Non-canonical odor coding ensures unbreakable mosquito attraction to humans

New York, NY 10065 USA, 3Howard Hughes Medical Institute, New York, NY 10065 USA 4Current address: New York University School of Medicine, 550 First Ave, New York, NY 10016 5Current address: Helen Wills Neuroscience Institute, University of California, Berkeley, Berkeley, CA 94720 USA 6Current address: University of British Columbia, Department of Zoology, Vancouver, BC V6S 0K3 Canada 7These authors contributed equally *Correspondence: meg.younger@gmail.com (M.A.Y.), margo.herre@gmail.com (M.H.), leslie.vosshall@rockefeller.edu (L.B.V.) Non-canonical odor coding ensures unbreakable mosquito attraction to humans


INTRODUCTION
Increased global travel, a growing world population, and rising temperatures are all combining to increase the emergence and transmission of novel disease-causing pathogens.This includes vector-borne diseases, in which pathogens are spread by "vector" organisms such as mosquitoes, ticks, sandflies, and fleas.Diseases spread by these arthropods collectively account for more than 700,000 deaths every year (WHO, 2020).Female Aedes aegypti mosquitoes spread arboviruses including dengue, Zika, yellow fever, and chikungunya.Only female mosquitoes bite, and they do so because they require a blood-meal for reproduction (Allan et al., 1987).Ae. aegypti prefer to bite human hosts, which contributes to their effectiveness as a disease vector (Brown et al., 2014;Gouck, 1972;McBride et al., 2014).To identify human hosts, mosquitoes rely heavily on chemosensory cues, including carbon dioxide (CO2) emitted from breath, and human body odor, which is a mixture of hundreds of different individual odorants including lactic acid and volatile amines such as ammonia (Acree et al., 1968;Bernier et al., 2000;Davis, 1984;Gallagher et al., 2008;Geier et al., 1999;Smallegange et al., 2005;Smith et al., 1970).
Because mosquitoes specialize on humans and require blood to reproduce, the drive to find humans is strong and innate.Indeed, even mosquitoes genetically engineered to eliminate genes critical for peripheral detection of host sensory cues are able to find and bite people.Animals lacking the olfactory receptor co-receptor (Orco), the obligate co-receptor required for the function of the entire family of ORs, show strong attraction to humans (DeGennaro et al., 2013).Deleting the Ir8a co-receptor reduces but does not eliminate attraction to humans (Raji et al., 2019).Similarly, while mosquitoes lacking the obligate CO2 receptor subunit, Gr3, do not respond to CO2 and show impaired behavioral responses in laboratory assays, they are highly effective in finding humans in a more naturalistic semi-field setting (McMeniman et al., 2014).It is not known how the mosquito olfactory system is seemingly infallible in its ability to detect humans.
The cloning of the first odorant receptors by Buck andAxel in 1991 (Buck andAxel, 1991) led to the subsequent discovery that each olfactory sensory neuron expresses a unique odorant receptor that specifies its functional properties.In the well-studied olfactory systems of Drosophila melanogaster flies (Clyne et al., 1999;Gao and Chess, 1999;Vosshall et al., 1999) and Mus musculus mice (Bashkirova and Lomvardas, 2019;Chess et al., 1994), olfactory sensory neurons are thought to express a single olfactory receptor.In both species, neurons expressing a given receptor project axons to dedicated olfactory glomeruli in the first sensory processing center in the brain, the antennal lobe in insects (Couto et al., 2005;Fishilevich and Vosshall, 2005;Vosshall et al., 2000) and the olfactory bulb in vertebrates (Mombaerts et al., 1996;Ressler et al., 1994;Vassar et al., 1994).This "one-receptor -to -one -neuron -to -oneglomerulus" organization is believed to be a widespread motif in both invertebrate and vertebrate olfactory systems, and the convergence onto discrete glomeruli is hypothesized to permit the brain to parse which subpopulation of olfactory neurons is activated by a given odor (Rubin and Katz, 1999;Uchida et al., 2000;Wang et al., 2003).
Consistent with this organization, the number of expressed chemosensory receptors in the OR and IR gene families roughly correlates to the number of olfactory glomeruli.This holds true in Drosophila melanogaster flies (~60 receptors/~55 glomeruli) (Benton et al., 2009;Laissue et al., 1999;Robertson et al., 2003), the honey bee Apis mellifera (~180 receptors/~160 glomeruli) (Flanagan and Mercer, 1989;Robertson et al., 2010), and the tobacco hornworm Manduca sexta (~60 receptors/~70 glomeruli) (Grosse-Wilde et al., 2011).Based on these studies, it is widely thought that merely counting the number of antennal lobe glomeruli in a new species would be reasonably predictive of the number of chemosensory receptors found in its genome.In Ae. aegypti, however, there is a striking mismatch between the number of expressed chemosensory receptors and the number of antennal lobe glomeruli, with at least twice as many receptors as available glomeruli (Bohbot et al., 2007;Ignell et al., 2005;Matthews et al., 2018;Shankar and McMeniman, 2020;Zhao et al., 2020).How is the mosquito olfactory system organized to accommodate so many receptors and does this departure from rules established in other species explain their exquisite ability to locate human hosts?
In this study, we developed a CRISPR-Cas9-based genetic knock-in strategy in Ae. aegypti to generate a suite of genetically-modified mosquito strains that label molecularly distinct populations of olfactory sensory neurons.We used these strains to understand how human odor is encoded in the mosquito olfactory system.Combining anatomical analysis of these lines with extensive RNA in situ expression analysis, we found that the olfactory system of Ae. aegypti does not have the expected "one-receptor-to-one-neuronto-one-glomerulus" organization seen in other organisms.We frequently observed co-expression of multiple chemosensory receptors from at least two of the three receptor gene superfamilies within individual olfactory sensory neurons.Antennal lobe glomeruli were often innervated by olfactory sensory neurons expressing different receptors.We hypothesized that this unconventional organization underlies the robust, seemingly unbreakable, properties of the Ae.aegypti olfactory system in detecting human odor.To test this, we examined a minimal mixture that drives hostseeking behavior, the combination of CO2 and lactic acid.We found that the neurons that sense CO2 express not only the CO2 receptor gene Gr3, but also members of the IR family of chemosensory receptors that detect acids and amines.Using in vivo functional calcium imaging we found that these CO2 neurons also sense volatile amines, consistent with their co-expression of Gr3 and IRs.Moreover, triethyl amine can substitute for CO2 to drive attraction to lactic acid in mutants that cannot sense CO2.This sensory organization, in which multiple receptors that respond to very different types of chemosensory stimuli are co-expressed, suggests a redundancy in the code for human odor.We speculate that this design supports the robust human host-seeking seen in this olfactory specialist.

Mismatch in chemosensory receptor and olfactory glomerulus number suggests a novel olfactory organization
In the mosquito, olfactory cues are sensed by olfactory sensory neurons in the antenna and the maxillary palp, whose axons project to the ipsilateral antennal lobe of the brain (Distler and Boeckh, 1997;Ignell et al., 2005) (Figure 1A-B,  S1A-C).The antennal lobe, the insect equivalent of the vertebrate olfactory bulb, is organized into discrete olfactory glomeruli in which axons from peripheral olfactory sensory neurons terminate and synapse with local interneurons and projection neurons that relay olfactory information to the higher brain (Stocker, 1994).Previous studies used morphological criteria to define 50 (Ignell et al., 2005), 60 (Zhao et al., 2020), or 81 (Shankar and McMeniman, 2020) discrete olfactory glomeruli in the female Ae.aegypti antennal lobe.In this study, we define approximately 65 olfactory glomeruli (64.9 ± 0.9, mean ± SEM), obtained by counting antennal lobe glomeruli in the left hemisphere of 12 female Ae.aegypti brains stained to reveal synaptic neuropil (Figure 1B,I,K, S2-5).The glomerulus count ranged from 60-72 glomeruli per antennal lobe, indicating a high level of variability in the organization of the antennal lobe (see Methods).We generated 3-D reconstructions of complete antennal lobes and saw considerable variability in the size and shape of the glomeruli (Figure S1).We were able to consistently identify certain landmark glomeruli, most notably the three glomeruli that are innervated by the maxillary palp (Ignell et al., 2005;Shankar and McMeniman, 2020) The canonical "one-receptor-to-one-neuron-to-one glomerulus" organization posits that the number of chemosensory receptors should roughly match the number of glomeruli in the antennal lobe (Figure 1D).While there is not yet a clear consensus on the number of olfactory glomeruli in Ae. aegypti, it ranges from 50 to 81.How does this relate to the number of chemosensory receptors expressed?In the updated Ae. aegypti genome (Matthews et al., 2018), there are 117 OR, 135 IR, and 72 GR genes for a total of 324 structural genes that could function in the olfactory system (Figure 1E).We have reanalyzed published RNA expression data (Matthews et al., 2016) using multiple thresholds to estimate the number of receptors expressed in the antenna and maxillary palp.Even at the conservative threshold of 5 transcripts per million (TPM), the mosquito olfactory system expresses 102 chemosensory receptors, and moving the threshold to 2, 1, or 0.5 TPM increases the number of receptors plausibly expressed to 134, 156, and 178, respectively (Figure 1F,G, Data File 1).Thus, there are many more chemosensory receptors expressed in the olfactory system than available antennal lobe glomeruli, suggesting that the organization of the Ae.aegypti olfactory system must differ from the canonical scheme.We speculate that the mismatch can be resolved by expressing multiple receptors per neuron or having multiple different types of neurons co-converge on a single glomerulus or both (Figure 1H).
To begin to distinguish between these two organizational principles, we generated a collection of CRISPR-Cas9 gene-targeted strains that label subpopulations of olfactory neurons using the Q-system, a binary expression system similar to Gal4/UAS (Brand and Perrimon, 1993) that uses cell type-specific expression of the QF2 transcription factor to induce expression of an effector from the QF2 binding QUAS enhancer (Potter et al., 2010).We introduced an inframe insertion that replaced the stop codon of each of the co-receptors Orco, Ir25a, Ir8a, and Ir76b, as well as the CO2 receptor subunit Gr3 with the transcription factor QF2 (Figure 1I, S2-S6.See Data File 1 for a full description of all genotypes by figure) (Matthews et al., 2019;Potter et al., 2010;Riabinina et al., 2016).These five new gene-sparing knock-in strains were designed to cause minimal disruption to the locus to increase the likelihood that they would faithfully report expression of the endogenous gene.We crossed these QF2 driver lines individually to a QUAS-CD8:GFP reporter to label neuronal membranes and visualized axonal projection patterns in the antennal lobe.
Orco, Ir25a, Ir8a, and Ir76b co-receptor driver lines were expressed in olfactory sensory neurons with distinct projection patterns in the antennal lobe (Figure 1I-K).Unexpectedly, neurons that expressed Ir25a projected to almost all of the glomeruli in the antennal lobe (89.9 ± 1.4%, mean ± SEM, n = 3)(Figure 1I-K, S3), and expression overlapped extensively with glomeruli labeled by Orco (Figure 1I-K, S2, S3).While these co-receptor driver lines labeled glomeruli in the same regions from brain to brain, the interindividual expression patterns were not identical, consistent with the variability in glomerular anatomy that we have observed (see Methods).Neurons that detect CO2 are located in the maxillary palp (Grant et al., 1995;Lu et al., 2007;Omer and Gillies, 1971) and we saw that Gr3-expressing neurons projected to a large glomerulus in the posterior antennal lobe, Glomerulus 1 (Figure 1I-K, see Methods for full discussion of glomerulus nomenclature) which is also innervated by Ir25a-expressing neurons.We also noted the presence of a second small glomerulus that was often innervated by Gr3-expressing neurons in the antenna (Figure S6B).These initial findings point to the overlap of OR-, IR-, and GR-expressing neurons in the antennal lobe of Ae. aegypti, which are believed to remain segregated in Drosophila melanogaster.

