CO2 per se activates carbon dioxide receptors

Journal Pre-proof CO2 per se activates carbon dioxide receptors Pingxi Xu, Xiaolan Wen, Walter S. Leal PII:

S0965-1748(19)30398-4

DOI:

https://doi.org/10.1016/j.ibmb.2019.103284

Reference:

IB 103284

To appear in:

Insect Biochemistry and Molecular Biology

Received Date: 11 October 2019 Revised Date:

14 November 2019

Accepted Date: 15 November 2019

Please cite this article as: Xu, P., Wen, X., Leal, W.S., CO2 per se activates carbon dioxide receptors, Insect Biochemistry and Molecular Biology, https://doi.org/10.1016/j.ibmb.2019.103284. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 The Author(s). Published by Elsevier Ltd.

1

CO2 per se activates carbon dioxide receptors

2

Pingxi Xu, Xiaolan Wen1, and Walter S. Leal*

3

Department of Molecular and Cellular Biology, University of California-Davis

4

Davis, CA 95616, USA

5 6

1

7

China

8

*Correspondence: [email protected]

9

Lead Contact: Walter S. Leal, [email protected], phone: 530-752-7755

Current address: School of Preclinical Medicine, Guangxi Medical University, Nanning 530021,

10 11 12 13 14 15 16 17 18

1

19

Abstract

20

Carbon dioxide has been used in traps for more than six decades to monitor mosquito

21

populations and help make informed vector management decisions. CO2 is sensed by gustatory

22

receptors (GRs) housed in neurons in the maxillary palps. CO2-sensitive GRs have been

23

identified from the vinegar fly and mosquitoes, but it remains to be resolved whether these

24

receptors respond to CO2 or bicarbonate. As opposed to the vinegar fly, mosquitoes have three

25

GR subunits, but it is assumed that subunits GR1 and GR3 form functional receptors. In our

26

attempt to identify the chemical species that bind these receptors, we discovered that GR2 and

27

GR3 are essential for receptor function and that GR1 appears to function as a modulator. While

28

Xenopus oocytes coexpressing Culex quinquefasciatus subunits CquiGR1/3 and CquiGR1/2

29

were not activated, CquiGR2/3 gave robust responses to sodium bicarbonate. Interestingly,

30

CquiGR1/2/3-coexpressing oocytes gave significantly lower responses. That the ternary

31

combination is markedly less sensitive than the GR2/GR3 combination was also observed with

32

orthologs from the yellow fever and the malaria mosquito. By comparing responses of

33

CquiGR2/CquiGR3-coexpressing oocytes to sodium bicarbonate samples (with or without

34

acidification) and measuring the concentration of aqueous CO2, we showed that there is a direct

35

correlation between dissolved CO2 and receptor response. We then concluded that subunits

36

GR2 and GR3 are essential for these carbon dioxide-sensitive receptors and that they are

37

activated by CO2 per se, not bicarbonate.

38 39

Keywords

40

Culex quinquefasciatus, Anopheles gambiae, Aedes aegypti, CquiGR1, CquiGR2, CquiGR3

41

2

42

1. Introduction

43 44

Carbon dioxide is the oldest known and the most powerful mosquito attractant (Van Thiel and

45

Weurman, 1947). It has been widely used for decades for trapping host-seeking female

46

mosquitoes (Reeves, 1951, 1953) to monitoring populations and to assist in integrated vector

47

management programs. In addition to being an attractant sensu stricto, CO2 gates mosquito

48

perception of other sensory modalities (Gillies, 1980; Liu and Vosshall, 2019; McMeniman et

49

al., 2014). In the vinegar fly, Drosophila melanogaster, CO2 elicits through different pathways

50

attraction and aversion behaviors (van Breugel et al., 2018). Mosquitoes sense CO2 with

51

olfactory receptor neurons (ORNs) (Grant et al., 1995; Majeed et al., 2017; Syed and Leal,

52

2007) housed in peg sensilla (= capitate pegs) in the maxillary palps (McIver and Charlton,

53

1970). It has been unambiguously demonstrated that CO2 is sensed by the gustatory receptors

54

DmelGR21a and DmelGR63a, housed in antennal neuron ab1C of the vinegar fly (Jones et al.,

55

2007; Kwon et al., 2007; Suh et al., 2004), but it remains to be resolved whether carbon dioxide

56

receptors are activated by CO2 per se or bicarbonate (Jones et al., 2007; Kwon et al., 2007;

57

Ning et al., 2016; Xu and Anderson, 2015).

58

Mosquitoes have three closely related homologs of DmelGR21a and DmelGR63a (Robertson

59

and Kent, 2009), GR1, GR2, and GR3. Because they were identified before the new

60

nomenclature for these GR genes was proposed (Robertson and Kent, 2009), GR1-3 in the

61

malaria mosquito, Anopheles gambiae, are still named AgamGR22, AgamGR23, and

62

AgamGR24, respectively (Hill et al., 2002). While it has been clearly demonstrated that GR3 is

63

essential for CO2 reception in the yellow fever mosquito, Aedes aegypti (McMeniman et al.,

3

64

2014), it is not yet known whether GR1 and GR2 are functionally redundant (McMeniman et

65

al., 2014). While no response to CO2 was recorded when both AgamGR22 and AgamGR24

66

(one copy of each) were coexpressed in the empty neuron system of the vinegar fly (Hallem et

67

al., 2004), by increasing the dosage of both transgenes by two-fold led to significant response

68

(Lu et al., 2007). Additionally, flies carrying one copy of each of the three subunits showed a

69

significant response to CO2, albeit not as strong as responses recorded from the flies carrying

70

two-fold of AgamGR22 and 24 (Lu et al., 2007). Thus, AgamGR23 was implicated in CO2

71

reception in the malaria mosquito, but its role was not clarified. On the other hand, by using

72

RNAi combined with behavioral measurements, it has been suggested that AaegGR2 had no

73

role in CO2 reception by the yellow fever mosquito (Erdelyan et al., 2012). In summary, it

74

remains to be determined whether all three GR subunits are functionally required for CO2

75

detection in mosquitoes or whether GR1 and GR2 are functionally redundant (McMeniman et

76

al., 2014).

