proteomic project, have been suggested to function as 582

semiochemical transporters in chelicerates (Pelosi et al., 2014; Renthal et al. 2017). 583

This study found that NPC2 was expressed not only in chemosensory sensilla, but 584

also in other parts of the body, suggesting roles in both chemical detection as well as 585

in other functions. The results of this study were consistent with other studies in 586

arthropods, which showed wide expression of NPC2 proteins across the body (Ishida 587

et al., 2014; Pelosi et al., 2017; Zhu et al., 2018). 588

589

590

In the transcriptome analysis we also identified sensory neuron membrane proteins 591

(SNMPs), which have key functions in pheromone detection in insects. The role of 592

SNMPs in chemoreception in acari is still unknown. Similarly, there is little 593

information available on the identified GPCRs, and the molecular characterization of 594

most GPCRs remains unknown in acari. Identification of G proteins are also relevant 595

for the understanding of acari biology in general, as these proteins could be targeted 596

by novel compounds. Most information on acari GPCRs were obtained from tick 597

transcriptome datasets and recombinant receptor systems (Gross et al., 2015; Xiong 598

et al, 2019). Previous experiments in ticks indicated that prediction of GPCRs using 599

transcriptome analysis might be a challenging task. As such, the transcriptome study of 600

the bovine tick Boophilus microplus did not allow to detect the kinin receptor, a 601

neuropeptide GPCR (Guerrero et al., 2016), which should be present (Holmes et al., 602

2000), underscoring the often low expression of GPCRs as a challenge for detection. 603

Likewise, only two G proteins were identified in I. scapularis ticks and all were 604


https://doi.org/10.1101/2020.04.09.034587

expressed at low to non-existent levels in legs (Josek et al., 2018). All of these 605

findings acknowledge limitations in the prediction of tick GPCRs. In agreement with 606

this, only one GPCR transcripts was detected in the present study. In addition, the 607

presence of putative rhodopsin-like GPCR transcripts, having a foreleg bias 608

expression in Dermanyssus mite, might code for olfaction. A similar finding was 609

observed in the American dog tick Dermacentor variabilis with foreleg-biased 610

expression of GPCRs, suggesting that tick olfaction via the Haller’s organ involves a 611

GPCR-mediated pathway like that of vertebrates and nematodes (Carr et al., 2017). 612

We thus suggest that mite olfactory receptors might represent a completely novel type 613

of 7-transmembrane receptor (7TM) family proteins that have yet to be identified. 614

615

After identification of putative genes of interest, functional genomics tools such as 616

RNA interference (RNAi) or CRISPR can be used to assess gene function. The RNAi 617

approach has been employed widely to demonstrate knockdown of target genes, 618

which subsequently validate the function of genes in order to identify novel candidate 619

genes for drug or vaccine development (Palevich et al., 2018). This functional 620

genomic approach can be applicable to other parasites, such as in mites where 621

annotated functional genomic information is still limited (Gu and Knipple, 2013; Marr 622

et al., 2014). A typical RNAi can be induced by a variety of delivery systems 623

including noninvasive immersion, electroporation, micro-injection and feeding. The 624

micro-injection alternative has widely been used for research in insects and ticks in 625

spite of having high mortality due to trauma. However, the small size of mites (av. 626

900 × 510 μm) prevent micro-injection of dsRNA from being practical. Our present 627

study demonstrated that a less invasive technique, using the non-invasive immersion 628

method, can be used and is simple and less labor-intensive. Using the conserved 629

odorant binding protein as a test case for RNA interference study, a significant 630

reduction of gene transcript was achieved following 14 hours immersion in double 631

stranded RNA encoding the OBP-1 gene. For efficient silencing, the dsRNA dosage 632

and incubation at 4°C were chosen based on previous experiments (Marr et al., 633

2015), but this should be optimized for each future experiment. The development of 634

gene silencing in PRM could be considered as a method of developing protocols for 635