Co-expression and co-convergence of Orco and Ir25a in the mosquito olfactory system
The high degree of overlap between glomeruli labeled by Orco-and Ir25a-expressing olfactory sensory neurons suggests that there is either widespread Orco and Ir25a co-expression within individual sensory neurons or that Orco and Ir25a are expressed in different neurons whose axons coconverge onto individual antennal lobe glomeruli or both (Figure 1H).To distinguish between these possibilities, we adapted the Split-QF2 system (Riabinina et al., 2019) for use in the mosquito.This system "splits" the transcription factor QF2 into two components, the DNA binding domain (QF2-DBD) and the activation domain (QF2-AD) each tagged with a synthetic leucine zipper (Figure 2A,B).When both the QF2-DBD and QF2-AD are co-expressed in the same cell, the two domains associate via the leucine zipper, reconstitute a functional QF2 protein, initiate transcription at the QUAS enhancer, and drive expression of a reporter gene (Figure 2C).
Using the same stop-codon replacement approach that we used to generate the QF2-lines, we inserted the QF2-AD into the Ir25a locus (here referred to as IR25a-QF2-AD) and the QF2-DBD into the Orco locus (here referred to as Orco-QF2-DBD).When either IR25a-QF2-AD or Orco-QF2-DBD was used to drive expression of dTomato, we did not see fluorescence in the antenna, maxillary palp, or the antennal lobe (Figure 2D-F, S7).Therefore, neither QF2-DBD nor QF2-AD alone can activate expression from the QUAS enhancer.However, when Orco-QF2-DBD and IR25a-QF2-AD were crossed into the same animal, we saw expression of dTomato in antennal and maxillary palp neurons, as well as axonal projections in the antennal lobe (Figure 2D-F, S7, S8).Nearly half of the glomeruli in the antennal lobe were labelled with dTomato (Figure 2G-I, S7, S8).This points to widespread Orco and Ir25a co-expression within Ae. aegypti olfactory sensory neurons.
While examining these antennal lobes we consistently noticed that a cluster of anterior dorsal glomeruli was unlabeled when the Orco-QF2-DBD, IR25a-QF2-AD combination was used to drive expression in animals.However, these glomeruli were labelled when either Orco-QF2 or IR25a-QF2 were used to drive expression (Figure 2J,K).This observation strongly suggests that separate populations of Orco-positive and Ir25a-positive neurons co-converge onto these glomeruli.Therefore, the mosquito olfactory system shows evidence of both co-expression and co-convergence, motifs that are not typically seen in the conventional model organisms Drosophila melanogaster flies or Mus musculus mice (Figure 2L).It is also possible that a given glomerulus shows both co-convergence of molecularly distinct neurons while also receiving projections from neurons co-expressing multiple chemosensory co-receptors.

Extensive co-expression of chemosensory co-receptors in the antenna
The observation that nearly half of antennal lobe glomeruli receive projections from neurons co-expressing Ir25a and Orco in gene-targeted strains suggested that there is extensive co-expression of the IR and OR chemoreceptor families.To confirm the existence of such co-expressing neurons, we carried out whole mount antennal immunostaining with an antibody to the endogenous Orco protein and GFP expressed from each sensory neuron driver.We observed extensive co-expression of Orco and Ir25a, with substantially fewer cells co-expressing either Orco and Ir8a or Orco and Ir76b (Figure 3A-D), even after accounting for fewer total Ir76b and Ir8a cells (Figure 3B-D, Data File 1).
To confirm and extend these results, we performed RNA in situ hybridization in wild-type antennae with probes designed to target endogenous Orco, Ir76b, and Ir25a transcripts (Figure 3E).These experiments replicated patterns of co-expression observed in immunostained antennae (Figure 3A-D), with almost half of Orco cells co-expressing Ir25a, and few Orco cells co-expressing Ir76b (Figure 3F-H, Data File 1) indicating that widespread co-expression is not an artifact of the genetic reporter lines.
By examining the expression pattern of these three transcripts simultaneously, we were also able to elucidate the relationship between Ir25a and Ir76b expression.In Drosophila melanogaster, Ir76b is co-expressed with Ir25a in olfactory sensory neurons in the antenna and in gustatory neurons in the labellum (Abuin et al., 2011;Lee et al., 2018).In Ae. aegypti, we found that 100% of Ir76b-expressing neurons in the antenna also express Ir25a (Figure 3F,H).We also observed that in addition to widespread co-receptor coexpression, some mosquito olfactory neurons express just one co-receptor (Figure 3I).Our data demonstrate not only extensive co-expression between Orco and IR co-receptors, but also complexity in the rules that govern receptor expression in antennal olfactory sensory neurons.

Coordinated co-expression of chemosensory receptors in the maxillary palp
We have documented extensive chemosensory co-receptor co-expression, but to form functional odorant-sensitive IR or OR complexes, olfactory sensory neurons must express both co-receptors and ligand-sensitive receptors (Abuin et al., 2011;Benton et al., 2009;Larsson et al., 2004;Neuhaus et al., 2005).To simultaneously monitor the extent of co-expression of both co-receptors and ligand-sensitive receptors, we carried out multiplexed whole mount RNA in situ hybridization (Choi et al., 2018) in the maxillary palp, the olfactory organ that detects the potent host cue CO2 as well as other host odors (Grant et al., 1995;Lu et al., 2007;McMeniman et al., 2014;Omer and Gillies, 1971) (Figure 4A,B).The maxillary palp expresses many fewer chemosensory receptors genes than the antenna, with 18 receptors detected at the 1 TPM threshold in the maxillary palp compared to 138 in the antenna at the same threshold, simplifying the task of selecting genes for expression analysis (Figure 1F,G).We performed RNA in situ hybridization with probes for 10 of the 18 chemosensory co-receptors and ligand-sensitive receptors that were present in maxillary palp RNA-seq at a threshold of TPM>1.This technique visualized gene expression with sufficient sensitivity that even Or71 and Ir75g (present at 1.93 and 1.67 TPM, respectively) were readily detected (Figure 4C-I, S9).
We found no overlap in expression of Orco and Gr3 in the maxillary palp, but Ir25a was expressed in all Orco and all Gr3 cells (Figure 4D).Previous work in Anopheles gambiae suggested that Orco-expressing neurons in the maxillary palp can be evenly divided into two non-overlapping groups: an Or8 population and an Or49 population (Lu et al., 2007).We show definitively that Or8 and Or49 are in segregated populations of Orco-expressing neurons in Ae. aegypti (Figure 4E), and when combined with the results of the previous experiment (Figure 4D), that these cells are also all Ir25apositive.Additional RNA in situ hybridization experiments revealed that Or8-and Or49-expressing cells also often express Ir76b, with a bias towards expression in Or8-  expressing cells (Figure 4F, S9, Data File 1).Taken together these data show that olfactory sensory neurons express coreceptors Orco, Ir25a, Ir76b and either of the ligand-sensitive subunits Or49 or Or8.
When we analyzed IR ligand-sensitive subunit expression, we found that Ir100a and Ir93a are also expressed in a subset of Or49-expressing neurons (Figure 4G,H, S9, Data File 1), suggesting that these cells can form functional OR and IR complexes with their respective co-receptors in the same neuron.Further, we identified co-expression of Or71 and Or49, demonstrating that more than one ligand-sensitive OR can also be expressed in an olfactory sensory neuron in Ae. aegypti (Figure S9O).Lastly, we found that 65% of Gr3-expressing cells also express Ir75g.Taken together with the finding that Gr3-expressing cells express Ir25a (Figure 4D, 4I, S9N, Data File 1) this raises the possibility that a functional IR complex can be found in CO2-sensitive neurons.

CO2 receptor
We set out to investigate the functional relationships of these extensive receptor co-expression patterns by focusing on the maxillary palp, which is both simpler in organization than the antenna and also uniquely important because it is the sensory appendage that detects CO2.
To delineate the organization of maxillary palp projections in the brain, we used our QF2 and Split-QF2 driver lines to examine sensory innervation of antennal lobe glomeruli (Figure 5A-H).Glomerulus 2 and Glomerulus 3 received input from both Orco-expressing neurons and Ir25a-and Ir76bexpressing neurons (Figure 5A-G), consistent with data presented in Figure 4. Co-expression of Orco and Ir25a in neurons that project to these two glomeruli was confirmed using the Split-QF2 system (Figure 2).In Orco-QF2-DBD, IR25a-QF2-AD animals, Glomerulus 2 and Glomerulus 3 were labeled, but Glomerulus 1 was not (Figure 5H).We also discovered that Glomerulus 1, which is the largest glomerulus in the antennal lobe (Shankar and McMeniman, 2020), received input from Gr3-expressing sensory afferents.Glomerulus 1 was also innervated by Ir25a-expressing sensory neurons (Figure 5A-H), consistent with the co-expression of Gr3 and Ir25a in olfactory neurons (Figure 4D).This finding raised the possibility that Glomerulus 1 may respond to additional odorant ligands independent of CO2-sensitive Gr3.
Given the extensive receptor co-expression observed in the maxillary palp (Figure 4,5I), we next explored the possibility that functional expression of multiple receptor types would allow maxillary palp CO2-sensitive neurons to respond to additional odorants.We set out to examine odorants detected by maxillary palp neurons by developing an in vivo calcium imaging preparation with the genetically-encoded calcium sensor GCaMP6s (Chen et al., 2013), allowing us to monitor stimulus-evoked activity in the antennal lobe of a mosquito as it was exposed to different odorants or to CO2 (Figure 5J).We imaged the axon terminals of maxillary palp sensory neurons in the antennal lobe with two-photon microscopy using a custom-built olfactometer to deliver odorants and CO2 within a continuous air stream.We expected that Glomerulus 1 would respond to CO2 because it receives projections from Gr3-expressing neurons.Indeed, in preliminary experiments when we imaged GCaMP6s responses in axonal termini of Gr3-expressing neurons in Glomerulus 1, we observed low fluorescence at baseline and concentration-dependent responses to CO2, with no apparent desensitization across trials (data not shown).
To image activity simultaneously in CO2-sensitive Glomerulus 1 and the non-CO2-sensitive Glomerulus 3, we used the Ir25a>GCaMP6s strain, which expresses GCaMP6s in all maxillary palp olfactory neurons (Figure 4).We saw consistent activation of Glomerulus 3 by 1-octen-3-ol and 3-octanone (Figure 5K,L), odorants that have been shown in in vitro heterologous expression experiments to activate Or8-Orco from both Ae.aegypti and Anopheles gambiae (Bohbot and Dickens, 2009;Lu et al., 2007).Therefore Glomerulus 3 is likely to be innervated by Or8-expressing neurons (Fig- ure 4E).From a panel of candidate odorants, we found that Glomerulus 3 also responded to two volatile amines, hexyl amine and triethyl amine (Figure 5K,L).Volatile amines, including polyamines, have been proposed to be IR ligands in Drosophila melanogaster (Geier et al., 1999;Hussain et al., 2016;Min et al., 2013;Silbering et al., 2011).These findings are consistent with the hypothesis that IRs and ORs are functionally co-expressed in maxillary palp neurons and that this co-expression enables these cells to respond to ligands that activate both classes of receptors.
We next turned to the CO2-sensitive Glomerulus 1 (Figure 5M) and imaged stimulus-evoked activity in Gr3 heterozygotes, which have no defects in CO2 sensation, and in Gr3 homozygous mutants, which lose all physiological and behavioral responses to CO2 (McMeniman et al., 2014).We used Gr3 heterozygotes as the control genotype in these experiments because this mutation was introduced in a different wild-type strain than our QF2/QUAS reagents and this genetic configuration better controls for genetic background (see Methods).As expected, we saw robust activation of Glomerulus 1 by CO2 in Gr3 heterozygotes, and a complete loss of response to CO2 in Gr3 homozygous mutant animals (Figure 5N-P).CO2-sensing neurons in the maxillary palp respond to multiple odorants in Aedes, Culex, and Anopheles mosquitoes (Lu et al., 2007;Tauxe et al., 2013;Turner et al., 2011) and it has been proposed that Gr3 is a broadlytuned receptor that responds to many odorants.We examined the response to a recently identified CO2-neuron activator, acetone (Ghaninia et al., 2019).This odorant activated Glomerulus 1 in the Gr3 heterozygote but the response to acetone was abolished in the homozygous Gr3 mutant (Figure 5Q,R).This suggests that the CO2 receptor can interact with non-CO2 ligands.We examined the response to volatile amines, and saw consistent responses to hexyl amine and triethyl amine (Figure 5Q,R).The response to both of these volatile amines was unaffected in Gr3 homozygous mutant animals (Figure 5Q,R).We speculate that hexyl amine and triethyl amine are activating the CO2-sensitive neuron by stimulating IRs expressed in this neuron,