77

We envisioned that having a functional carbon dioxide-detecting system in an aqueous

78

environment as in the Xenopus oocyte recording system would allow us to address with the Le

79

Chatelier’s principle the long-standing question whether carbon dioxide receptors are activated

80

by CO2 per se or by bicarbonate (Jones et al., 2007; Kwon et al., 2007; Ning et al., 2016; Xu

81

and Anderson, 2015). Because dissolved CO2 forms an equilibrium with bicarbonate in water

82

⇄

+

, a shift of the equilibrium towards dissolved CO2 would increase a

83

receptor response if CO2 per se binds carbon dioxide receptors, whereas a reduced response

84

would indicate activation by bicarbonate. Here, we show that CO2 per se, not bicarbonate,

85

activates the receptors. While preparing to address this question, we discovered that the

4

86

gustatory receptor CquiGR2 from the southern house mosquito, Culex quinquefasciatus, is

87

essential for function, whereas CquiGR1 appears to be a modulator. Additionally, we show a

88

similar GR1 effect in the Ae. aegypti and An. gambiae CO2-sensing system. Specifically, CO2

89

elicited weaker responses in the ternary receptor systems than in the respective binary

90

counterparts devoid of GR1 (GR22 in the case of the malaria mosquito receptors).

91 92

2. Material and methods

93

2.1. RNA extraction, DNA synthesis, and cloning

94

Three hundred pairs of palps from 4- to 6-d-old female Culex mosquitoes were dissected under

95

a stereomicroscope (Zeiss, Stemi DR 1663) and collected in 75% (vol/vol) ethanol diluted in

96

diethylpyrocarbonate (DEPC)-treated water on ice. The samples were centrifuged, and ethanol

97

was removed before total RNA was extracted using TRIzol reagent (Invitrogen). cDNA was

98

synthesized from 1 µg of total RNA using GoScript™ Reverse Transcription kit according to

99

the manufacturer’s instructions (Promega). For An. gambiae and Ae. aegypti, gene sequences of

100

AgamGR22, 23, 24 and AaegGR1, 2, 3 from Genbank were synthesized by

101

GenScript USA Inc. (Piscataway, NJ). Full-length gene-specific cloning primers were designed

102

based on the sequences from Genbank and added 15-bp In-Fusion cloning adapters.

103

CqGR1-Fw GATCAATTCCCCGGGaccATGATTCACAGTCAGATGGAGGACG

104

CqGR1-Rv CAAGCTTGCTCTAGACTATTGTGGGTTCGTTTTGGCCG

105

CqGR2-Fw GATCAATTCCCCGGGaccATGGTCATCAAGGACAGTGACTTTGAC

106

CqGR2-Rv CAAGCTTGCTCTAGATTAGTGAGCGTGAGCTTTCTGTAAATTCTC

107

CqGR3-Fw GATCAATTCCCCGGGaccATGAGCATATTTCCGGATACTCTGCG

5

108

CqGR3-Rv CAAGCTTGCTCTAGATCAATGCGCTGCCGTCG

109

AgGR22-Fw GATCAATTCCCCGGGaccATGATTCACACACAGATGGAAG

110

AgGR22-Rv CAAGCTTGCTCTAGATTAGGTGTTCACTTTGTCTGC

111

AgGR23-Fw GATCAATTCCCCGGGaccATGGTTATCAAGGAAAGTGAGTTC

112

AgGR23-Rv CAAGCTTGCTCTAGATTACTGTTTGTGTAGCAGCTTAACA

113

AgGR24-Fw GATCAATTCCCCGGGaccATGAGTCTCTACTTCAACGCGG

114

AgGR24-Rv CAAGCTTGCTCTAGACTAAGAATGAGACGAATTACTGTGC

115

AaGR1-Fw GATCAATTCCCCGGGaccATGATTCACAGCCAGATGGAAG

116

AaGR1-Rv CAAGCTTGCTCTAGACTAGTTCTCCTTCAGCTTAGTTAGA

117

AaGR2-Fw GATCAATTCCCCGGGaccATGGTCATCAAAGACAGTGAGT

118

AaGR2-Rv CAAGCTTGCTCTAGACTATCCCTTATGACTGTGCTTGATT

119

AaGR3-Fw GATCAATTCCCCGGGaccATGAATCTCAACCAAGACCCCA

120

AaGR3-Rv CAAGCTTGCTCTAGACTACTCGCGATATGAACCCGTCATA

121

Adapter sequences are underlined, and lower case acc is Kozak consensus. PCR products were

122

purified by Monarch® DNA Gel Extraction Kit (NEBioLab) and directly inserted to linearized

123

destination vector-pGEMHE by In-Fusion reaction (Clontech). The colonies were picked, and

124

vectors were extracted using Monarch® Plasmid Miniprep Kit. Inserts were submitted to DNA

125

Sequencing Facility, UC Berkeley for verification of sequences.