RNAi in other astigmatid mites of veterinary and medical importance. 636


https://doi.org/10.1101/2020.04.09.034587

637

6. Conclusions 638

639

The chemosensory genes reported here represent significant contribution to the pool 640

of identified olfactory genes in non-insect arthropods in general. Mites consistently 641

move their first pair of legs in the air in the manner of the antennae of insects, along 642

with ultrastructural features of the sensory organs, and the results of transcriptomic 643

analyses are all strongly in favour of the assumption that the forelegs serve an 644

olfactory function. These results highlight the importance of an integrative approach 645

in systematic studies, combining morphological and molecular characteristics, 646

previously not being conducted in other mite and tick species. Further, the ability to 647

induce RNAi through a simple non-invasive immersion-based dsRNA delivery in D. 648

gallinae will allow researchers to study olfactory function in detail. Further work on 649

other relevant genes needs to be done to verify these results. 650

651

652

Acknowledgement 653

654

We sincerely thanks to Xie Yi (Chinese academy of tropical agricultural sciences, 655

Hainan, China) for her help in SEM analysis. The authors thank Dr. Sunil Kumar 656

Sahu (BGI-Shenzhen, China) and Jing Haiyu (BGI-Shenzhen, China) for RNA-seq 657

analysis. We are grateful to Anindya Sen, Bijoy Sen, Kawser Ayon, Jing Chen, Lei 658

Zhang, Tianlin Bi, Min Chen for their help in experiment and assistance with the 659

figure production. 660

661

Authors’ contributions 662

QH designed this study, obtained funding for this study, and supervised all research 663

activities. BB drafted and edited the manuscript, and conducted experiments. YT, FL, 664


https://doi.org/10.1101/2020.04.09.034587

JZ and CL performed bioinformatics analysis and figure production. ØØ, RI, BSH 665

and QH commented on experimental design, methodology, and substantially revised 666

the manuscript. All authors read and approved this final manuscript. 667

Funding 668

This work was funded by the National Key R&D Program of China 669

(2017YFD0501204 and 2017YFD0501200). 670

Availability of data and materials 671

RNA-seq data discussed in this article are publicly available at the NCBI database 672

under accession number PRJNA602095. 673

Ethics approval and consent to participate 674

Not applicable. 675

Consent for publication 676

Not applicable. 677

Competing interests 678

The authors declare that they have no competing interests. 679

680

References 681

682

683

1. Altner, H., Sass, H., Altner, I., 1977. Relationship between structure and function 684 of antennal chemo-, hygro-, and thermoreceptive sensilla in Periplaneta 685 americana. Cell Tissue Res. 176, 389e405. 686 http://dx.doi.org/10.1007/BF00221796. 687

2. Altner H, Prillinger L. 1980. Ultrastructure of invertebrate chemo-, thermo-, and 688 hygroreceptors and its functional significance. Int. Rev. Cytol. 1980;67:69-139 689

3. Anisimova M, Gascuel O. Approximate likelihood-ratio test for branches: a fast, 690 accurate, and powerful alternative. syst. biol. 55(4):539–552, 2006 691


https://doi.org/10.1101/2020.04.09.034587

4. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, 692 Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis 693 A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G, 694 Consortium TGO: Gene Ontology: tool for the unification of biology. Nat Genet 695 2000, 25:25–29. 696

5. Bartley, K., Huntley, J. F., Wright, H. W., Nath, M., Nisbet, A. J. (2012). 697 Assessment of cathepsin D and L-like proteinases of poultry red mite, 698 Dermanyssus gallinae (De Geer), as potential vaccine antigens. Parasitology, 699 139(6), 755-765. doi:10.1017/s0031182011002356 700

6. Barth FG. 2002. Spider senses - technical perfection and biology. Zoology 701 2002;105:271–85. 702

7. Benton R, Vannice KS, Vosshall LB. An essential role for a CD36-related 703 receptor in pheromone detection in Drosophila. Nature. 2007;450:289–93. 704

8. Benton, R., Vannice, K. S., Gomez-Diaz, C., Vosshall, L. B. Variant ionotropic 705 glutamate receptors as chemosensory receptors in Drosophila. Cell 136, 149–162 706 (2009). 707

9. Bhowmick B, Zhao J, Øines Ø, Bi T, Liao C, Zhang L, Han Q. Molecular 708 characterization and genetic diversity of Ornithonyssus sylviarum in chickens 709 (Gallus gallus) from Hainan Island, China. Parasit. Vectors. 2019;12:553. 710