Volatile amines substitute for CO2 in activating mosquito attraction to humans
Female mosquitoes integrate multiple sensory cues to drive strong attraction to humans.Human odor is a blend of hundreds of volatile odorants, including lactic acid and volatile amines (Acree et al., 1968;Bernier et al., 2000;Dormont et al., 2013).While body heat, CO2 emitted in the breath, and individual odorants produced by human skin individually elicit weak to no mosquito attraction, the combination of multiple sensory cues leads to robust attraction in laboratory behavioral assays (Corfas and Vosshall, 2015;Dekker et al., 2005;McMeniman et al., 2014).Among these cues, CO2 is the most potent synergistic stimulus, sensitizing mosquitoes to gate attraction to both heat and odorants (Dekker et al., 2005;McMeniman et al., 2014).The discovery that triethyl amine and hexyl amine activate Glomerulus 1 in Gr3 homozygous mutants that are unable to detect CO2 led us to hypothesize that volatile amines could substitute for CO2 in sensitizing Gr3 mutant mosquitoes to human body odor (Figure 6A).
To test this, we used a behavioral assay (Basrur et al., 2020) to measure mosquito attraction to various stimuli (Figure 6B).In control experiments, we baited the assay with either CO2 or lactic acid, a human host volatile, or both CO2 and lactic acid combined, and monitored how many mosquitoes accumulated in the trap.As previously shown, either CO2 or lactic acid alone only weakly attracted mosquitoes, but the combination of lactic acid with CO2 caused potent dose-dependent attraction in two different wild-type strains of Ae. aegypti (Figure 6C, Data File 1) (Acree et al., 1968;Bernier et al., 2000;Davis, 1984;Geier et al., 1999;Majeed et al., 2014;McMeniman et al., 2014;Smith et al., 1970).Female mosquito attraction to humans is suppressed for 72-96 hours after ingestion of a blood-meal (Davis, 1984;Duvall et al., 2019;Klowden, 1981).We saw the same state-dependent suppression of attraction to CO2 and lactic acid by bloodfed mosquitoes, confirming that these stimuli serve as a proxy for the attraction to human host odor (Figure 6C).
We next tested the hypothesis that activating Glomerulus 1 with volatile amines in CO2-insensitive Gr3 mutants could sensitize these animals to respond to lactic acid.In control experiments, we observed that neither wild-type nor Gr3 mutants were attracted to CO2 alone, and that only wild-type animals and not Gr3 mutants showed synergistic attraction to CO2 and lactic acid (Figure 6D).As previously observed, both wild-type and Gr3 mutants showed strong attraction to a live human host in the presence of CO2 (McMeniman et al., 2014).Live human hosts emit many sensory cues, effectively compensating for the inability of Gr3 mutant mosquitoes to detect CO2 (McMeniman et al., 2014).
Both CO2 and triethyl amine activate Glomerulus 1 in wildtype animals but in Gr3 mutants only triethyl amine and not CO2 activates this glomerulus.We asked if triethyl amine alone or in combination with CO2 was attractive to wild type or Gr3 mutant mosquitoes.Neither wild type nor Gr3 mutant mosquitoes showed any attraction to CO2 alone, triethyl amine alone, or the combination of both (Figure 6E).This was expected because CO2 alone-and by extension the activation of Glomerulus 1-is insufficient to mediate attraction in this assay.We next asked if triethyl amine can substitute for CO2 in synergizing with lactic acid to attract Gr3 mutant mosquitoes.We tested a range of concentrations of triethyl amine with a single concentration of lactic acid and quantified the number of attracted mosquitoes.Lactic acid alone and both low and high concentrations of triethyl amine were not attractive to Gr3 mutants.However, moderate concentrations of triethyl amine stimulated strong attraction when supplied together with lactic acid (Figure 6F).This sensitization to lactic acid triggered by triethyl amine was not further enhanced by the addition of CO2, a stimulus that Gr3 mutants do not detect (Figure 6G).We speculate that CO2insensitive Gr3 mutants rely on a compensatory pathway that uses Ir25a and unknown ligand-sensitive IRs to activate Glomerulus 1 with volatile amines.
To ask if this compensatory pathway is active in wild-type mosquitoes, we repeated these experiments and were surprised to discover that the concentrations of triethyl amine that stimulated Gr3 mutant attraction to lactic acid had no effect on wild-type animals (Figure 6H,I).These results suggest that when mosquitoes lack receptors for potent host cues such as CO2, they rely more heavily on redundant cues detected by the same neurons via the expression of a diverse array of chemosensory receptors to ensure that hostseeking behavior is not compromised (Figure 6J).

DISCUSSION
A compensatory pathway promotes attraction behavior in the CO2 receptor mutant It has only recently become possible to generate targeted mutations in chemosensory genes in Ae. aegypti, and a striking finding from studies that generated such mutants is how difficult it is to disrupt human host-seeking.Mutations in the co-receptor Orco (DeGennaro et al., 2013) and the coreceptor Ir8a (Raji et al., 2019) both eliminate entire classes of ligand-specific receptors.In spite of this, Orco mutants show normal attraction to humans and mutations in Ir8a reduce but do not abolish mosquito attraction to humans.Gr3 mutants are completely unable to sense CO2 yet they can also find human hosts, even when released into a large semi-field environment (McMeniman et al., 2014).These findings raise the question: Why is it so difficult to break the mosquito olfactory system and prevent mosquitoes from finding humans?We propose that this is due to redundancy in receptor expression at the periphery.A recent paper (Zhao et al., 2020) examined Ae. aegypti antennal lobe responses of Orco-expressing neurons to physiological concentrations of odor collected from several individual humans, a variety of non-human animals, and floral odors.They identified a glomerulus that responded to odor blends from all human individuals, animals, as well as honey and milkweed odor.It is tempting to speculate that this glomerulus is innervated by olfactory sensory neurons that co-express multiple receptors.We have probed IR and GR redundancy in the CO2 system and found that Gr3 mutant mosquitoes that do not detect CO2 can still detect triethyl amine, and have demonstrated that triethyl amine can act as a behavioral substitute for CO2 in driving attraction to lactic acid in Gr3 mutants.This activation was strongly dependent on the dose of triethyl amine.Although human skin releases a large number of odorants, including volatile amines, we do not know how the concentrations used in our laboratory behavioral assays relate to those produced by human skin.
Surprisingly, this substitution of triethyl amine for CO2 does not drive attraction in wild-type animals.One possible mechanism to explain this finding might be that in the absence of Gr3, the sensitivity of the circuit shifts to enhance the behavioral effect of compounds that activate the IRs co-expressed in the CO2-sensitive neuron.If this were the case, it remains to be seen how the CO2-sensing circuitry is altered, whether it is the sensory neuron or downstream neurons involved in host-seeking, and if changes occur at the level of the receptor, synapse, or the intrinsic firing properties of the sensory neurons themselves (Marder and Goaillard, 2006).CO2sensing neurons in the maxillary palp rapidly reset their intrinsic firing to adjust to changes in background levels of CO2 (Majeed et al., 2014) and perhaps the absence of background CO2 sensation in the Gr3 mutant enhances the signal to noise ratio of other odorants.

Coordinated co-expression between IR, OR, and GR ligand-sensitive receptors
We identified co-expression of co-receptors and ligand-sensitive receptors belonging to distinct chemosensory families in single neurons in the maxillary palp.This co-expression poses a gene regulatory problem for an olfactory neuron.For ORs and IRs to form functional chemoreceptors, at least one co-receptor and one ligand-sensitive receptor must be expressed in a cell.We have demonstrated that two IRs, Ir93a and Ir100a, are expressed nearly exclusively with Or49 and the co-receptors Orco and Ir25a, whereas Ir75g is co-expressed with the coreceptor Ir25a and the CO2 receptor subunit Gr3.Thus, the transcriptional landscape in Ae. aegypti maxillary palp olfactory neurons is not only permissive to co-expression, but ensures certain receptors are only expressed with others.
How might this complex code of chemosensory receptor coexpression be regulated?In vertebrates, an elaborate epigenetic silencing mechanism ensures that each olfactory neuron expresses only a single allele of a single odorant receptor (Bashkirova and Lomvardas, 2019).In contrast, Drosophila is thought to use a more conventional transcription factor code in which the specification of a neuron and the expression of its chemosensory receptor is tightly regulated (Jafari and Alenius, 2015;Li et al., 2016;Ray et al., 2008).Single-cell sequencing data generated from developing Drosophila melanogaster olfactory neurons demonstrates a complex regulatory landscape wherein dependent on cell type, a set of transcription factors govern receptor expression, axon targeting, or both (Li et al., 2020).Two recent studies (McLaughlin et al., 2020;Task et al., 2020) also document extensive co-expression of receptors in Drosophila melanogaster, calling into question the rules that regulate olfactory organization in this insect.It is yet to be determined if mosquito orthologues of these transcription factors have been co-opted to regulate the co-expression we observe or if this novel olfactory organization demands a distinct transcriptional mechanism.
The maxillary palp has been implicated in responses to CO2 (Acree et al., 1968;Gillies, 1980;Grant et al., 1995), thermosensation (Roth, 1951), mechanosensation (Bohbot et al., 2014), detection of attractive monomolecular odorants such as 1-octen-3-ol (Syed and Leal, 2007;Takken and Kline, 1989;Vythilingam et al., 1992), as well as blends of odorants extracted from human hosts (Tauxe et al., 2013).Maxillary palp transcriptomes offer insights into how these tiny sensory organs detect a broad range of stimuli (Bohbot et al., 2014;Matthews et al., 2018), pointing to a transcriptional landscape that is difficult to reconcile with the prior understanding of neuronal identity in the maxillary palp.It was thought that all basiconic sensilla, which detect volatile odorants, house the sensory dendrites of three neurons that form identical repeating units: one large CO2-sensitive neuron that expresses Gr3, and two smaller neurons that express either an Orco-Or8 receptor or an Orco-Or49 receptor (Lu et al., 2007;McIver, 1972).How do these repeating units of just three neurons enable the detection of so many stimuli and accommodate the expression of so many chemosensory genes?We demonstrate through multiplexed RNA in situ hybridization experiments that the receptor composition of these neurons is far more complex, and they can be subdivided into many more than three cell types.We did not evaluate the relationship of every combination of chemosensory genes expressed in the maxillary palp and hypothesize that comprehensive gene expression profiling including the application of single cell RNA-sequencing will reveal even more cell types.The extensive co-expression of IRs, ORs, and GRs documented in this work likely enables maxillary palp neurons to detect volatile odorants belonging to multiple chemical classes and may explain how CO2-sensitive neurons also respond to human skin odor (Dormont et al., 2013;Tauxe et al., 2013), and why mosquitoes unable to detect CO2 can still locate a human host (McMeniman et al., 2014).