126 127

2.2. Oocytes preparations and two-electrode voltage clamp recordings

128

The vectors carrying GR genes were linearized with restriction endonuclease NheI, except for

129

AgamGR22, which was cut by SphI. The linearized vectors with gene inserts were used as

6

130

templates to transcribe capped cRNAs with poly(A) using an mMESSAGE mMACHINE T7 kit

131

(Ambion) following the manufacturer's protocol. The cRNAs were dissolved in RNase-free

132

water and adjusted at a concentration of 200 µg/mL by UV spectrophotometry (NanoDrop™

133

Lite Spectrophotometer, ThermoFisher). 9.2 nl of each cRNA samples were microinjected into

134

stage V or VI Xenopus oocytes (purchased from EcoCyte Bioscience, Austin, TX). Then

135

injected oocytes were incubated at 18 °C for 3–7 days in modified Barth's solution [in mM: 88

136

NaCl, 1 KCl, 2.4 NaHCO3, 0.82 MgSO4, 0.33 Ca(NO3)2, 0.41 CaCl2, 10 HEPES, pH 7.4]

137

supplemented with 10 µg/mL of gentamycin, 10 µg/mL of streptomycin, and 1.8 mM sodium

138

pyruvate. For two-electrode voltage clamp (TEVC) recordings, as previously done (Xu et al.,

139

2019), oocytes were placed in a perfusion chamber and challenged with test compounds.

140

Compound-induced currents were amplified with an OC-725C amplifier (Warner Instruments,

141

Hamden, CT), the voltage held at −80 mV, low-pass filtered at 50 Hz and digitized at 1 kHz.

142

Data acquisition and analysis were carried out with Digidata 1440A and pClamp10 software

143

(Molecular Devices, LLC, Sunnyvale, CA).

144 145

2.3. Sample preparations and CO2 measurement

146

Sodium chloride (Fisher Scientific, >99%) and sodium bicarbonate (Sigma-Aldrich, 99.7%)

147

were used to prepare fresh samples by dissolving the appropriate amounts of these salts in

148

perfusion Ringer buffer (NaCl 96 mM, KCl 2 mM, CaCl2 1.8 mM, MgCl2 1 mM, HEPES 5

149

mM, pH 7.6) to make 0.5 M solutions. Then, the desired concentrations were prepared by

150

diluting with perfusion Ringer buffer. CO2 samples were prepared by bubbling MediPure™

7

151

CO2 (Praxair, Danbury, CT) directly into perfusion buffer for 5 s, and subsequently adjusting

152

the pH.

153

The concentration of dissolved CO2, ie, CO2 (aq), was calculated by equation I:

154 155

Koverall, which is sometimes referred to as Ka, is a constant, which incorporates the constant of

156

hydrolysis of CO2 and the first dissociation constant of carbonic acid, Ka1, ie, Koverall = Kh x Ka1.

157

pKa is 6.3 (https://pubchem.ncbi.nlm.nih.gov/compound/Sodium-bicarbonate#section=pKa).

158

The molar concentration of dissolved CO2 is, therefore, related to the pH of the solution and the

159

nominal concentration of sodium bicarbonate. Thus, the concentration of dissolved CO2 in a 10

160

mM NaHCO3 solution at pH 7.73 is approximately 0.358 mM or 15.76 ppm. At pH 7.32, the

161

calculated concentration of dissolved CO2 is 0.872 mM or 38.36 ppm.

162

pH and carbon dioxide were measured with an Orion Star™ A214 pH/ISE Benchtop Meter

163

(ThermoFisher Scientific, Waltham, MA 02451). A carbon dioxide electrode was used for CO2

164

measurements. The internal filling solution of this electrode is separated from the sample

165

solution with a gas-permeable membrane. CO2 dissolved in a sample solution diffuses through

166

this membrane until an equilibrium between CO2 in the electrode filling and the sample solution

167

is reached. The influx of CO2 causes a change in hydrogen ion concentration in the electrode

168

filling solution, which is measured by a sensing element behind the membrane. The system was

169

calibrated before measurements with freshly prepared 22, 55 and 110 ppm solutions. Standard

170

solutions were sealed with parafilm before measurements. To minimize losses of carbon

8

171

dioxide during measurements, we used 50 ml Falcon tubes to house sample solutions. A hole

172

was opened in the tube cap, and the electrode was inserted and kept in place with two O-ring

173

seals below and two O-ring seals on the top of the cap. Then the cap was sealed with Teflon

174

tapes and finally covered with electrical tape. Fifty milliliter Falcon tubes housing test samples

175

were attached to the electrode and fastened tightly. The concentrations of dissolved CO2 in

176

bicarbonate solutions were measured by preparing 10, 25, 50, 100, 150, 200, 250, and 300 mM

177

solutions in perfusion Ringer buffer, with four replicates for each sample. To compare the

178

concentrations of dissolved CO2 without adjustment and after acidification, 7 ml aliquots of 10

179

mM sodium bicarbonate buffers were prepared in pairs. One of the two aliquots in a pair was

180

used to determine CO2 concentration without pH change. To the other pair, 150 µl of 0.1 N HCl

181

was added and immediately used for CO2 measurement, and then the pH was recorded. To

182

compare oocyte response, samples were similarly prepared by using smaller aliquots (700 µl) in

183

V-vials. In this case, 0.1 N HCl was injected through the Teflon liner of the open-top caps.

184

After that, an aliquot was collected to be applied to the Xenopus oocyte recording system, and

185

the remainder was used for pH measurement.

186 187

2.4. Statistical analysis and graphical preparations

188

Prism 8.2.0 from GraphPad Software (La Joya, CA) was used for both statistical analysis and

189

graphical preparations. A dataset that passed the Shapiro-Wilk normality test was analyzed by t-

190

test; otherwise, data were analyzed by Wilcoxon matched-pairs signed-rank test. All data are

191

presented as mean ± SEM.