10. Campbell, E.M., Budge, G.E., Bowman, A.S., 2012. Gene-knockdown in the 711 honeybee mite Varroa destructor by a non-invasive approach: studies on 712 aglutathione S-transferase. Parasit Vectors 3, 73. 713

11. Carr AL, Roe M. Acarine attractants: Chemoreception, bioassay, chemistry and 714 control. Pestic Biochem Physiol. 2016;131:60–79. 715

12. Carr, A.L., Mitchell III, R.D., Dhammi, A., Bissinger, B.W., Sonenshine, D.E., 716 Roe, R.M., 2017. Tick Haller’s organ, a new paradigm for arthropod olfaction: 717 how ticks differ from insects. Int. J. Mol. Sci. 18, 1563. 718

13. Conesa, A., Götz, S., García-Gómez, J. M., Terol, J., Talón, M., and Robles, M. 719 (2005). Blast2GO: a universal tool for annotation, visualization and analysis in 720 functional genomics research. Bioinformatics 21, 3674–3676. 721

14. Croset, V., Rytz, R., Cummins, S.F., Budd, A., Brawand, D., Kaessmann, H., 722 Gibson, T.J., Benton, R., 2010. Ancient protostome origin of chemosensory 723 ionotropic glutamate receptors and the evolution of insect taste and olfaction. 724 PLoS Genet. 6, e1001064. 725

15. Cruz MDS, Vega Robles MC, Jespersen JB, Kilpinen O, Birkett M, Dewhirst S. 726 et al. (2005). Scanning electron microscopy of foreleg tarsal sense organs of the 727 poultry red mite, Dermanyssus gallinae (DeGeer) (Acari:Dermanyssidae). 728 Micron. 2005;36:415–21. 729

16. Davis JC, Camin JH. Setae of the anterior tarsi of the martin mite, Dermanyssus 730 prognephilus (Acari: Dermanyssidae). J Kans Entomol Soc. 1976;49:441–49. 731


https://doi.org/10.1101/2020.04.09.034587

17. Dillier FX, Fluri P, Imdorf A. Review of the orientation behaviour in the bee 732 parasitic mite Varroa destructor: sensory equipment and cell invasion behaviour. 733 Rev. Suisse Zool. 2006;113:857–77. 734

18. Eliash, N., Singh, N.K., Thangarajan, S., Sela, N., Leshkowitz, D., Kamer, Y., 735 Zaidman, I., Rafaeli, A., Soroker, V. Chemosensing of honeybee parasite, Varroa 736 destructor: Transcriptomic analysis. Sci Rep. 2017, 7(1):13091. 737

19. Flochlay AS, Thomas E, Sparagano O. Poultry red mite (Dermanyssus gallinae) 738 infestation, a broad impact parasitological disease that still remains a significant 739 challenge for the egg-laying industry in Europe. Parasit Vectors. 2017;10:357. 740

20. George DR, Finn RD, Graham KM, Mul M, Maurer V, Valiente Moro C, 741 Sparagano OA. Should the poultry red mite Dermanyssus gallinae be of wider 742 concern for veterinary medical science. Parasit Vectors. 2015;8:178. 743

21. Götz S, García-Gómez JM, Terol J, Williams TD, Nagaraj SH, Nueda MJ, 744 Robles M, Talón M, Dopazo J, Conesa A: High-throughput functional annotation 745 and data mining with the Blast2GO suite. Nucleic Acids Res 2008, 746 36:3420–3435. 747

22. Groh-Lunow KC, Getahun MN, Grosse-Wilde E, Hansson BS. Expression of 748 ionotropic receptors in terrestrial hermit crab’s olfactory sensory neurons. 749 2015;8:448 750

23. Gainett, G., Michalik, P., Müller, C.H.G., Giribet, G., Talarico, G., Willemart, 751 R.H., 2017. 752

24. Ultrastructure of chemoreceptive tarsal sensilla in an armored harvestman and 753 evidence of olfaction across Laniatores (Arachnida, Opiliones). Arthropod Struct. 754 Dev. 46, 178–195. doi:10.1016/j.asd.2016.12.005 755