Receptor co-expression as a mechanism for redundancy or blend integration
We hypothesize that receptor co-expression is used broadly to detect redundant cues that are present in human odor, a blend that can vary from individual to individual and contains hundreds of different chemicals (Bernier et al., 1999;Bernier et al., 2000).We have not yet explored the role of OR-IR coexpression or the expression of multiple IRs or ORs in single olfactory neurons.It is possible that receptor co-expression is always used to form a highly redundant detection system for different cues that represent the same ecological target: human hosts.This motif has the benefit of limiting the number of neurons needed to detect varied odorants with the same meaning, however, in exchange it may sacrifice the ability to distinguish between cues detected by receptors expressed in the same sensory neurons.In Drosophila melanogaster there are rare cases of OR-OR or OR-IR co-expression, which is the exception rather than the rule.Or35a is co-expressed with Ir76b (Silbering et al., 2011), and while these neurons respond to many odorants (Silbering et al., 2011;Yao et al., 2005), the role of co-expression remains unknown.In Drosophila melanogaster, Or49a and Or85f are co-expressed in a specific olfactory sensory neuron population where they play redundant roles in predator avoidance (Ebrahim et al., 2015).Intriguingly, recent data from Drosophila melanogaster also show Orco and Ir25a co-expression in the antenna and maxillary palp (McLaughlin et al., 2020;Task et al., 2020), suggesting that broad co-receptor co-expression may be a conserved motif in insects and upending previous data that suggested strict segregation of OR and IR sensory neuron populations in flies.
While extensive co-expression of different classes of olfactory receptors within olfactory neurons documented in this study is a novel organizational principle, there are counterexamples of receptor co-expression within a single class of chemosensory receptor.Karner et al. have documented polycistronic expression of multiple odorant receptors in Anopheles gambiae sensory neurons (Karner et al., 2015).The nematode C. elegans copes with a very large number of chemosensory receptor genes and a very small number of sensory neurons by extensive receptor co-expression (Troemel et al., 1995;Vidal et al., 2018).
What is the functional significance of co-expression of different classes of olfactory receptors, the ORs, IRs, and GRs, in Aedes aegypti?It is possible that co-expression is exploited to integrate the detection of multiple odors in the sensory neuron itself.In Drosophila melanogaster, IRs that form functional complexes can depolarize resting membrane potential in vitro (Ai et al., 2013).If the same phenomenon is occurring in Ae. aegypti neurons that express both IR and OR complexes, it is possible that an IR complex could render a neuron more sensitive to OR ligands.Certain odors repress OR activity relative to baseline in Drosophila melanogaster (Hallem and Carlson, 2006;Yao et al., 2005) and it is possible that multiple receptors housed in a single neuron could be a source of blend-specific repression.Individual odorant molecules found in human odor exist in many blends and this might be a way to limit the attraction to monomolecular components of human odor that are found in aversive blends or are produced by other animals.Recent data in Drosophila show that, in maxillary palp sensilla innervated by Orco neurons, responses to several odorants change in Ir25a mutant flies (Task et al., 2020).The direction of the response is sensilla-specific.In some neurons, responses are higher in Ir25a mutants, in other neurons, responses are lower, suggesting that Ir25a may be implicated in gain control in Orco neurons (Task et al., 2020).These ideas remain to be tested in the mosquito, but it is tempting to speculate that this could be used as a method to detect ratios of odors present in a blend, a key determinant in how attractive many monomolecular compounds are to mosquitoes (Bernier et al., 2007).
We demonstrate that Ae. aegypti mosquitoes employ both receptor co-expression in primary sensory neurons as well as co-convergence of neurons that express distinct classes of chemosensory receptors onto the same glomerulus.Consistent with these data, we also identified olfactory neurons in the antenna that express ORs without IRs, IRs without ORs, and neurons that express both ORs and IRs.What is the functional difference between integrating odorant information in the primary sensory neuron versus at the first synapse in the antennal lobe?Co-convergence of olfactory sensory neurons onto the antennal lobe could allow for the early integration of olfactory cues while still retaining discrete input channels that could be selectively modulated during changes in behavioral state, such as the suppression of host-seeking after a blood meal.The role of co-convergence in this system remains to be tested.
In summary, our study reveals unexpected complexity in the gene expression and functional organization of the mosquito olfactory system.The potential redundancy afforded by a system in which many neurons express receptors that respond to non-overlapping sets of odorant ligands may greatly increase the sensitivity and robustness of the mosquito olfactory system.Future attempts to refine the design of repellents to ward off mosquitoes or attractant traps to lure them will have to reckon with the complexity of this system.troport olfactometer; Javier Marquina-Solis for antennal lobe tracing assistance; Christina Pyrgaki, Carlos Rico, Katarzyna (Kate) Cialowicz, and Alison North at the Rockefeller Bio-Imaging Resource Center for assistance with confocal imaging; Daniel Gross, James Petrillo, and Peer Strogies at the Rockefeller Precision Instrumental Technologies (PIT) resource center for advice and fabrication of the olfactometer and other imaging components used in the GCaMP imaging experiments; Olena Riabinina and Christopher Potter for providing advice and unpublished Q-system reagents; Carolyn McBride, Matthew DeGennaro and members of the Aedes Toolkit Group for advice and discussion; Caroline Jiang for advice on statistical analysis; Nipun Basrur, Priya Rajasethupathy, Andrea Terceros, Harry Choi and Molecular Instruments for advice on RNA in situ hybridization experiments; Rob A Harrell II at the Insect Transgenesis Facility at the University of Maryland for embryo injections; Cory Root, Raphael Cohn, Ari Zolin, Gaby Maimon, and Vanessa Ruta for advice on calcium imaging and data analysis; Alan Grant and Greg Suh for advice on CO2 delivery; Frances Weis-Garcia and the members of the MSKCC Antibody and Bioresource Core Facility for preparation of the nc82/Brp monoclonal antibody.

MATERIALS AND METHODS
Human and animal ethics statement: Blood-feeding procedures and behavioral experiments with live hosts were approved and monitored by The Rockefeller University Institutional Animal Care and Use Committee (IACUC protocol 17018) and Institutional Review Board (IRB protocol LV-0652), respectively.Human volunteers gave their written informed consent to participate.

Mosquito rearing and maintenance
Aedes aegypti wild-type laboratory strains (Liverpool and Orlando), CRISPR-Cas9 knock-in, and piggyBAC QUAS transgenic strains were maintained and reared at 25 -28 o C, 70-80% relative humidity with a photoperiod of 14 hr light: 10 hr dark as previously described (DeGennaro et al., 2013).Adult mosquitoes were provided constant access to 10% sucrose.For routine strain maintenance, animals were primarily blood-fed on live mice and occasionally on live human volunteers.Newly generated strains were blood-fed on human volunteers until they were established.All experiments were conducted on female mosquitoes.Detailed genotypes used in each figure, and the use of Liverpool or Orlando as the wild-type strain in a given figure can be found in Data File 1. Generation of chemosensory receptor QF2 and Split-QF2 knock-in strains T2A-QF2 gene-sparing stop codon replacement lines were generated using the strategy outlined in Matthews et al. (Matthews et al., 2019).sgRNAs were placed as close to the stop codon as possible and donor constructs were designed to remove the stop codon and replace it with an in-frame cassette containing the T2A ribosomal skipping sequence and the QF2 transcription factor or Split-QF2 domains, comprising the QF2 activation domain QF2-AD, or the QF2 DNA-binding domain QF2-DBD.This strategy spares the function of the gene at the locus being targeted, expresses QF2 or Split-QF2 domains in the cells specified by enhancers at the locus.Insertions were marked by the 3xP3 enhancer expressing a fluorescent protein.

Transcript abundance estimates of Ae. aegypti OR, IR, and GR genes
Expression values for adult sugar-fed, non-blood-fed female sensory tissues were retrieved from the Ae.aegypti L5 genome GitHub repository (https://github.com/VosshallLab/AGWG-AaegL5)at this link: https://github.com/VosshallLab/AGWG-AaegL5/raw/master/AGWG%20AaegL5%20Chemoreceptor%20TPM.xlsx.These expression values reflect libraries from a previous transcriptome study (Matthews et al., 2016) that had been aligned to the Ae.aegypti genome (AaegL5) and chemoreceptor geneset annotation reported in units of Transcripts Per Million (TPM) (Matthews et al., 2018).The number of genes from each of three gene families (ORs, IRs, and GRs) with expression values above the indicated threshold were plotted in Figure 1F,G and are available in Data File 1.

Whole brain fixation and immunostaining
Dissection of adult brains and immunostaining was done as previously described (Matthews et al., 2019).6-14 day-old mosquitoes were anesthetized on ice.Heads were carefully removed from the body by pinching at the neck with sharp forceps.Heads were placed in a 1.5 mL tube for fixation with 4% paraformaldehyde, 0.1 M Millonig's Phosphate Buffer (pH 7.4), 0.25% Triton X-100, and nutated for 3 hr.Brains were then dissected out of the head capsule in ice-cold Ca +2 -, Mg +2 -free phosphate buffered saline (PBS, Lonza 17-517Q) and transferred to a 24-well plate.All subsequent steps were done on a low-speed orbital shaker.Brains were washed in PBS containing 0.25% Triton X-100 (PBT) at room temperature 6 times for 15 min.Brains were permeabilized with PBS, 4% Triton X-100, 2% normal goat serum (Jackson Immu-noResearch #005-000-121) for ~48 hr (2 nights) at 4°C.Brains were rinsed once and then washed with PBT at room temperature 6 times for 15 min.Primary antibodies were diluted in PBS, 0.25% Triton X-100, 2% normal goat serum for ~48 hr (2 nights) at 4°C.Brains were rinsed once then washed in PBT at room temperature 6 times for 15 min.Secondary antibodies were diluted in PBS, 0.25% Triton X-100, 2% normal goat serum for ~48 hr (2 nights) at 4°C.Brains were rinsed once then washed in PBT at room temperature 6 times for 15 min.Brains were equilibrated overnight in Vectashield (Vector Laboratories H-1000) and were mounted in Vectashield.
The following primary antibodies were used: anti-Brp/nc82 (mouse; 1:50, Developmental Studies Hybridoma Bank -see below) and/or anti-GFP (rabbit: 1:10,000; Life Technologies A11122).The secondary antibodies used in all experiments except Figure S1 and Figure S6 were anti-mouse-Cy5 (1:250; Life Technologies A-10524) and anti-rabbit-Alexa Fluor 488 (1:500; Life Technologies A-11034).In Figure S1, the secondary antibody was anti-mouse-Alexa Fluor 488 (1:500; Life Technologies A-11001) and in Figure S6, the secondary antibodies were anti-mouse-Alexa Fluor 594 (1:500; Life Technologies A-11005) and anti-rabbit-Alexa Fluor 488 (1:500; Life Technologies A-11034).Purification of nc82/Brp monoclonal antibody Hybridoma cells expressing monoclonal antibody nc82 (Antibody Registry ID: AB_2314866), which recognizes the Drosophila melanogaster Brp protein (Wagh et al., 2006) developed by Erich Buchner were obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242.Frances Weis-Garcia and the members of the MSKCC Antibody and Bioresource Core Facility subsequently used these hybridoma cells to purify this monoclonal antibody.The hybridoma was adapted to Gibco™ Hybridoma-SFM (Cat # 12045084) and 1% fetal bovine serum prescreened for ultra-low levels of bovine Ig.Antibody expression was confirmed and the adapted hybridoma was inoculated into the cell compartment of the Corning™ CELLine Disposable Bioreactor (Cat # 353137) in 15 ml of Hybridoma-SFM + 0.5% fetal bovine serum (production media) at 3 million viable cells / ml.The media compartment of the flask contained 350 ml of production media.The bioreactor was incubated at 37 o C with 7% CO2 for 3 days, at which time the cells and media containing nc82 were harvested.30 million viable cells from the harvest were re-inoculated back into the cell compartment in 30 ml fresh production media.The media in the media compartment was replaced the following day with 650 ml production media.
Three days later, the media in the media compartment was replaced with 1,000 ml production media, with the next harvest 3 days later (7 days after the previous harvest).Cells were harvested weekly and fed bi-weekly until the desired amount of monoclonal antibody was reached.After the first harvest, each one contained about 3 mg of monoclonal antibody nc82/ml production media.The harvests to be purified were pooled, centrifuged at 12,855 x g for 15 min.6.5 mg / run were loaded onto a Cytiva (formerly GE Life Sciences) 1 ml HiTrap Protein G HP antibody purification column (Cat # 29048581) at 1 ml / min.The column was then washed with 0.02 M Sodium Phosphate (pH 7.0) before the monoclonal antibody was eluted with 0.1 M Glycine-HCl (pH 2.7).One ml fractions were collected and immediately neutralized with 60 ml of 1.0 M Tris-HCl (pH 9.0).The harvest, flow through and fractions from the peak were run on an a 10% SDS-PAGE (Bio-Rad Cat # 345-0010) to confirm purity and determine which should be pooled.The pooled fractions of monoclonal antibody were dialyzed into PBS overnight using dialysis tubing (Spectrum™ 132544) with a 50 kDa MWCO.Another 10% SDS-PAGE was run and the concentration determined using the absorbance at 280 using an extinction coefficient of 1.43.