192

9

193

3. Results and discussion

194

3.1. Cx. quinquefasciatus gustatory receptor GR2/GR3 is sensitive to carbon dioxide

195

We cloned three gustatory receptors from the southern house mosquito, specifically CquiGR1

196

(VectorBase, CPIJ006622), CquiGR2 (GenBak, MN428502), and CquiGR3 (MN428503),

197

which are likely to be involved in CO2 reception. Initially, we coexpressed combinations of

198

these three GRs in the Xenopus laevis oocytes and recorded their responses to sodium

199

bicarbonate. Of note, the application of sodium bicarbonate increases the sodium concentration

200

in the perfusion Ringer, thus causing inward currents through endogenous oocyte channels due

201

to an increase in the extracellular concentration of Na+ (Fig. 1A, Fig. S1). We, therefore,

202

applied equivalent doses of NaCl and NaHCO3 in all experiments to control for the Na+

203

response. Equal responses to NaCl and NaHCO3 at the same dose, as in the case of naked

204

oocytes (Fig. 1A), indicate that the currents were generated by Na+ ions. Surprisingly, oocytes

205

coexpressing CquiGR1 and CquiGR3 did not respond to bicarbonate as the response to sodium

206

bicarbonate did not differ from the response to sodium chloride (Fig. 1B, Fig. S1). A lack of

207

specific response to sodium bicarbonate was also observed with CquiGR1/CquiGR2-

208

coexpressing oocytes (Fig. 1C, Fig. S1). By contrast, CquiGR2/CquiGR3-expressing oocytes

209

gave robust dose-dependent responses to sodium bicarbonate (Figs. 1D, 2, Fig. S1). Although it

210

is not surprising that injections of sodium chloride generated dose-dependent currents, they

211

were significantly smaller (t-test, P=0.0008) than those generated by sodium bicarbonate at the

212

same dose (Fig. 2).

10

213

We then compared CquiGR2/CquiGR3 and CquiGR1/CquiGR2/CquiGR3 (Fig. 3). The ternary

214

receptor system showed a remarkably lower (t-test, P=0.0320) response to sodium bicarbonate

215

than the binary receptor (Fig. 3).

216

Next, we examined whether variants of CquiGR1 would have different effects. We cloned five

217

CquiGR1 variants, one of them showed an amino acid sequence identical to the sequence

218

appearing in VectorBase (CPIJ 006622). We named this variant the wild type form. Three other

219

variants differed in one amino acid residue each, specifically CquiGR1K456N (GenBank,

220

MN418391), CquiGR1M365V (MN418394), and CquiGR1L246F (MN418393). Lastly, one

221

variant differed in two amino acid residues, ie, CquiGR1D143G;R434K (MN418392). We then

222

expressed variants along with CquiGR2+CquiGR3 in the Xenopus oocyte recording system and

223

compared their responses to those obtained with CquiGR2/CquiGR3-coexpressing oocytes

224

when stimulated with the same dose of sodium bicarbonate (Fig. S2). None of the five ternary

225

receptor systems showed a stronger response than those recorded from CquiGR2/CquiGR3-

226

coexpressing oocytes (Fig. 4). Responses recorded from CquiGR1D143G;R434K were not

227

significantly different from those measured from the binary receptor system (n = 4, unpaired t-

228

test, P=0.0710). All other variants, including the wild type, generated significantly lower

229

responses than those recorded from CquiGR2/CquiGR3-coexpressing oocytes (Fig. 4). These

230

findings suggest that CquiGR1 may modulate the receptor system response to sodium

231

bicarbonate.

232 233

3.2. Gustatory receptors are activated by carbon dioxide per se, not bicarbonate

11

234

To address this question, we used a carbon dioxide electrode, which was designed to measure

235

the total amount of dissolved carbon dioxide in solution. The solution in the internal filling of

236

the electrode is separated from the analytical sample by a CO2 permeable membrane. Typically,

237

a carbon dioxide buffer is added to the analytical sample to lower the pH to 4.8-5.2 thus shifting

238

the CO2/bicarbonate equilibrium

239

dioxide diffuses through the membrane, and an equilibrium is reached between the

240

concentration of CO2 in the analyte and the internal filling. The increase of CO2 in the internal

241

filling causes an increase in hydrogen ion concentration, which is ultimately measured by the

242

pH electrode. To measure the actual (equilibrium concentration) rather than the total

243

concentration of carbon dioxide, we used this electrode in a sealed system and without the

244

carbon dioxide buffer. To calibrate our measurement system, we compared the calculated

245

concentrations of CO2 in sodium bicarbonate buffers with those obtained by direct

246

measurements with the carbon dioxide electrode (Fig. 5). The results showed a faithful

247

relationship (two-tailed t-test, P=0.1037) between calculated and measured CO2 concentrations

248

(Fig. 5). We then used this system to measure the concentrations of dissolved CO2 in samples

249

obtained by bubbling CO2 in perfusion Ringer buffer. The higher the concentration of CO2, the

250

larger the currents recorded from CquiGR2/CquiGR3-coexpressing oocytes (Fig. S3). Of note,

251

currents recorded from sodium bicarbonate is the summation of carbon dioxide and Na+

252

responses. Although we adjusted the pH of the tested bubbled CO2 samples (Fig. S3) to rule out

253

the possible effect of hydrogen ion concentrations on oocyte responses, these recordings did not

254

unambiguously determine whether the receptor system is responding to carbon dioxide per se or

255

bicarbonate.