25. Guerrero FD, Kellogg A, Ogrey AN, Heekin AM, Barrero R, Bellgard MI, Dowd 756 SE, Leung M-Y. Prediction of G protein-coupled receptor encoding sequences 757 from the synganglion transcriptome of the cattle tick, Rhipicephalus microplus. 758 Ticks and Tick-borne Diseases. 2016; 7:670–677. 759

26. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. New 760 algorithms and methods to estimate maximum-likelihood phylogenies: assessing 761 the performance of PhyML 3.0. Syst Biol. 2010; 59(3):307–21. 762

27. Gu L, Knipple DC. Recent advances in RNA interference research in insects: 763 implications for future insect pest management strategies. Crop Prot 2013; 764 45:36–40. 765

28. Guo W, Rao B. On optimality of the Benjamini-Hochberg procedure for the false 766 discovery rate. Stat Probabil Lett. 2008;78(14):2024–30. 767

29. Gross, A. D., Temeyer, K. B., Day, T. A., Pérez de León, A. A., Kimber, M. J., 768 and Coats, J. R. (2015). Pharmacological characterization of a tyramine receptor 769 from the southern cattle tick, Rhipicephalus (Boophilus) microplus. Insect 770 Biochem. Molec. 63, 47–53. doi: 10.1016/j.ibmb.2015.04.008 771


https://doi.org/10.1101/2020.04.09.034587

30. Hall TA. BioEdit: a user-friendly biological sequence alignment editor and 772 analysis program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 773 1999;41:95–98. 774

31. Häußermann CK, Ziegelmann B, Bergmann P, Rosenkranz P. Male mites (Varroa 775 destructor) perceive the female sex pheromone with the sensory pit organ on the 776 front leg tarsi. Apidologie. 2015;46:771–78. 777

32. Holmes S, He H, Chen A, Lvie G, Pietrantonio P. Cloning and transcriptional 778 expression of a leucokinin�like peptide receptor from the Southern cattle tick, 779 Boophilus microplus (Acari: Ixodidae). Insect Molecular Biology. 2000; 780 9:457–465. 781

33. Ishida, Y., Tsuchiya, W., Fujii, T., Fujimoto, Z., Miyazawa, M., Ishibashi, J., 782 Yamazaki, T., 2014. Niemann-Pick type C2 protein mediating chemical 783 communication in the worker ant. Proc. Natl. Acad. Sci. U.S.A. 111, 3847–3852. 784

34. Iovinella, I., Ban, L., Song, L., Pelosi, P., Dani, F.R., 2016. Proteomic analysis of 785 castor bean tick Ixodes ricinus: a focus on chemosensory organs. Insect 786 Biochem. Mol. 78,58–68. 787

35. Javed S, Khan F, Ramirez-Fort M, Tyring SK. Bites and mites: prevention and 788 protection of vector-borne disease. Curr Opin Pediatr. 2013;25:488-91. 789

36. Josek, T., Walden, K. K. O., Allan, B. F., Alleyne, M., and Robertson, H. M. 790 (2018). A foreleg transcriptome for Ixodes scapularis ticks: candidates for 791 chemoreceptors and binding proteins that might be expressed in the sensory 792 Haller’s organ. Ticks Tick Borne Dis. 9, 1317–1327. 793

37. Kamau LM, Wright HW, Nisbet AJ, Bowman AS. Development of an 794 RNAinterference procedure for gene knockdown in the poultry red mite, 795 Dermanysus gallinae: studies on histamine releasing factor and Cathepsin-D. Afr 796 J Biotechnol. 2013;12(12):1350–6. 797

38. Katoh K, Misawa K, Kuma K-I, Miyata T. MAFFT: a novel method for rapid 798 multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res. 799 2002;30(14):3059–66. 800

39. Kim, D., Langmead, B. & Salzberg, SL HISAT: a fast spliced aligner with low 801 memory requirements. Nat. Methods 12, 357-360 (2015). 802

40. Kim, D., Šimo, L., and Park, Y. (2018). Molecular characterization of 803 neuropeptide elevenin and two elevenin receptors, IsElevR1 and IsElevR2, from 804 the blacklegged tick, Ixodes scapularis. Insect Biochem. Molec. 101, 66–75. doi: 805 10.1016/j.ibmb.2018.07.005 806

41. Kumar, S., Stecher, G., Tamura, K., 2016. MEGA7: molecular evolutionary 807 genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 808 1870–1874. 809