Antennal lobe confocal imaging
All brains were imaged using a Zeiss Inverted LSM 880 laser scanning confocal microscope with a 25x / 0.8 NA immersion-corrected objective unless otherwise noted.Glycerol was used as the immersion medium to most closely match the refractive index of the mounting medium Vectashield.Antennal lobes in Figure 1, 2, 5, S2-S8 were imaged at either 1024 x 1024 or 2048 x 2048 pixel resolution in X and Y with 0.5 µm Z-steps for a final voxel size of either 0.0615 x 0.0615 x 0.5 µm 3 or 0.1230 x 0.1230 x 0.5 µm 3 .Both conditions oversampled relative to the objective resolution and no differences were noted between imaging conditions.The laser intensity and gain were adjusted along the Z-axis to account for a loss of intensity due to depth and care was taken to avoid saturation and ensure that the deepest glomeruli were visible for segmentation.3xP3 was used as a promoter to express fluorescent proteins as markers for the knock-ins and QUAS transgenes used in this study, and care was taken to distinguish expression derived from the 3XP3 promoter from the expression of the QF2 driver and QUAS effector lines under investigation.3xP3 drives expression in the optic lobes, as well as some cells in the dorsal brain.Neither area overlaps with the antennal lobes, and as reported previously (Matthews et al., 2019), we saw no 3xP3-driven expression in the antennal lobes in the reporter lines alone (data not shown).Representative antennal lobe images presented in the figures were cropped to remove 3xP3-driven expression elsewhere in the brain.

Antennal lobe glomerulus quantification
Confocal images of the antennal lobes in Figure 1, 2, 5, S2-S8 were processed in ImageJ/FIJI (NIH).The number of glomeruli was quantified as follows: a single region of interest (ROI) was manually drawn around each glomerulus at a section approximately central along the Z-axis.Every glomerulus was outlined and an ROI set was collected that contained the outlines of all glomeruli.Glomeruli were then separated into two groups, GFP-positive and GFP-negative glomeruli.A count of each was made to determine the number of glomeruli labeled by each line as well as the total number of glomeruli.The ROIs were flattened along the Z-axis to enable representation of the data in two dimensions in Figure 1, 2, S2-S5, S7.The left antennal lobe in 3 brains was analyzed for each genotype in Figure 1 except for Gr3, for which the left antennal lobe was analyzed in 1 brain, and both left and right antennal lobes were analyzed in an additional 4 brains in Figure S6.Although we were able to recognize general regions of the antennal lobe, the interindividual variability made it impossible to identify most glomeruli by anatomy alone.We therefore have not attempted to name and number every glomerulus in Ae. aegypti as has been done in previous studies (Ignell et al., 2005;Shankar and McMeniman, 2020).As noted by Ito et al. (Ito et al., 2014), there is considerable confusion about the use of coordinate axes in the brains of animals in general and insects in particular.The glomeruli in the antennal lobe of Ae. aegypti were originally named by Ignell et al. (Ignell et al., 2005) using a set of coordinate axes that differ from those consistently used in Drosophila melanogaster (Couto et al., 2005;Fishilevich and Vosshall, 2005;Grabe et al., 2015;Laissue et al., 1999;Stocker et al., 1990).A recent study of the antennal lobe of Ae. aegypti renamed glomeruli to account for this discrepancy in coordinate axes (Shankar and McMeniman, 2020), and throughout this paper we use the same coordinate axes they have implemented.While Shankar and McMeniman renamed most antennal lobe regions and glomeruli, they chose not to rename the MD (Medio-Dorsal) cluster of glomeruli comprising MD1, MD2, and MD3 whose sensory input derives from the maxillary palp.We have observed in our study that the MD glomeruli are medial, but they are not notably dorsal, and therefore refer to them as Glomerulus 1, Glomerulus 2, and Glomerulus 3 in this paper for simplicity.While there is utility in naming glomeruli, we suspect that the Ae.aegypti mosquito antennal lobe atlas will be refined in the future with the advent of new genetic tools that will unambiguously allow the field to distinguish and name genetically identifiable glomeruli.We found that the size, shape, and number of antennal lobe glomeruli in Ae. aegypti was variable from animal to animal.It is possible that the boundaries between glomeruli are not easily distinguished by synaptic staining and that specific glomeruli will become identifiable once there are genetic tools available that label smaller populations of olfactory sensory neurons.The anatomical variability we see is consistent with both the original map that identified 50 glomeruli (Ignell et al., 2005), which divided glomeruli into 3 classes based on their variability in location, as well as a recent study that looked specifically at the size and shape of glomeruli across animals (Shankar and McMeniman, 2020) and revised the original map to a count of ~80 glomeruli.Shankar and McMeniman named and numbered these glomeruli across animals, but they noted that they were only able to consistently identify 63 glomeruli.This is similar to the ~65 glomeruli we observed in our work.While there is not yet a clear consensus on the exact number of antennal lobe glomeruli in Ae. aegypti, the number of chemosensory receptors expressed in the antenna and maxillary palp is at least twice as large as any of the estimates of glomerulus number.The variability in antennal lobe structure appears at first to contrast with Drosophila melanogaster, where each glomerulus can be clearly identified and named.However, we note that the antennal lobe map in Drosophila melanogaster has been refined with the advent of new genetic techniques, starting with 35 glomeruli in the original atlas (Stocker et al., 1990), then modified to 40 glomeruli (Laissue et al., 1999), and further refined in numerous studies (Couto et al., 2005;Fishilevich and Vosshall, 2005;Tanaka et al., 2012) including a recent count of 54 (Grabe et al., 2015) and 58 (Task et al., 2020) glomeruli.We have refrained from naming glomeruli in Ae. aegypti at this time because we believe that a more stereotyped anatomy will emerge as new genetic lines are generated that allow cell-type-specific labelling.A recent study in the mosquito Anopheles gambiae using mosquitoes that label Orco-expressing olfactory neurons also noted that the antennal lobe was variable between animals relative to Drosophila melanogaster (Riabinina et al., 2016).It is therefore possible that mosquito antennal lobes are more variable than Drosophilids (Grabe et al., 2015;Prieto-Godino et al., 2017).Variability in olfactory bulb structure is seen even in the mouse, Mus musculus, where the principles of olfactory organization were first established (Schaefer et al., 2001;Strotmann et al., 2000;Zou et al., 2009).The exact size and location of glomeruli can vary between animals more than initially appreciated and appears to be determined by both genetic factors and activity in olfactory sensory neurons during the early life of the animal.In Drosophila melanogaster, glomerulus size is highly genetically determined and correlates strongly with the number of olfactory sensory neurons that innervates each glomerulus (Grabe et al., 2015).Whether the variability in glomerulus size in the mosquito is due to activity-dependent changes in structure or other factors remains to be seen.

Additional technical notes on expression and projection patterns of chemosensory receptor knock-in strains
Orco-QF2>QUAS-mCD8:GFP: We noted that the intensity of GFP varies between glomeruli in this driver line, with some bright and others comparably dim.We speculate that this is due to a combination of the variability in Orco expression levels in individual neurons and variability in the density of innervation in individual glomeruli.A large region of the anterior ventral antennal lobe was previously referred to as the Johnston's organ center and was thought to comprise a single large glomerulus (Ignell et al., 2005).In other insect species, Johnston's organ mediates detection of auditory cues.Consistent with a recent study (Shankar and McMeniman, 2020), we segmented this region into multiple glomeruli based on anatomical boundaries revealed with Brp immunofluorescence.Glomeruli in this region are innervated by Orco-expressing neurons, calling into doubt the original report that these glomeruli process auditory stimuli and suggesting instead that they serve an olfactory function.In support of this hypothesis, the analogous area of the Anopheles coluzzii antennal lobe has been shown to receive projections from Orco-expressing olfactory sensory neurons (Riabinina et al., 2016).We also observed GFP projections into the subesophageal zone in Orco-QF2>QUAS-mCD8:GFP animals, which appear to derive from expression in the proboscis, the primary taste organ in insects.This is consistent with similar expression in Anopheles coluzzii (Riabinina et al., 2016) and functional data in Anopheles gambiae showing that olfactory responses are detected in this gustatory organ (Kwon et al., 2006).Ir25a-QF2>QUAS-mCD8:GFP: The intensity of GFP projections varies between glomeruli in this driver line, with some bright and other comparably dim, as noted for Orco-QF2.The brightest glomeruli are primarily medial and anterior.We see the dimmest innervation in the area previously described as Johnston's organ center as well as in the central antennal lobe.Labeling was also seen in other areas of the brain, most notably the subesophageal zone and anterior mechanosensory motor center.Ir8a-QF2>QUAS-mCD8:GFP: Depending on the brain being analyzed there were either 2 or 3 medial glomeruli labelled in this line.In the cases where there were 3 medial glomeruli, this third medial glomerulus was innervated by a few large-diameter axons.These were larger and sparser than the smaller axons that densely innervated most other glomeruli in this line.We also note that there are 2-3 cell bodies that express GFP located in the cell body rind lateral to the antennal lobe (rALl).We are unable to definitively describe where these cells project without genetic reagents that selectively label these cells, but they appear to send bilateral processes that cross the midline within what appears to be the saddle to innervate the anterior mechanosensory motor center outside the antennal lobe.All naming is in accordance with the new insect brain nomenclature presented in Ito et al. (Ito et al., 2014).Ir76b-QF2>QUAS-mCD8:GFP: In addition to projections to the antennal lobe, this line shows innervation of the subesophageal zone of the brain.Gr3-QF2>QUAS-mCD8:GFP: All antennal lobes in this line show innervation of a single glomerulus (also referred to as "MD1" and here referred to as "Glomerulus 1"; (Ignell et al., 2005;Shankar and McMeniman, 2020).In several brains, we saw a second small medial glomerulus that derives its innervation from the antenna and is in a small medial cluster of landmark glomeruli midway down the anterior-posterior axis closest to the center of the brain.Innervation appears to come from only a few axons.This low and variable reporter expression is consistent with the low level of expression of Gr3 in the antennal transcriptome (Matthews et al., 2016).Because this line only shows innervation of these 1-2 glomeruli, we analyzed all glomeruli only in the single brain in Figure 1I, and additionally analyzed 8 more antennal lobes in 4 brains for the presence or absence of labelling in these two glomeruli.We analyzed both left and right antennal lobes from 4 brains and found that in 3 of the 4 brains there was a second glomerulus in one or both antennal lobes (Figure S6).The presence of the second glomerulus was not specific within a single animal as we found all variations of presence and absence of this glomerulus across both antennal lobes in these 4 animals.In some Gr3-QF2>QUAS-mCD8:GFP animals, we detected a small number of processes that extended beyond the antennal lobe and into the higher brain, although the exact termination site varied.We never saw CO2evoked activity in the variable second glomerulus or these projections outside the antennal lobe.Images in Figure S6 were taken as described above with the following changes: Secondary antibodies used were anti-mouse-Alexa Fluor 594 (1:500; Life Technologies A-11005) and anti-Rabbit-Alexa Fluor 488 (1:500; Life Technologies A-11034).Images were taken using a Zeiss Inverted LSM 880 laser scanning confocal microscope with a Plan-Apochromat 40x/1.4Oil DIC objective.Images were taken at 1024 x 1024 in XY to generate images with a final voxel size of 0.1384 x 0.1384 x 0.5 µm 3 .Images were scored as containing GFP in one or two glomeruli.