⇄

+

towards free CO2. Carbon

12

256

The next experiment further examined the possible role of pH on recorded currents. We

257

injected Ringer and 50 mM sodium chloride samples with their pH adjusted to that of the

258

perfusion Ringer buffer, as well as to 7, 6.5, 6, and 5.5. To make certain that the

259

CquiGR2/CquiGR3-coexpressing oocyte system was functional throughout the tests, we

260

injected Ringer, sodium chloride, and sodium bicarbonate at the beginning and the end of each

261

run (Fig. S4). These experiments (n = 6) further demonstrated that hydrogen ion concentration

262

at the pH range of 7.94 to 5.5 had little or no effect on sodium bicarbonate-elicited response

263

(Fig. S4).

264

Having developed tools to measure the amount of free carbon dioxide in solution, CO2 (aq), and

265

demonstrated that pH change had minimal or no effect on oocyte responses, we next tested

266

whether sifting the equilibrium of CO2/bicarbonate buffers would cause an increase or a

267

decrease in the receptor response. We prepared in duplicates aliquots of 10 mM sodium

268

bicarbonate buffers and placed them in sealed V-vials with Teflon liner, open-top caps. One of

269

the experimenters took a sample from one of a pair of vials, applied to a CquiGR2/CquiGR3-

270

coexpressing oocyte, and recorded the elicited current. Meanwhile, the other experimenter

271

injected an aliquot of 0.1 N HCl to the second pair of the same sample, vortexed, and provided

272

the sample to the other experimenter to apply immediately to the oocyte preparation (Fig. S5).

273

The pairs of unadjusted and pH adjusted samples were kept sealed until their pH values were

274

recorded. As expected, the initial pH was nearly the same (pH = 7.73) because they were

275

aliquots from the same 10 mM sodium bicarbonate sample. The final pH of the tested samples

276

was 7.39 ± 0.05.

13

277

Quantification of the recorded currents from paired samples showed a significant increase (two-

278

tailed t-test, n = 12, p = 0.0005) in the lower pH samples (Fig. 6A). Similar experiments were

279

performed to measure the actual concentration of dissolved CO2 in paired samples with initial

280

and lowered pH. The initial and final pH values of these samples (n = 4) were 7.73 and 7.32 ±

281

0.03, respectively. As expected, the quantification of CO2 (aq) showed an increase in dissolved

282

CO2 at lower pH (Fig. 6B). The measured CO2 concentrations in the samples without

283

adjustment (pHinitial) and with pH adjusted (pHlowered) were 17.90 ± 1.55 ppm and 32.40 ± 2.40

284

ppm, whereas the calculated CO2 concentrations were approximately 15.8 and 38.4 ppm,

285

respectively. An increase in dissolved CO2 concentration by lowering pH is predicted by Le

286

Chatelier’s Principle, given that an increase in H+ concentration shifts the equilibrium towards

287

CO2 (aq). Because the increased oocyte response (Fig. 6A) resulted from the increase of in situ

288

CO2 concentration (Fig. 6B), we concluded that CO2 per se, not bicarbonate, activates the Cx.

289

quinquefasciatus gustatory receptor system when expressed in Xenopus oocytes. It has also

290

been reported that the BAG neurons of the nematode Caenorhabditis elegans are activated by

291

molecular CO2, although they can be activated by acid stimuli (Smith et al., 2013). We found

292

no evidence for the activation of the Culex mosquito GR system by hydrogen ions.

293 294

3.3. CquiGR1 orthologs in An. gambiae and Ae. aegypti are putative modulators

295

Given the unexpected results that CquiGR1 was not necessary for receptor function, but rather

296

attenuated the responses of the CquiGR2/CquiGR3 receptor system to CO2, we asked whether

297

this would be a common feature of mosquito biology. We prepared oocytes to coexpress the

298

three GR subunits from the malaria mosquito, specifically AgamGR22, AgamGR23, and

14

299

AgamGR24. Using the same batch of oocytes, we prepared a binary combination of these

300

receptors devoid of AgamGR22, ie, AgamGR23/AgamGR24-expressing oocytes. Then we

301

compared the responses of these oocytes to the same concentrations of sodium chloride and

302

sodium bicarbonate. Specifically, we challenged each oocyte preparation of the same age and

303

the same batch of oocytes with 50, 100, 200, and 300 mM solutions of NaCl and NaHCO3.

304

AgamGR23/AgamGR24-coexpressing oocytes gave stronger, concentration-dependent

305

responses to sodium bicarbonate (Fig. 7A) than the oocytes coexpressing a ternary receptor

306

system, AgamGR22/AgamGR23/AgamGR24 (Fig. 7B). These experiments were replicated

307

with another batch of oocytes (Fig. 7C,D). These results suggest that the CquiGR1 ortholog

308

from the malaria mosquito, ie, AgamGR22, functions as a modulator.

309

Next, we prepared oocytes coexpressing the equivalent binary and ternary receptor systems

310

with GRs from the yellow fever mosquito. Specifically, we compared the responses of

311

AaegGR2/AaegGR3-coexpressing oocytes to oocytes of the same batch and the same age that

312

coexpressed AaegGR1, AaegGR2, and AaegGR3. Again, the receptor system devoid of

313

AaegGR1 generated stronger, dose-dependent responses to bicarbonate (Fig. S6A) than the

314

ternary system (Fig. S6B), thus implicating AaegGR1 in a modulatory role.