42. Lei J, Liu Q, Kadowaki T (2019) Honey Bee Parasitic Mite Contains the 810 Sensilla-Rich Sensory Organ on the Foreleg Tarsus Expressing Ionotropic 811


https://doi.org/10.1101/2020.04.09.034587

Receptors With Conserved Functions. Front. Physiol. 10:556. doi: 812 10.3389/fphys.2019.00556 813

43. Letunic I, Bork P (2007) Interactive Tree Of Life (iTOL): an online tool for 814 phylogenetic tree display and annotation. Bioinformatics 23:127–128. 815

44. Langmead, B. et al. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 816 357-359 (2012). 817

45. Li, B. & Dewey, CN RSEM: accurate transcript quantification from RNA-Seq 818 data with or without a reference genome. BMC Bioinformatics 12, 323 (2011). 819

46. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using 820 real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 821 25:402–408 822

47. Machanick, P., Bailey, T.L., 2011. Meme-chip: motif analysis of large DNA 823 datasets. Bioinformatics 27, 1696–1697. 824

48. Marr EJ, Sargison ND, Nisbet AJ, Burgess ST. RNA interference for the 825 identification of ectoparasite vaccine candidates, Parasite Immunol. 36 (2014) 826 616e626. 827

49. Marr, E. J., Sargison,N.D.,Nisbet, A. J., and Burgess, S. T.G. (2015).Gene 828 silencing by RNA interference in the house dust mite, Dermatophagoides 829 pteronyssinus. Mol. Cell. Probes. 29, 522–526. doi: 10.1016/j.mcp.2015.07.008 830

50. Palevich N, Britton C, Kamenetzky L, Mitreva M, de Moraes Mourão M, 831 Bennuru S, Quack T, Scholte LLS, Tyagi R, Slatko BE. 2018. Tackling 832 hypotheticals in helminth genomes. Trends Parasitol. 34:179–183. 833

51. Pándy-Szekeres G, Munk C, Tsonkov TM, Mordalski S, Harpsøe K, Hauser AS, 834 Bojarski AJ, Gloriam DE. GPCRdb in 2018: adding GPCR structure models and 835 ligands. Nucleic Acids Research. 2018; 46:D440–D446. 836

52. Pelosi, P., Iovinella, I., Felicioli, A., Dani, F.R., 2014. Soluble proteins of 837 chemical communication: an overview across arthropods. Front. Physiol. 5, 320. 838

53. Pelosi, P., Iovinella, I., Zhu, J., Wang, G., and Dani, F. R. (2017). Beyond 839 chemoreception: diverse tasks of soluble olfactory proteins in insects. Biol. Rev. 840 Camb. Philos. Soc. 93, 184–200. doi: 10.1111/brv.12339 841

54. Renthal, R., Manghnani, L., Bernal, S., Qu, Y., Griffith, W.P., Lohmeyer, K. et al. 842 (2017) The chemosensory appendage proteome of Amblyomma americanum 843 (Acari: Ixodidae) reveals putative odorant-binding and other 844 chemoreception-related proteins. Insect Science, 24, 730–742. 845

55. Sánchez-Gracia A, Vieira FG, Rozas J. Molecular evolution of the major 846 chemosensory gene families in insects. Heredity. 2009;103:208–16. 847

56. Sharma PP. Chelicerates. Curr Biol. 2018;28:774–8. 848


https://doi.org/10.1101/2020.04.09.034587

57. Schicht, S.; Qi, W.; Poveda, L.; Strube, C. Whole transcriptome analysis of the 849 poultry red mite Dermanyssus gallinae (De Geer, 1778). Parasitology 2014, 141, 850 336–346, doi:10.1017/S0031182013001467. 851

58. Thorne N, Chromey C, Bray S, Amrein H. Taste perception and coding in 852 Drosophila. Current biology : CB. 2004; 14: 1065-79. 853

59. Tichy, H., Barth, F.G., 1992. Fine structure of olfactory sensilla in myriapods and 854 arachnids. Micr. Res. Tech. 22, 372-391. 855

60. Van Giesen, L., and Garrity, P. A. (2017) More than meets the IR: the expanding 856 roles of variant ionotropic glutamate receptors in sensing odor, taste, temperature 857 and moisture. F1000Res. 6, 1753 CrossRef Medline 858