Additional technical notes on expression and projection patterns of Split-QF2 strains
Ir25a-QF2-AD,Orco-QF2-DBD > QUAS-dTomato-T2A-GCaMP6s: All antennal lobe immunostaining in Figure 2, 5, S7, S8 was carried out as described above with slight modifications to utilize the 15xQUAS-dTomato-T2A-GCaMP6s effector line.The same primary antibodies were used because of the structural similarity between GCaMP6s and GFP.Intrinsic dTomato was detected without antibody amplification, as it retained fluorescence throughout fixation and staining.Brp (Cy5), dTomato, and GCaMP6s (Alexa Fluor 488) were imaged as three separate confocal channels as described above.Glomeruli labelled by dTomato completely overlapped with those labelled by GCaMP6s immunofluorescence, so both channels were used during the quantification of positive and negative glomeruli.dTomato labeling was used to generate sample images.There was no staining in the antennal lobes of the individual split effector lines crossed to 15xQUAS-dTomato-T2A-GCaMP6s (n=3 / genotype, Figure 2, S7).Due to an initial observation that certain glomeruli were consistently unlabeled in the Ir25a-AD;Orco-DBD line but were labelled in the Orco-QF2 and IR25a-QF2 lines, particularly in an anterior region of the brain, we conducted a second experiment to compare glomeruli labelled by the Orco-QF2, IR25a-QF2 and IR25a-AD;Orco-DBD drivers side-by-side.All drivers were crossed to the 15xQUAS-dTomato-T2A-GCaMP6s effector (Figure 2J). 3 brains per genotype were immunostained and imaged as described above.

Antennal lobe anterograde dye fill
For images in Figure S1, mosquitoes were anesthetized on ice until immobile and then transferred to a cold dissection dish.A single antenna or maxillary palp was loaded with Texas-red conjugated dextran (Molecular Probes D3328) diluted 10 mg in 100 μL external saline (103 mM NaCl, 3 mM KCl, 5 mM 2-[Tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid (TES), 1.5 mM CaCl2, 4 mM MgCl2, 26 mM NaHCO3, 1 mM NaH2PO4, 10 mM trehalose, 10 mM glucose, pH 7.3, osmolality adjusted to 275 mOsm/kg).To load the dye a small drop (approximately 0.5-1 µL) of dye was placed onto the surface of the dish and the animal was moved such that the intended cut-site on a single antenna or maxillary palp was placed in the drop of dye.The antenna or maxillary palp was then removed with sharp forceps and a fine scalpel (F.S.T 10315-12) while it was submerged in the dye.Care was taken to remove the maxillary palp proximal to the 4 th segment, to include all the basiconic sensilla, and to remove the antenna near the base but to leave the antennal pedicel completely intact.The animal remained immobile on ice with the antenna or maxillary palp submerged and the dye was loaded for 2-5 min.After this time the animal was placed in a small soup cup with access to 10% sucrose and returned to standard rearing conditions overnight to give the dye time to diffuse throughout the neurons and fill the length of the axon.The next morning dissection of adult brains and immunostaining was carried out as described above.

Antennal lobe 3-D reconstructions
In an attempt to develop a map of the Ae.aegypti antennal lobe, 3 brains from the Liverpool strain were immunolabeled with Brp to identify the boundaries between antennal lobe glomeruli.The left antennal lobe in each brain was independently reconstructed from confocal sections taken with a Plan-Apochromat 63x/1.40NAoil immersion objective, at 1024 x 1024 pixel resolution in X and Y with 0.5 µm Z-steps for a final voxel size of either 0.1318 x 0.1318 x 0.5 µm 3 using the software Imaris (Bitplane).Although the area previously termed Johnston's organ center was considered a single glomerulus in a previous study (Ignell et al., 2005), we noted anatomical boundaries in this region, suggesting that it contains multiple glomeruli.This observation is consistent with recently published work (Shankar and McMeniman, 2020) and this area was segmented by an individual researcher to generate the final reconstructions.Two of these are shown in Figure S1.Each glomerulus was manually segmented into an individual surface using Surpass View.We were consistently able to identify the three glomeruli innervated by the maxillary palp, previously termed MD1, MD2 and MD3 (Ignell et al., 2005) which we refer to in this study as Glomerulus 1, Glomerulus 2, and Glomerulus 3 (Figure 1, Figure 5).The overall structure of the antennal lobe varied considerably from animal to animal and although we were able to identify certain regions and certain landmark glomeruli including those that are targeted by the maxillary palp, we were unable to assign an unambiguous identity to every glomerulus, as is possible in Drosophila melanogaster (Couto et al., 2005;Fishilevich and Vosshall, 2005).This variability makes it essentially impossible to identify a given glomerulus between animals and we therefore have decided to avoid referring to glomeruli by previous naming schemes, including MD1, MD2, MD3.An authoritative atlas of the Ae.aegypti antennal lobe awaits genetic reagents that label subpopulations of sensory neurons that will permit the field to refer to glomeruli by their molecular identity.

Antennal whole mount immunofluorescence
Whole-mount immunostaining of adult antennae was performed as described (Riabinina et al., 2016) with modifications.7-11 day-old Liverpool mosquitoes were immobilized on ice, decapitated and heads and placed in 1 mL ZnFA fixative solution (0.25% ZnCl2, 2% paraformaldehyde, 135 mM NaCl, 1.2% sucrose and 0.03% Triton X-100) for 20-24 h at room temperature in the dark.Next, the heads were washed three times for 30 min each with HBS buffer (150 mM NaCl, 5 mM KCl, 25 mM sucrose, 10 mM HEPES, 5 mM CaCl2 and 0.03% Triton X-100).Antennae were carefully removed in HBS on ice and placed in 400 μL HBS in 0.5 mL Eppendorf tubes.After a brief wash in HBS, the tissue was incubated in 400 μL 80% methanol/20% dimethyl sulfoxide (DMSO) solution for 1 h at room temperature, washed for 5 min in 400 μL 0.1 M Tris pH 7.4, 0.03% Triton X-100 solution and incubated in 400 μL blocking solution (PBS, 5% normal goat serum (Jackson 005-000-121), 1% DMSO and 0.3% Triton X-100) for at least 3 h at room temperature or overnight at 4 o C. Next, the tissue was placed in a 0.5 mL Eppendorf tubes containing 400 μL blocking solution with primary antibodies [rabbit anti-Orco EC2 (Larsson et al., 2004), 1:50, Vosshall lab; chicken anti-GFP,1:200, Aves GFP-1020] and submerged and held in a water bath sonicator (Branson m1800) for 30 sec at the high setting.Next, the tubes were placed on a rotator for 2 days at 4 o C in the dark, after which the sonication procedure was repeated.The tubes were placed on a rotator for 2 additional days (for a total of 4 days) at 4 o C in the dark.Next, the tissue was washed 5X 30 min each at room temperature in PBS, 1% DMSO and 0.3% Triton X-100.Secondary antibodies (anti-rabbit Alexa Fluor 555 Plus, 1:200, Thermo Fisher A32732, anti-chicken Alexa Fluor 488, 1:200, Thermo Fisher A-11039) and nuclear dye (TO PRO 3 Iodide, 1:400, Thermo Fisher T3605) were added to the blocking solution, and tubes were sonicated as described above and incubated for 4 days at 4 o C in the dark with the sonication repeated after 2 days of incubation.The tissue was then washed 5X 30 min at room temperature in PBS, 1% DMSO and 0.3% Triton X-100, rinsed in PBS and mounted in Slow Fade Diamond for confocal imaging.Whole mount antennal and maxillary palp RNA in situ hybridization RNA was detected in whole mount antenna and maxillary palp using the hybridization chain reaction (HCR) technique as previously described (Choi et al., 2018) with modifications.Probes, amplifiers, Probe Hybridization Buffer, Amplification Buffer, and Probe Wash Buffer were purchased from Molecular Instruments.Full list of probe lot numbers can be found in Date File 1. 5-8 day-old Liverpool mosquitoes were cold anesthetized, manually decapitated with forceps, and heads with antennae and the proboscis were digested in a chitinase-chymotrypsin solution (119 mM NaCl, 48 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 25 mM HEPES, 5 U/mL chitinase (Sigma-Aldrich C6137-50UN), 100 U/mL alpha-chymotrypsin (Sigma-Aldrich CHY5S-10VL), 2% DMSO) (Manning and Doe, 2017) at 37 o C for 30 min (antennae) or 1 hr (maxillary palps) in a Fisher Isotemp oven and subsequently fixed in 4% paraformaldehyde, 1X PBS, 0.03% Triton X-100 on a rotator at 4 o C overnight.Heads were washed 4 times on ice for 10 min each in 0.1% PBS-Tween-20.Antennae or maxillary palps were dissected in 0.1% PBS-Tween-20 on ice and dehydrated with a graded series of methanol/0.1% PBS-Tween: 25% methanol in 0.1% PBS-Tween-20 for 10 min on ice, 50% methanol in 0.1% PBS-Tween-20 for 10 min on ice, 75% methanol in 0.1% PBS-Tween-20 for 10 min on ice, and two washes of 100% methanol for 10 min on ice.Tissues were incubated overnight in 100% methanol at -20 o C and were subsequently rehydrated with a series of graded methanol/0.1% PBS-Tween-20: 75% methanol in 0.1% PBS-Tween-20 for 10 min on ice, 50% methanol in 0.1% PBS-Tween-20 for 10 min on ice, 25% methanol in 0.1% PBS-Tween-20 for 10 min on ice, and two washes of 0.1% PBS-Tween-20 for 10 min each on ice.Tissue was digested in 20 µg/mL Proteinase-K (Thermo Fisher AM2548) in 0.1% PBS-Tween for 30 min at room temperature and washed twice with 0.1% PBS-Tween-20 for 10 min each at room temperature.Tissue was fixed in 4% paraformaldehyde in 0.1% PBS-Tween-20 for 20 min at room temperature and washed 3 times for 10 min each in 0.1% PBS-Tween-20 at room temperature.Tissue was incubated in Probe Hybridization Buffer at room temperature for 5 min and then in 37 o C pre-warmed Probe Hybridization Buffer rotating in a hybridization oven for 30 min.8 pmol of each probe set was prepared in 37 o C pre-warmed Probe Hybridization Buffer and tissue was incubated in probe solution at 37 o C in a hybridization oven for 2 nights.Tissues were washed in 37 o C prewarmed Probe Wash Buffer 5 times for 10 min each at 37 o C. Tissues were washed twice in 5X SSC 0.1% Tween-20 at room temperature for 10 min each.Tissues were pre-amplified in room temperature Amplification Buffer for 10 min.18 pmol hairpins were separately prepared by heating 6 µl of 3 µM stock of hairpins H1 and H2 at 95 o C for 90 sec on an Eppendorf Mastercycler and allowing to cool to room temperature in a dark drawer for 30 min.Hairpins were resuspended in 100 µl amplification buffer and tissues were incubated in this hairpin solution in the dark on a rotator at room temperature overnight.Tissues were washed 5 times for 10 min each in 5X SSC 0.1% Tween-20 and mounted in SlowFade Diamond (Thermo Fisher S36972) on glass slides with coverslips for confocal imaging.Whole mount antennal and maxillary palp dTomato visualization 7-14 day-old Ir25a-QF2, Orco-QF2, Ir25a-QF2AD, Orco-QFDBD, and Ir25a-QF2AD,Orco-QFDBD > 15XQUAS-dTomato-T2A-GCaMP6s mosquitoes were cold anesthetized, manually decapitated with forceps and heads with antennae and maxillary palps were immediately fixed in 1 mL 4% paraformaldehyde, 1X PBS, 0.03% Triton X-100, on a rotator in the dark at 4 o C overnight.Heads were washed 3X 30 min each in 1X PBS, 0.03% Triton X-100 at room temperature, then antennae and maxillary palps were carefully removed and placed in 1X PBS, 0.03% Triton X-100.Next, antennae and maxillary palps were placed in a solution of 1X PBS, 0.03% Triton X-100, 1% DMSO, and a 1:400 dilution of TO PRO 3 (Thermo Fisher T3605) for 24 hr at 4 o C in the dark.Antennae and maxillary palps were then washed 5X 30 min each in 1X PBS, 0.03% Triton X-100 at room temperature in the dark, washed once with 1X PBS, transferred to a well of SlowFade diamond to remove excess PBS, and mounted in SlowFade Diamond for confocal imaging.