315 316

3.4. Overall conclusions

317 318

Our research plans to address a long-standing question regarding CO2 reception led to

319

unexpected results. Because it has been well-established in the literature that the reception of

320

CO2 in the vinegar fly antennae is mediated by DmelGR21a and DmelGR63a (Jones et al.,

15

321

2007; Kwon et al., 2007; Suh et al., 2004), it has been assumed that CO2 is detected by

322

GR1+GR3 in mosquitoes. This notion has been supported by RNAi-based experimental data

323

suggesting that GR2 was not involved in CO2 reception in Ae. aegypti (and Cx.

324

quinquefasciatus) (Erdelyan et al., 2012). On the other hand, AgamGR23, the CquiGR2

325

ortholog from the malaria mosquito An. gambiae, was implicated in CO2 response (Lu et al.,

326

2007), but the role of AgamGR22 is yet to be unraveled. Our findings suggest that the three

327

subunits might be involved, with GR1 (GR22 in the case of An. gambiae) being a putative

328

modulator. Future studies will determine whether GR2 (or GR23) knock out mosquitoes are

329

indeed insensitive to CO2, whereas GR1 (or GR22) lines will allow us to get a better

330

understanding of the mechanism(s) of modulation. It is conceivable that only the reception of

331

CO2 at high doses might be different in GR1 (or GR22) knock out mosquitoes. With a

332

functional carbon dioxide-detecting receptor system in an aqueous environment, we applied a

333

fundamental principle in chemistry to interrogate which of two species in equilibrium (CO2 or

334

HCO3-) might activate the receptors. We compared receptor responses as well as CO2

335

concentrations in paired samples without acidification and after acidification and showed that

336

an increase in the concentrations of dissolved CO2 was associated with an increase in receptor

337

response. We then concluded that CO2 per se, neither HCO3- nor protons, activates the carbon

338

dioxide-detecting system in mosquitoes.

339 340

Funding

16

341

X.W. was supported by the Chinese Scholarship Council. Research reported in this publication

342

was supported by the National Institute of Allergy and Infectious Diseases (NIAID) of the

343

National Institutes of Health under award number R01AI095514. The content is solely the

344

responsibility of the authors and does not necessarily represent the official views of the NIH.

345 346

Author contributions

347

P.X., X.W., and W.S.L. performed research; W.S.L. designed research; P.X. and WS.L.

348

analyzed the data; W.S.L. wrote the paper; all authors read and approved the final manuscript.

349 350

Competing financial interest

351

The authors declare no conflict of interest.

352 353

Acknowledgments

354

We thank Dr. Su Liu for comments on a draft version of the manuscript.

355 356

Supplementary information

357

Supplementary figures: Fig. S1-S5

358 359

17

360

References

361

Erdelyan, C.N.G., Mahood, T.H., Bader, T.S.Y., Whyard, S., 2012. Functional validation of the

362

carbon dioxide receptor genes in Aedes aegypti mosquitoes using RNA interference. Insect Mol

363

Biol 21, 119-127.

364

Gillies, M.T., 1980. The role of carbon-dioxide in host-finding by mosquitos (Diptera,

365

Culicidae) - a Review. B Entomol Res 70, 525-532.

366

Grant, A.J., Wigton, B.E., Aghajanian, J.G., O'Connell, R.J., 1995. Electrophysiological

367

responses of receptor neurons in mosquito maxillary palp sensilla to carbon dioxide. J Comp

368

Physiol A Sens Neural Behav Physiol 177, 389-396.

369

Hallem, E.A., Ho, M.G., Carlson, J.R., 2004. The molecular basis of odor coding in the

370

Drosophila antenna. Cell 117, 965-979.

371

Hill, C., Fox, A., Pitts, R., Kent, L., Tan, P., Chrystal, M., Cravchik, A., Collins, F., Robertson,

372

H., Zwiebel, L., 2002. G protein-coupled receptors in Anopheles gambiae. Science 298, 176-

373

178.

374

Jones, W.D., Cayirlioglu, P., Kadow, I.G., Vosshall, L.B., 2007. Two chemosensory receptors

375

together mediate carbon dioxide detection in Drosophila. Nature 445, 86-90.

376

Kwon, J.Y., Dahanukar, A., Weiss, L.A., Carlson, J.R., 2007. The molecular basis of CO2

377

reception in Drosophila. Proc Natl Acad Sci U S A 104, 3574-3578.

378

Liu, M.Z., Vosshall, L.B., 2019. General visual and contingent thermal cues interact to elicit

379

attraction in female Aedes aegypti mosquitoes. Curr Biol 29, 2250-2257 e2254.

18

380

Lu, T., Qiu, Y.T., Wang, G., Kwon, J.Y., Rutzler, M., Kwon, H.W., Pitts, R.J., van Loon, J.J.,

381

Takken, W., Carlson, J.R., Zwiebel, L.J., 2007. Odor coding in the maxillary palp of the

382

malaria vector mosquito Anopheles gambiae. Curr Biol 17, 1533-1544.

383

Majeed, S., Hill, S.R., Dekker, T., Ignell, R., 2017. Detection and perception of generic host

384

volatiles by mosquitoes: responses to CO2 constrains host-seeking behaviour. Roy Soc Open

385

Sci 4.

386

McIver, S., Charlton, C., 1970. Studies on the sense organs on the palps of selected culicine

387

mosquitoes. Can J Zool 48, 293-295.

388

McMeniman, C.J., Corfas, R.A., Matthews, B.J., Ritchie, S.A., Vosshall, L.B., 2014.

389

Multimodal integration of carbon dioxide and other sensory cues drives mosquito attraction to

390

humans. Cell 156, 1060-1071.

391

Ning, C., Yang, K., Xu, M., Huang, L.Q., Wang, C.Z., 2016. Functional validation of the

392

carbon dioxide receptor in labial palps of Helicoverpa armigera moths. Insect Biochem Molec

393

73, 12-19.