61. Van Leeuwen T, Dermauw W. 2016. The molecular evolution of xenobiotic 859 metabolism and resistance in chelicerate mites. Annu Rev Entomol. 860 2016;61:475–98. 861

62. Vosshall LB, Stocker RF. Molecular architecture of smell and taste in 862 Drosophila. Annu Rev Neurosci. 2007;30:505-33. 863

63. Vogt RG, Prestwich GD, Lerner MR. Odorant-binding-protein subfamilies 864 associate with distinct classes of olfactory receptor neurons in insects. J 865 Neurobiol. 1991;22:74–84. 866

64. Wang L, Feng Z, Wang X, et al. DEGseq: an R package for identifying 867 differentially expressed genes from RNA-seq data [J]. Bioinformatics, 2010, 868 26(1): 136-138. 869

65. Wang C, Ma Y, Huang Y, Xu J, Cai J, Pan B. An efficient rearing system rapidly 870 producing large quantities of poultry red mites, Dermanyssus gallinae (Acari: 871 Dermanyssidae), under laboratory conditions. Vet Parasitol. 2018;258:38–45. 872

66. Willemart RH, Chelini MC, de Andrade R, Gnaspini P. An ethological approach to 873 a SEM survey on sensory structures and tegumental gland openings of two 874 neotropical harvestmen (Arachnida, Opiliones, Gonyleptidae). Ital. J. Zool. 875 2007;74:39–54. 876

67. Whyard S, Erdelyan CNG, Partridge AL, Singh AD, Beebe NW, Capina R. 877 Silencing the buzz: a new approach to population suppression of mosquitoes by 878 feeding larvae double-stranded RNAs. Parasit Vectors. 2015;8:96. 879

68. Xiong C, Baker D and Pietrantonio PV (2019) The Cattle Fever Tick, 880 Rhipicephalus microplus, as a Model for Forward Pharmacology to Elucidate 881 Kinin GPCR Function in the Acari. Front. Physiol. 10:1008. doi: 882 10.3389/fphys.2019.01008 883

69. Zhu J, Guo M, Ban L, Song L-M, Liu Y, Pelosi P and Wang G (2018) 884 Niemann-Pick C2 Proteins: A New Function for an Old Family. Front. Physiol. 885 9:52. doi: 10.3389/fphys.2018.00052 886

887

888


https://doi.org/10.1101/2020.04.09.034587

889

Fig. 1 Scanning electron micrographs of PRM. A Ventral view of an adult PRM. B 890

The first leg (LI), indicating the distal sensilla-rich structure (circle). C Different 891

types of sensilla on the distal-most tarsomere (DT-I) of LI. D The second leg (LII). E 892

The third leg (LIII) and fourth leg (LIV). Different types of sensilla are not 893

discriminated in this figure because of low magnifications. Scale-bars: A, 100 μm B, 894

D, E, F, 20 μm, C, 2 μm. Abbreviations: st, sternal plate; gn, gnathostoma; as, anal 895

shield; ao, anal opening; ch, chelicera; Pa, pedipalp; (LI–IV), leg; gs, genitoventral 896

shield 897

898

899

900

Fig. 2 SEM of sensilla chaetica with or without tip-pore (Sc-tp). A. Circles indicating 901

sensilla chaetica with a tip pore or without pores on distalmost tarsomere of first leg. 902

Different types of sensilla are not discriminated in this figure because of low 903

magnifications. B. Shaft broken in its apical part revealing a pore and structure of 904

socket (�) C. no pore sensilla. D. Lateral view of the Sc-tp displaying longitudinal 905

grooves (arrows). E. Tip-pore sensilla showed indistinct pore on tarsomere I (circle) 906