Antennal and maxillary palp confocal imaging and cell quantification
Images of peripheral tissues were acquired with a Zeiss Axio Observer Z1 Inverted LSM 880 NLO laser scanning confocal microscope (Zeiss) with a 25x/0.8NA or 63x/1.4NA immersion-corrected objective at a resolution of 3096 x 3096 pixels or 2048 x 2048 pixels.When comparing dTomato fluorescence across genotypes, image acquisition parameters were kept consistent.When necessary, tiled images were stitched with 20% overlap.Confocal images were processed in ImageJ (NIH).Because the antenna is a cylindrical structure, when whole antennal segments are mounted on a slide and imaged on a confocal microscope, signal can be easily detected from the region closest to the coverslip and confocal objective, but signal is weaker when imaging the side further from the coverslip and objective.For the purposes of consistent quantification, we only quantified cell numbers from the region closest to the coverslip (orange region in Figure S9A).For quantifying expression in the maxillary palp, only the dorso-lateral region of the 4 th maxillary palp segment was analyzed.(yellow region in Figure S9B).Quantification of co-expression in antennae and maxillary palps was done in ImageJ (NIH) using the Cell Counter plugin.Cells in each channel were manually marked independently of the signal in the other channels.After cells in each channel are marked, and markers were then merged.Cells that were labeled with multiple markers (co-expressing cells) were then marked with a third marker (Figure S9C-H).Cell counts were then imported into Microsoft Excel and R for analysis.

In vivo 2-photon GCaMP calcium imaging preparation
Calcium imaging was performed on an Ultima IV two-photon laserscanning microscope (Bruker Nanosystems) equipped with galvanometers and illuminated by a Chameleon Ultra II Ti:Sapphire laser (Coherent).GaAsP photomultiplier tubes (Hamamatsu) were used to collect emitted fluorescence.Images were acquired with a 60X/1.0N.A.Long Working Distance Water-Immersion Objective (Olympus) at a resolution of 256 x 256 pixels.GCaMP6s was expressed via the Ir25a-QF2 knock-in which labels all three maxillary palp glomeruli.Preliminary imaging studies were conducted in animals that were wild-type at the Gr3 locus.In Figure 5 all Gr3+/and Gr3-/-animals were reared and imaged in parallel.Because the Gr3 mutants were generated in the Orlando wild-type background and all Q-system lines were generated in the Liverpool wildtype background, care was taken to compensate for background effects in the following way.A single stock was created containing the 15x-QUAS-dTomato-T2A-GCamP6s-3xP3-ECFP transgene (Matthews et al., 2019), the Ir25a-QF2-3xP3-dsRed gene-sparing knock-in, and the Gr3 ECFP knock-out allele that is marked with ECFP (McMeniman et al., 2014).This parental line was then crossed to either the unmarked Gr3 Δ4 mutant that is in the Orlando background (McMeniman et al., 2014) to generate Gr3-/-animals, or to wild-type Orlando to generate Gr3+/-animals.The Gr3 mutation is recessive (McMeniman et al., 2014), so this comparison of heteroallelic and heterozygous animals both controls for genetic background and makes it possible to assay the Gr3 mutant phenotype.Calcium imaging experiments were performed on 7-14 day-old female mosquitoes.Mosquitoes were fed on 10% sucrose and switched to water overnight prior to imaging.Mosquitoes were anesthetized at 4 o C for dissection.The mosquito was fixed to a custom Delrin plastic holder with UV-curable glue (Bondic).The mosquito was inserted into a hole in the holder, such that the head and thorax were exposed above the surface of the holder, with the rest of the mosquito below.The mosquito was secured with a few points of glue (Bondic) on the thorax and head.The antennae and maxillary palp remained below the plate and were kept free of saline to prevent damage to the tissue.The top of the dish was then filled with external saline, which is based on Drosophila melanogaster imaging saline (Fişek and Wilson, 2014) with the following composition: 103 mM NaCl, 3 mM KCl, 5 mM 2-[Tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid (TES), 1.5 mM CaCl2, 4 mM MgCl2, 26 mM NaHCO3, 1 mM NaH2PO4, 10 mM trehalose, 10 mM glucose, pH 7.3, osmolality adjusted to 275 mOsm/kg).A small window was gently opened in the head and the cuticle was removed anterior to the antennal pedicles and the tissue was secured with dental wax to expose the antennal lobes.Opaque nonneural tissue, primarily fat cells, was removed if they obstructed the antennal lobes.Great care was taken not to damage the antennal nerves.The preparation was placed on a custom laser cut acrylic holder that was secured to the stage and placed the mosquito in close proximity to the output port of the olfactometer.Plans for the fabrication of the imaging holder are available on Github: https://github.com/VosshallLab/Younger_Herre_Vosshall2020.On rare occasions mosquitoes showed no movement or odor responses and were discarded.Image volumes were taken at higher resolution to identify glomeruli innervated by the maxillary palp, including the nerves that innervate these glomeruli.Glomerulus 1 was always easily identifiable by eye.A single plane through the center of each glomerulus was scanned at 4.22 frames per second with a 920 nm excitation wavelength imaged through a 680 nm shortpass infrared (IR) blocking filter, a 565 nm longpass dichroic and 595/50 nm or 525/70 nm bandpass filters.GCaMP6s and dTomato emission was collected simultaneously for at least 70 frames per trial.

In vivo 2-photon GCaMP calcium imaging stimulus delivery
Odor stimulation was achieved by directing a continuous stream (400-800 mL/min) of clean air, originating from a breathing air tank (Praxair), through 0.125 inch inner diameter/0.250inch outer diameter perfluoroalkoxy alkaline (PFA) tubing.The antenna and maxillary palp of the mosquitoes were positioned close to the tube such that they were exposed to constant air flow (carrier stream).4-60% of the total air stream was diverted through the headspace of a 20 mL borosilicate glass vial (EnviroWare #03-339-14E) containing the odorant diluted in either paraffin oil or water (odor stream).At a trigger, a custom-built solenoid valve controller system redirected the odor stream from a blank vial to a vial containing various odorants diluted in paraffin oil or water to a final volume of 3 mL.Unless otherwise noted, 40% of the air stream was used for odor delivery.Prior to conducting experiments, a mini-PID (Aurora Scientific, model 201A) was used to measure odorant waveforms and ensure the consistency of odor presentations across trials.CO2 was introduced into this system from a tank of 5% CO2 in a 20.9% oxygen/ 79.1% nitrogen balance (Praxair).At a trigger, CO2 was diverted through the headspace of a second clean 20 mL borosilicate glass vial (CO2 stream) that was redirected from a bypass stream into the system.This was coupled to a clean air stream with a matched flow rate that was simultaneously removed from the system.In experiments where odor concentration of CO2 concentration was varied, the fraction of odor stream directed to the mosquito was adjusted.CO2 was delivered at 4-60% of the total air stream for total concentrations of 0.2-3.0%CO2.In Figure 5, a final concentration of 2% CO2 was used in all experiments.Prior to conducting experiments, a CO2 meter (Vaisala GM70) was used to measure CO2 concentration in the final air stream and we found that the expected concentration was delivered.Odor and CO2 delivery was controlled using custom software written in Matlab R2020a (MathWorks).The odor delivery system and software was based on a previously used system and updated for CO2 delivery (Cohn et al., 2015).After the mosquito was transferred to the microscope and placed within the air stream, there was a 10-min acclimation period before any imaging began.For each trial, 5 sec of GCaMP6s fluorescence at baseline was imaged and then a one second odor or CO2 pulse was delivered while imaging continued for at least 10 more sec (at least 70 frames total).There was a one-min delay between imaging trials so that the mosquito could reacclimate to the air stream after exposure to each odor.In an initial experiment we did not see any change in the amplitude of the CO2 response across multiple trials within an animal.Imaging remained stable during the duration of the imaging session in all animals that were included in this study.We did not notice a decrease in the response to stimuli over time.Odorants were selected for the highest purity available for purchase (greater than 98%) and were either diluted in MilliQ ultrapure water (resistance 18 megaohm) or paraffin oil (EMD Millipore #PX0045-3) depending on solubility.All odorants were diluted into a large stock solution that was used throughout each entire experiment to avoid variability in concentrations.The odorants used in Figure 5 were: R-(-)-1-octen-3-ol (PubChem CID: 6992244, Penta Manufacturing 15-18900) diluted 1:10 in paraffin oil; 3-octanone (PubChem CID: 246782, Sigma 136913) diluted 1:10 in paraffin oil; acetone (PubChem CID: 180 Sigma A4206) diluted to 5%v/v in MilliQ ultrapure water; ammonia (Pub-Chem CID: 222, Suprapur, Supelco 105428) diluted to 2.5% v/v in MilliQ ultrapure water; hexyl amine (PubChem CID: 8102, Sigma 219703) diluted to 1% v/v in MilliQ ultrapure water; triethyl amine (PubChem CID: 1146, Sigma T0886) diluted to 1% v/v in MilliQ ultrapure water.In vivo 2-photon GCaMP calcium imaging data analysis All image processing was done using FIJI/ImageJ (NIH).Data analysis used Matlab R2020a (MathWorks), Excel (Microsoft), and Prism (GraphPad).Regions of interest were selected based on the dTomato fluorescence intensity and GCaMP6s was normalized to dTomato intensity for analysis.A Gaussian blur with a sigma value of 1 was performed on the GCaMP6s signal.In the calculation of ΔF/F, six frames were averaged before stimulus presentation to determine the baseline fluorescence.To determine maximum ΔF/F, the average of the maximum 3 frames at the peak after stimulus delivery was determined for each sweep.

Quattroport olfactometer behavioral assays
Mosquitoes that were blood fed for behavior were fed on warmed defibrinated sheep's blood (Hemostat Laboratories DSB100) through a parafilm membrane.Animals were sorted by eye to determine if they were blood fed by looking at engorgement of the abdomen, and non-fed animals were not used in subsequent behavior assays.Blood-fed animals were returned to the same rearing conditions as their non-blood-fed counterparts, and behavior was assayed ~48 hr after they had consumed a blood meal.Quattroport olfactometer experiments were conducted as previously described (Basrur et al., 2020) with modifications to enable the delivery of monomolecular odorants.Details of fabrication and operation of the Quattroport are available at https://github.com/Voss-hallLab/Basrur_Vosshall2020.All behavioral experiments were carried out in an environmental room set to 25 -28 o C, 70-80% relative humidity, and were carried out between ZT3-14.Trials with identical stimuli were interspersed throughout experimental days to control for circadian rhythm fluctuations.Mosquitoes in the Quattroport began each trial in a start canister, flew through a flight tube, and, if they were attracted to a stimulus, into a trap proximal to the stimulus.Filtered air, with or without additional CO2, was pumped through the stimulus box and into the flying tube.Mosquitoes were prevented from contacting the stimulus by a mesh barrier, and adjustable gates between each chamber allow the experimenter to count the number of mosquitoes in each compartment.The night before each experiment, approximately 20 7-22 days old mosquitoes were cold-anesthetized and placed in each canister.They were sugar-starved with access to water overnight.The next day, at the start of each trial a set of 4 canisters was attached to the Quattroport and mosquitoes were given 10 min to acclimate with filtered air flowing at the 25 mm setting on the flowmeter throughout the assay (see https://github.com/VosshallLab/Bas-rur_Vosshall2020 for details on filtered air and CO2 flowmeter settings).They were then exposed to the stimulus with or without added CO2 flowing at the 50 mm flowmeter setting for 30 sec, after which they were given access to the flying tube and attraction trap for 5 min in the continued presence of filtered air with or without added CO2 and the stimulus.At the end of the trial, the gates enclosing the start canister and the attraction trap were closed, and mosquitoes that remained in the start canister, entered the flying tube, or flew into the attraction trap were counted.For live human host-seeking assays, a single human volunteer placed their forearm in the stimulus box.For the delivery of lactic acid and triethyl amine, the odorant solutions were delivered using 35 mm Petri dishes (Fisher Scientific #08-757-100A) placed within the air stream in the stimulus box.The odorants used in this assay were L-(+)-Lactic acid solution, 88-92% (PubChem CID: 107689, Sigma-Aldrich #27714-500mL) and Triethyl amine (PubChem CID: 1146, Sigma T0886) diluted to 1% v/v in MilliQ ultrapure water.In cases where both odorants were used as a stimulus, care was taken not to mix the liquids when they were placed in the stimulus box.Different volumes of each liquid odorant were delivered to generate the dose response curves in Figure 6F,H and Data File 1. Depending on the experiment, between 1 μL and 2 mL of liquid odorant was used.When less than 1 mL of liquid was used, the lid of the Petri dish was inverted, and the liquid odorant was pipetted onto the center of the inverted lid.The inverted lid was then placed on the bottom of the Petri dish, which elevated it slightly, placing the odorants in the path of the air stream that ran through the stimulus box.When 1 mL or greater of liquid was used, the liquid odorant was pipetted into the base of the Petri dish, which was set atop the inverted lid of the Petri dish to elevate it.If the volume of odor was 1 mL or greater for either odorant, they were delivered using 2 separate Petri dishes to avoid possible mixing.If the volume of each liquid was less than 1 mL, both liquid odorants were delivered as separated drops of liquid on a single inverted Petri dish lid, and the drops remained sperate because of surface tension.When neither lactic acid nor triethyl amine was delivered, an empty Petri dish was placed inside the stimulus box to serve as a control for any odors that might be emitted by the dish itself and/or any perturbations in air flow introduced by the presence of the Petri dish lid and base