394

Reeves, W.C., 1951. Field studies on carbon dioxide as a possible host simulant to mosquitoes.

395

Proc Soc Exp Biol Med 77, 64-66.

396

Reeves, W.C., 1953. Quantitative field studies on a carbon dioxide chemotropism of

397

mosquitoes. Am J Trop Med Hyg 2, 325-331.

398

Robertson, H.M., Kent, L.B., 2009. Evolution of the gene lineage encoding the carbon dioxide

399

receptor in insects. J Insect Sci 9.

400

Smith, E.S., Martinez-Velazquez, L., Ringstad, N., 2013. A chemoreceptor that detects

401

molecular carbon dioxide. J Biol Chem 288, 37071-37081.

19

402

Suh, G.S., Wong, A.M., Hergarden, A.C., Wang, J.W., Simon, A.F., Benzer, S., Axel, R.,

403

Anderson, D.J., 2004. A single population of olfactory sensory neurons mediates an innate

404

avoidance behaviour in Drosophila. Nature 431, 854-859.

405

Syed, Z., Leal, W.S., 2007. Maxillary palps are broad spectrum odorant detectors in Culex

406

quinquefasciatus. Chemical Senses 32, 727-738.

407

van Breugel, F., Huda, A., Dickinson, M.H., 2018. Distinct activity-gated pathways mediate

408

attraction and aversion to CO2 in Drosophila. Nature 564, 420-424.

409

Van Thiel, P.H., Weurman, C., 1947. Attraction exerted on Anopheles maculipennis atroparvus

410

by carbon dioxide. Acta Leiden 18 (1 Vol.), 219-228.

411

Xu, P., Choo, Y.M., Chen, Z., Zeng, F., Tan, K., Chen, T.Y., Cornel, A.J., Liu, N., Leal, W.S.,

412

2019. Odorant inhibition in mosquito olfaction. iScience 19, 25-38.

413

Xu, W., Anderson, A., 2015. Carbon dioxide receptor genes in cotton bollworm Helicoverpa

414

armigera. Sci Nat-Heidelberg 102.

415 416 417 418 419

Figure legends

420

Fig. 1. Responses of Oocytes to Sodium Chloride and Sodium Bicarbonate

421

Representative traces obtained with the following types of oocytes from the same batch and

422

age:

423

(A) Intact (naked) Xenopus oocytes.

20

424

(B) Oocyte coexpressing CquiGR1 and CquiGR3

425

(C) Oocyte coexpressing CquiGR1 and CquiGR2

426

(D) Oocyte coexpressing CquiGR2 and CquiGR3.

427

All scales are 200 nA and 0.5 min. Oocytes were first challenged with 200 mM NaCl and then

428

200 mM NaHCO3.

429 430

Fig. 2. Responses of CquiGR2+CquiGR3-coexpressing oocytes to sodium chloride and

431

sodium bicarbonate

432

(A) Representative trace from a single oocyte preparation stimulated with increasing doses (10

433

– 300 mM) of the two compounds. For clarity, responses to sodium chloride and sodium

434

bicarbonate were colored red and blue, respectively.

435

(B) Concentration-dependent curves obtained with five different oocytes coexpressing

436

CquiGR2 and CquiGR3.

437

Data are presented as mean ± SEM. Some error bars for NaCl curve do not appear, because

438

they are shorter than the size of the symbol. From left to right: 0.81, 2.59, 3.24, 5.35, 6.92,

439

11.24, 9.59, and 14.42.

440 441

Fig. 3. Comparative concentration-dependent curves for oocytes coexpressing either two

442

or three functional subunits

443

Results were obtained with four oocytes coexpressing CquiGR2 and CquiGR3 and four oocytes

444

coexpressing the three GRs. For simplicity, in figures, we use two-letter symbols for proteins as

445

opposed to four-letter symbols (eg, Cq = Cqui) used elsewhere. Data are presented as mean ±

21

446

SEM. Error bars for the green curve do not appear because they are shorter than the size of the

447

symbol. From left to right: 0, 0.63, 3.24, 1.65, and 3.84.

448 449

Fig. 4. Effect of different variants of CquiGR1 on the response of CquiGR2+CquiGR3-

450

coexpressing oocytes

451

All experiments were performed with four replicates with oocytes from the same batch

452

coexpressing either CquiGR2+CquiGR3 or CquiGR2+CquiGR3 plus one of the CquiGR1

453

variants (K456N, M365V, D143G;R434K, and L246F). The clone with sequence identical to

454

that in VectorBase (CJPI 006622) was named wild type (wt). Responses recorded from

455

CquiGR2+CquiGR3-coexpressing oocytes and those recorded from oocytes coexpressing

456

CquiGR2+CquiGR3 plus one CquiGR1variant after stimulus with 200 mM sodium bicarbonate

457

(n = 4) were compared by unpaired t-test. P values are presented on the top of each variant

458

column. Data are presented as mean ± SEM.

459 460

Fig. 5. CO2 concentrations in solutions of sodium bicarbonate prepared in Ringer buffer

461

The calculated concentrations based on the final pH of each preparation (n = 4) are shown in

462

red. The actual concentration of CO2 in each solution was measured with a CO2 Ion Selective

463

Electrode. Data are presented as mean ± SEM.