F Tip-pore sensilla showed visible tip pore on tarsomere III and delimited by a round 907

margin (circle). Scale-bars: A, 20 μm B, 2 μm, C, 300 nm, D,E,F, 1 μm. 908

909


https://doi.org/10.1101/2020.04.09.034587

910

911

Fig. 3 SEM of sensilla chaetica with wall-pores (Sc-wp). A. Lateral view of different 912

types of hair sensilla on the distalmost tarsus (DT-I), circles indicate Sc-tp and 913

arrowhead indicate Sc-wp sensillum B. Structure of socket with an incomplete 914

articulation (�) C. shaft with longitudinal ridges (white arrows) and numerous pores 915

housed in longitudinal grooves (black circle) D. comparing tip pore (circle) and wall 916

pore sensilla (white arrows) E. White arrows represent longitudinal ridges, and 917

circles indicate lines of wall pores. Scale-bars: A, 10 μm B, 2 μm, C, 1 μm, D, E, 200 918

nm. 919

920

921


https://doi.org/10.1101/2020.04.09.034587

922

Fig. 4 Gene ontology (GO) classifications of DEGs for forelegs vs hindlegs. X-axis 923

represents the number of DEGs involved in particular GO term, and Y-axis indicates 924

significantly enriched GO term. 925

926

927

928

Fig. 5 a The protein domain organization of iGluRs/IRs in PRM is shown in 929

histogram. Phylogenetic tree of the IR/iGluR family based on PRM amino acid 930

sequences, and the corresponding protein domains of each member. b Hierarchical 931

clustering of the differentially expressed genes (DEGs). Blue to red colors represent 932

gene expression levels (i.e., FPKM values from −2 to 2) 933

934


https://doi.org/10.1101/2020.04.09.034587

935

936

Fig. 6 Phylogenetic tree of putative D. gallinae iGluRs and IRs along with insect and 937

non-insect arthropods iGluRs and IRs. The D. gallinae genes are shown in light 938

orange. Branch support was estimated by an approximate likelihood-ratio test ( 939

aLRT) (circles > 0.80). The amino acid sequences and a comprehensive list of 940

acronyms are listed in S2. 941

942

943

944


https://doi.org/10.1101/2020.04.09.034587

945

Fig. 7 a Alignment of the five putative OBP-like sequences of PRM together with 946

those of V. destructor sequences, A. americanum and I. scapularis. The highly 947

conserved six cysteine (C) residues are shown in blue. b Alignment of the six putative 948

NPC2 sequences of PRM together with those of V. destructor sequences. The highly 949

conserved six cysteine (C) residues are shown in blue. 950

951

952

953

954

Fig. 8 a Phylogenetic unrooted tree of odorant-binding protein (OBPs). Included are 955

OBPs from D. gallinae (Dg: 5), V. destructor (Vd: 5), I. scapularis (Ix: 3), A. 956

americanum (Aam: 1), T. mercedesae (Tp: 4), Stegodyphus mimosarum (Sm: 2), 957

Tetranychus urticae (Tu: 4), Sarcoptes scabiei (Sc: 2), Trichomalopsis sarcophagae 958

(Ts: 2), Aedes albopictus (Aal: 1), Aedes aegypti (Aa: 2), Papilio xuthus (Px: 1), 959

Folsomia candida (Fc: 1), D. sechellia (Ds: 1), D. ficusphila (Df: 1), Euroglyphus 960

maynei (Em: 1), Orchesella cincta (Oc: 1) and D. willistoni (Dw: 1). Bootstrap values 961

were estimated using a recent fast approximate likelihood-ratio test ( aLRT) and are 962

displayed as circles on the branches, and are proportionate to the circle diameter 963

(0–500). Leaves are coloured according to the organisms classes: arachnida (black), 964

collembola (purple) and insecta (blue). Dermanyssus transcripts are highlighted in 965

green.The amino acid sequences and a comprehensive list of acronyms are listed in 966


https://doi.org/10.1101/2020.04.09.034587

S3. b Hierarchical clustering analysis of the differentially expressed genes (DEGs). 967

Blue to red colors represent gene expression levels (i.e., FPKM values from −1.5 to 968

1.5). 969

970

971

972

Fig. 9 Motif-pattern analysis of OBP protein from D. gallinae and other classes 973

including insecta and arachnida. Eight amino acid motifs with various widths and 974

corresponding e-values were identified, and the lower part indicates approximate 975

locations of each motif on the protein sequence. Different motifs are shown by 976

different colored boxes, where small number indicates high conservation. Protein 977

sequence of OBPs used in this analysis are listed in table S4. 978


https://doi.org/10.1101/2020.04.09.034587

979

980

Fig. 10 a An unrooted phylogenetic tree of NPC2 proteins. The evolutionary tree is 981

constructed by aligning the SNMP sequences of V. destructor (Vd: 6), Dermanyssus 982

gallinae (Dg: 6), Drosophila melanogaster (Dm: 6), Ixodes scapularis (Isc: 11) and 983