Figure 1 :
Figure 1: Mismatch in chemosensory receptor and olfactory glomerulus number suggests a novel olfactory organization (A) Ae. aegypti female with sensory structures highlighted in yellow boxes.(B) Approximate number of antennal lobe glomeruli per brain hemisphere innervated by the indicated sensory structure, derived from quantification of the left antennal lobe in 12 brains presented in (I-J) and Figure S2-S5.See also Figure S1.(C) Cartoons of insect chemosensory gene families.(D) Cartoon of canonical olfactory system organization.(E-G) Stacked bar plots of the number of chemosensory genes in the Ae.aegypti genome (E), and the number expressed above the indicated TPM thresholds in the antenna (F) and maxillary palp (G).(H) Two models of olfactory system organization that can account for the observation that there are far more chemosensory receptors than olfactory glomeruli in Ae. aegypti.(I) Maximum-intensity projections of confocal Z-stacks of antennal lobes in the left-brain hemisphere of the indicated genotype with immunofluorescent labeling of GFP (green) and the nc82 monoclonal antibody, which recognizes Brp (magenta).Brp is used throughout this paper as a synaptic marker (Wagh et al., 2006).Scale Bar: 50 µm.Orientation: d=dorsal, m=medial.(J) 2-D representation of the boundary of each glomerulus that is GFP positive and GFP negative.See also Figure S2-S6.(K) Cartoon schematic of the glomeruli receiving projections from olfactory sensory neurons expressing the indicated chemosensory receptor.

Figure 2 .
Figure 2. Genetic evidence for widespread Orco and Ir25a co-expression (A) Schematic of the Split-QF2 system (left) and diagrams of Orco and Ir25a gene loci with exons (grey boxes), introns (grey lines) and CRISPR-Cas9 gRNA site (arrowhead) used to insert T2A-QF2-AD (light blue) and T2A-QF2-DBD (medium blue).AD and DBD gene maps are not to scale.(B-C) Schematic of the Split-QF2 system (B) and outcome of gene expression in olfactory sensory neurons of the indicated genotypes (C).(D-E) Maximum-intensity projections of confocal Z-stacks of antennae (D) and maxillary palp (E) of the indicated genotypes showing intrinsic dTomato fluorescence, with transmitted light overlay.See also Figure S7A,B.(F-G) Maximum-intensity projections of confocal Z-stacks of antennal lobes from the left brain hemisphere of the indicated genotype with immunofluorescent labeling of dTomato (green) and Brp (synaptic marker, magenta).See also Figure S8.(H-I) 2-D representation of the boundary of each glomerulus in (G) that is GFP positive and GFP negative (H) and quantification (I).n=3, mean ± SEM.See also Figure S7C.(J) Single confocal sections of the anterior dorsal region of the antennal lobe in the left-brain hemisphere of the indicated genotype with immunofluorescent labeling of dTomato (green) and Brp (synaptic marker, magenta).(K) Schematic of the 5 glomeruli shown in (J) with GFP positive glomeruli (green).(L) Two models of Ae. aegypti olfactory system organization supported by data in this figure.Scale bars: 50 µm, except (J): 25 µm.Orientation: proximal left (D), distal left (E); (F,G,H,J): d=dorsal, m=medial.

Figure 3 .
Figure 3. Extensive co-expression of chemosensory co-receptors in the antenna (A) Orco and GFP immunostaining in antennae of the indicated genotypes with cartoon schematic indicating cell identity.(B-D) Quantification of antennal cells in the indicated genotypes co-expressing Orco protein and GFP presented as violin plots showing raw cell counts (B)(mean with range), stacked bar plots (C)(mean ± SEM), and Euler diagrams with area scaled to mean cells/region (D).n=6-8 antennal segments, 34-68 cells/region.(E) RNA in situ hybridization in wild-type antennae with the indicated probes.(F-H) Quantification of wildtype antennal cells expressing the indicated genes as violin plots showing raw cell counts (F)(mean with range), stacked bar plots (G)(mean ± SEM), and Euler diagrams with area scaled to mean cells/region (H).n=4 antennal segments, 45-63 cells/ region.(I) Cartoon schematic of olfactory neuron populations identified in this figure.Scale bars: 10 µm.

Figure 5 :
Figure 5: Volatile amine sensing by CO2 neurons that lack the Gr3 CO2 receptor (A-B) Cartoon (A) and left 3-D antennal lobe reconstruction (B) showing 3 glomeruli that are innervated by the maxillary palp.Image in (B) is reprinted in Figure S1G.(C-H) Single confocal sections through the center of Glomerulus 1 (top) or Glomerulus 2 and Glomerulus 3 (bottom) in left antennal lobes of the indicated genotypes.Sections are taken from Z-stacks presented in Figure 1I (C-G) and Figure 2G (H).(I) Schematic of sensory neuron gene expression and glomerular convergence based on (C-H).(J) Side view schematic of liveimaging preparation of Glomerulus 3. (K) GCaMP6s fluorescence traces from Glomerulus 3 presented with the indicated stimuli.Black bar indicates stimulus presentation, mean ± SEM, n=9-10 sweeps from 3 animals/stimulus.(L) Summary of data in (K) indicating the maximum ΔF/F (arbitrary units).Lines denote median and quartiles, *p<0.05one-way ANOVA with Bonferroni correction compared to water control.(M) Side view schematic of live-imaging preparation of Glomerulus 1. (N) Two-photon image of Glomerulus 1 in the left antennal lobe of Ir25a>dTomato-T2A-GCaMP6s animals with dTomato (left), GCaMP6s fluorescence at baseline (middle), and after CO2 presentation (right) with region of interest (yellow), in the indicated Gr3 genotypes.(O) GCaMP6s traces from Glomerulus 1 from animals of the indicated genotype presented with CO2 at the time indicated by the black bar, mean ± SEM, n=9-10 sweeps from 3 animals.(P) Summary of data in (O) indicating the maximum ΔF/F (arbitrary units).Lines denote median and quartiles, *p<0.05one-way ANOVA with Bonferroni correction.(Q) GCaMP6s fluorescence traces from Glomerulus 1 presented with the indicated stimuli.Black bar indicates stimulus presentation, mean ± SEM, n=5-10 sweeps from 3 animals/stimulus.(R) Summary of data in (Q) indicating the maximum ΔF/F (arbitrary units).Lines denote median and quartiles.*p<0.05one-way ANOVA with Bonferroni correction.(S) Summary of Glomerulus 1 responses with and without Gr3 function.CO2 concentration in all panels is 2%; see Methods for details of odorant concentrations.Scale bars: 25 µm.

Figure 6 :
Figure 6: Volatile amines substitute for CO2 in activating mosquito attraction to humans (A) Humans emit a complex blend of volatile odorants, including CO2, lactic acid, and volatile amines.(B-C) Schematic of behavioral assay (B) and percent wild-type mosquitoes of the indicated blood-feeding status attracted to the indicated stimuli in the assay (C).Lactic acid was supplied at 10 µl.Mean with range, n=11-15 trials, 19-21 mosquitoes/trial, *p<0.05,two-way ANOVA followed by Tukey's HSD.(D-I) Attraction of mosquitoes of the indicated genotype to the indicated stimuli.Lactic acid was supplied at 1 µl in D,F-I.Triethyl amine was presented at the indicated concentration in F, H and at 50 µl in E,G,I.Mean with range, n=7-16 trials, 15-22 mosquitoes/trial.*p<0.05,ns=not significant (*p>0.05),two-way ANOVA followed by Tukey's HSD.(J) Model of the compensatory pathway where volatile amines substitute for CO2 to enable attraction to humans by CO2-insensitive Gr3 mutants.

Figure S5 .
Figure S5.Projections of Ir76b-QF2-expressing neurons in the antennal lobe (Related to Figure 1) (A) Ir76b locus with exons (grey boxes), introns (grey lines) and CRISPR-Cas9 gRNA site (arrowhead) used to insert T2A-QF2 (light blue).(B) Quantification of the number of glomeruli that are GFP positive GFP negative (magenta), and total number of glomeruli (black).Analysis based on brains in (C-D) and Figure 1I,J.(C) Maximum-intensity projections of confocal Z-stacks of left antennal lobes from two different brains of the indicated genotype with immunofluorescent labeling of GFP (green) and Brp (synaptic marker, magenta) (top) and 2-D representation of the boundary of each glomerulus that is GFP positive and GFP negative (bottom).(D) Single confocal sections taken from the maximum-intensity projection confocal Z-stack of the left antennal lobe shown in Figure 1I with immunofluorescent labeling of GFP (green) and Brp (synaptic marker, magenta).A single plane is shown every 5 µm in Z to capture each glomerulus.Scale bar (C-D): 50 µm.Orientation: d=dorsal, m=medial.

Figure S7 .
Figure S7.Specificity of Split-QF2 reagents (Related to Figure 2) (A-B) Maximum-intensity projections of confocal Z-stacks of antennae (A) and maxillary palps (B) of the indicated genotypes showing intrinsic dTomato fluorescence and stained with the nuclear dye TO-PRO-3, with transmitted light overlay.(C) Maximum-intensity projections of confocal Z-stacks of left antennal lobes from two different brains of the indicated genotype with immunofluorescent labeling of dTomato (green) and Brp (synaptic marker, magenta) (top) and 2-D representation of the boundary of each glomerulus that is GFP positive and GFP negative (bottom).Scale bars: 50 µm.Orientation (C,D): d=dorsal, m=medial.

Figure S8 .
Figure S8.Projections of Orco-QF2-DBD; Ir25A-QF2-AD-expressing neurons in a single antennal lobe (Related to Figure 2) Single confocal sections taken from the maximum-intensity projection confocal Z-stack of the left antennal lobe shown in Figure 2G with immunofluorescent labeling of dTomato (green) and Brp (synaptic marker, magenta).A single plane is shown every 5 µm in Z to capture each glomerulus.Scale bar: 50 µm.Orientation: d=dorsal, m=medial.

Figure S9 .
Figure S9.Quantification of antennal and maxillary palp cell populations (Related to Figure 3 and Figure 4) (A-H) Workflow for cell quantification.Schematic of antennal region imaged on a confocal microscope (A) and image of maxillary palp with imaged area indicated with the yellow square (B).Whole-mount maxillary palp RNA in situ, yellow region from (C).Cells are manually marked independently as Orco+, Or49+, or Or8+ (red inset from B) using FIJI Cell Counter (D-F) and markers from each channel are merged (G).Cells with markers 1 and 2 are then scored as Orco+Or49+ with marker 4, and cells with markers 1 and 3 are then scored as Orco+Or8+ with marker 5 (H).Counts from each marker for each image are exported into Excel and R for further analysis.(I-N) Total cell counts from whole mount maxillary palp RNA in situs in Figure 4. Mean with range, n=5.(O) RNA in situ of whole-mount maxillary palp with the indicated probe and cell identity schematic.(P-Q) Whole-mount maxillary palp immunostaining showing Ir25a expression in "small" and "large" cells (P) and Gr3 expression in "large" cells and Orco protein in "small" cells (Q).Scale bars: 25 µm except (C): 50 µm.Orientation (B, P): distal right.