464 465

Fig. 6. Effect of pH on the oocyte response and the concentrations of CO2 in solution (A)

466

Responses of CquiGR2+CquiGR3-expressing oocytes to 10 mM of sodium bicarbonate without

467

acidification and after acidification (n = 12). The final pH (pHlowered) of the sodium bicarbonate

22

468

buffer was 7.39 ± 0.05. (B) Quantification of CO2 in sodium bicarbonate samples (n = 4)

469

without and after pH adjustment (acidification) as measured with a CO2 Ion Selective

470

Electrode. After acidification (pHlowered), the pH of the sodium bicarbonate samples was 7.32 ±

471

0.03. Data are presented as mean ± SEM.

472 473

Fig. 7. Responses of An. gambiae GRs-expressing oocytes to sodium chloride and sodium

474

bicarbonate

475

(A) Trace obtained with an oocyte coexpressing AgamGR23 and AgamGR24. (B) Responses of

476

another oocyte (from the same batch) co-expressing the three GRs from An. gambiae, GR22,

477

GR23, and GR24. (C) Replication of experiment in (A), but using a different batch of oocytes.

478

(D) Replication of the experiment in (B) with the same batch of oocytes used in (C). ∆ values

479

represent responses to sodium bicarbonate subtracted from responses to sodium chloride at the

480

same concentration. In a third replication with a different batch of oocytes the ∆ values

481

recorded from a GR23+GR24-coexpressing oocyte stimulated with 50, 100, 200, and 300 mM

482

were 120, 225, 250, and 780 nA, whereas the respective values recorded from an oocyte from

483

the same batch and coexpressing GR22+GR23+GR24 were 23, 89, 193, and 270 nA.

484 485

Figs1. NaCl and NaHCO3 responses recorded from oocytes only and different ooocytes

486

coexpressing combination of Culex GRs

487

From top to bottom: naked oocytes (black), oocytes coexpressing GR1+GR3 (red), GR1/GR2-

488

coexpressing oocytes (green), and oocytes coexpressing GR2 and GR3. NaCl and NaHCO3

489

were applied at 200 mM.

23

490 491

Figs2. Effect of CquiGR1 variants on the modulation of CO2 response

492

Traces obtained with the same batch of oocytes injected either with CquiGR2+CquiGR3 or a

493

ternary combination of CquiGR2+CquiGR3 plus one of the CquiGR1 variants. For clarity,

494

responses to sodium chloride and sodium bicarbonate are colored red and blue, respectively.

495 496

Figs3. Responses obtained with a CquiGR2+CquiGR3-expressing oocyte when stimulated

497

with carbon dioxide

498

Samples were prepared by bubbling CO2 in Ringer buffer and adjusting the final pH or by

499

dissolving either sodium chloride or sodium bicarbonate in Ringer buffer. The concentrations of

500

CO2 (in samples prepared by bubbling CO2 or by dissolving NaHCO3) were measured using a

501

CO2 Ion Selective Electrode. Of note, currents elicited by sodium carbonate are larger than

502

expected for the amount of CO2 (aq), because they include currents elicited by sodium.

503 504

Figs4. A continuous trace obtained with an oocyte coexpressing CquiGR2+CquiGR3

505

In this experiment, a freshly prepared perfusion Ringer buffer (pH 7.94) was used. Ringer

506

buffer and 50 mM samples of sodium chloride or sodium bicarbonate at the same pH were

507

applied at the beginning and the end of the experiment. The pH values of Ringer buffers and 50

508

mM NaCl solutions were adjusted to 7, 6.5, 6, and 5.5 before injection. These experiments were

509

repeated six times.

510 511

Figs5. Trace obtained from an oocyte coexpressing CquiGR2 and CquiGR3

24

512

The pH of injected samples (10 mM NaHCO3) was measured before and after acidification.

513 514

Figs6. Responses of Ae. aegypti GRs coexpressed in Xenopus oocytes and challenged by

515

sodium chloride and sodium bicarbonate

516

(A) Traces obtained with an AaegGR2/AaegGR3-coexpressing oocyte. (B) Responses recorded

517

from another oocyte of the same age and the same batch, but coexpressing the subunits

518

AaegGR1, AaegGR2, and AaegGR3. ∆ values represent responses to sodium bicarbonate

519

subtracted from responses to sodium chloride at the same concentration.

520

25

800

Current (nA)

700

NaHCO3

600 500 400 300

NaCl

200 100 0 0

50

100

150

200

Conc.(mM)

250

300

Current (nA)

500

400

300

CqGR2+GR3 200

CqGR1+GR2+GR3 100

0 0

25

50

NaHCO3 [mM]

75

100

6F

R 1L 24

4K

5V

36

43

;R

3G

1M

R

6N

P<0.0001

P<0.0001

P<0.0001

P=0.0048

200

G

14

1D

R

G

G

t

w

1-

100

45

1K

R

G

R

G

3

R

G

2+

R

G

Current (nA)

P=0.0710

250

150

50

0

GR2+GR3+variants of GR1

CO2 (aq) Concentration (in ppm)

250

200

150

100

CO2 (aq) meas. 50

CO2 (aq) clcd.

0 0

50

100

150

200

[NaHCO3] mM

250

300

er ed

w

lo

0

pH

50

al

100

iti

150 20

in

P=0.0005

CO2 (aq) Concentration (ppm)

250

pH

er ed

w

lo

al

iti

200

pH

in

pH

Current (nA) Elicited by 10 mM NaHCO3

35

30

25 P=0.0256

15

10 5

0

Highlights

Xenopus oocytes coexpressing Culex GR1+GR3 were insensitive to sodium bicarbonate

CquiGR2+CquiGR3 gave robust, dose-dependent responses to sodium bicarbonate

CquiGR1 seems to modulate the response of the CquiGR2+CquiGR3 receptor

The active receptors responded to aqueous CO2, not bicarbonate

Similar response profiles were observed with GRs from other mosquito species