Tropilaelaps mercedesae (Tm: 6). Bootstrap values were estimated using a recent fast 984

approximate approximate likelihood ratio test ( aLRT) and are displayed as circles on 985

the branches, and are proportionate to the circle diameter (0–500). Leaves are colored 986

according to the organisms classes: Insecta (blue) and Arachnida (black). 987

Dermanyssus transcripts are highlighted in green. The amino acid sequences and a 988

comprehensive list of acronyms are listed in S5. b Hierarchical clustering analysis of 989

the differentially expressed genes (DEGs). Blue to red colors represent gene 990

expression levels (i.e., FPKM values from −1.5 to 1.5). 991

992


https://doi.org/10.1101/2020.04.09.034587

993

994

Fig.11 Motif analysis of NPC2 protein from D. gallinae and other classes including 995

insecta and arachnida. Eight amino acid motifs with various widths and 996

corresponding e-values were identified, and the lower part indicates approximate 997

locations of each motif on the protein sequence. Different motifs are shown by 998

different colored boxes, where small number indicates high conservation. Protein 999

sequence of NPC2 used in this analysis are listed in table S6. 1000

1001

1002

1003


https://doi.org/10.1101/2020.04.09.034587

Fig. 12. a An unrooted phylogenetic tree of SNMP proteins. The tree is constructed 1004

by aligning the SNMP sequences of V. destructor (Var: 7), D. gallinae (Dg: 5) T. 1005

mercedesae (Trp: 12), Ixodes scapularis (Isc: 6) and Drosophila melanogaster (Dm: 1006

19). Bootstrap values were estimated using a recent fast approximate likelihood ratio 1007

test ( aLRT) and are displayed as circles on the branches, and are proportionate to the 1008

circle diameter (0–500). Leaves are coloured according to the organisms classes: 1009

Insecta (blue) and Arachnida (black). Dermanyssus transcripts are highlighted in 1010

green. b Hierarchical clustering analysis of the differentially expressed genes 1011

(DEGs). Blue to red colors represent gene expression levels (i.e., FPKM values from 1012

−2 to 2). 1013

1014

1015

1016

Fig. 13. Phylogenetic tree of rhodopsin-like GPCR of D. gallinae and other mite 1017

species. The amino acid sequences are listed in S8. 1018

1019


https://doi.org/10.1101/2020.04.09.034587

1020

1021

Fig. 14 Expression profiles of putative chemosensory genes in D. gallinae using 1022

qRT-PCR. FL, forelegs; HL, hindlegs. Error bars represent means ±SEM, and 1023

asterisks above indicate significant differences between forelegs and hindlegs (*p < 1024

0.05). 1025


https://doi.org/10.1101/2020.04.09.034587

1026

1027

Fig. 15 Immersion-based gene silencing of candidate DgOBP-1 in D. gallinae by 1028

RNAi. Results are presented as the mean ± the SEM. Statistical differences relative to 1029

the control group are presented as *p–value < 0.05; NS, no statistical significance. 1030

Abbreviations: dsRNA, treatment of mites with ds-OBP; GUS, treatment of mites 1031

with ds-GUS RNAs; BC, treatment of mites with DEPC-treated water (without any 1032

dsRNA). 1033

1034


https://doi.org/10.1101/2020.04.09.034587

1035

1036

Fig. 16 The number of chemosensory genes in different mite and tick species are 1037

presented in bar plot. A phylogenetic tree showing evolutionary relationships between 1038

various species are illustrated on the left. The number of chemosensory genes are 1039

obtained from V. destructor, I. scapularis, D. variabilis. 1040

1041

1042


https://doi.org/10.1101/2020.04.09.034587

Dermanyssus gallinae and knockdown by RNA84 chemoreception (Barth, 2002; Willemart et al., 2007)....

Documents

Transcript of Dermanyssus gallinae and knockdown by RNA84 chemoreception (Barth, 2002; Willemart et al., 2007)....