Like other mosquitoes, anophelines use olfaction to guide a range of behaviors. Females use odor cues to detect vertebrate hosts, sugar sources, and oviposition sites (Clements 1999). The role of olfaction in male anophelines is less well understood, although recent work by Mozüraitis et al. (2020) demonstrated that male An. gambiae Say and An. arabiensis Patton produce aggregating pheromones that attract males and virgin females from several of their sibling species, as well as the more distantly related An. funestus Giles. Furthermore, a contact pheromone, or one with low volatility has been postulated in this species (Pitts et al. 2014). Mating pheromones had been reported from several mosquito species previously. For example, Culiseta inornata (Williston, Diptera: Culicidae) uses a contact pheromone to identify conspecific females (Kliewer et al. 1966, Lang and Foster 1976, Lang 1977). This has also been demonstrated for various Aedes species (Nijhout and Craig 1971) and one study suggests that Ae. aegypti (L., Diptera: Culicidae) males produce volatiles that attract other males and females (Cabrera and Jaffe 2007).

Olfaction plays a role in male biology besides mating behavior as well. Like females, males use olfactory cues to locate sugar sources. These include volatiles produced by flowers and fermented fruit (reviewed in Pitts et al. 2014), with different plant species eliciting a variable response (Nyasembe et al. 2012). Based on a variety of studies looking at behavioral and electroantennographic responses, male and female culicines appear to have similar responses and preferences for plant volatiles (e.g., Jepson and Healy 1988, Jhumur et al. 2007, Otienoburu et al. 2012), and this has been found for An. arabiensis as well (Healy and Jepson 1988).

Finally, males of some mosquito species are attracted to vertebrate hosts to locate conspecific mates. For example, males of Aedes aegypti and Ae. albopictus Skuse congregate around humans (Lumsden 1957, McClelland 1959, Hartberg 1971, Gubler and Bhattacharya 1972). While this behavior has not been reported for anophelines, some data suggests that male An. coluzzii respond to human odor (Foster and Takken 2004). Although males were significantly less attracted to human odors than females during a dual olfactometer assay, a small proportion of males (10%) preferred human odor to clean air (Foster and Takken 2004). However, this result is not conclusive and needs confirmation. Certainly, male An. coluzzii are much less attracted to human hosts than are females, as human odor baited traps collect far fewer males than females, (e.g., Qui et al. 2007). Intriguingly however, an extensive effort to map An. coluzzii mating swarms showed that they are located around human habitation (Diabaté et al. 2011). Therefore, it is conceivable that human odor is used by males to position mating swarms in locations where females are more abundant.

Males of both An. coluzzii and An. quadriannulatus express the gustatory receptor Gr33 in their antennae at high levels, whereas this receptor is all but absent in female antennae (Pitts et al. 2011, Athrey et al. 2020). It is tempting to speculate that it plays a role in mating behavior, either during swarm formation or recognition of conspecific females.

The mosquito olfactory system is housed in three appendages: the antennae, the maxillary palps, and the proboscis (Pitts et al. 2004, Kwon et al. 2006). The neurons housed in the sensilla on these appendages express three types of olfactory receptors: the odorant receptors (Ors), variant ionotropic receptors (Irs), and gustatory receptors (Grs) (Clyne et al. 1999, Lu et al. 2007, Benton et al. 2009). Odorant binding proteins (Obps) in olfactory tissues are thought to carry hydrophobic odorants through the sensillum lymph to the receptors (Wojtasek and Leal 1999, Horst et al. 2001, Leal et al. 2005).

In contrast to the antennae, mosquitoes carry only a single type of olfactory sensillum on the maxillary palps: the capitate sensilla basiconica, also known as the capitate peg (McIver and Siemicki 1975). Anopheles gambiae sensu lato male palps have 28 capitate pegs, whereas female palps have 134 (McIver 1980). This relative scarcity of chemosensory sensilla suggests that the expression of chemosensory-related genes is lower in male maxillary palps. This hypothesis is supported by transcriptome analyses showing a lower abundance of reads associated with chemosensory-related genes in male maxillary palp than in those of females (Pitts et al. 2011).

Female An. coluzzii capitate pegs contain three chemosensory neurons. The cpA neuron co-expresses Gr22, Gr23, and Gr24, and responds to CO₂ and human skin odor (Lu et al. 2007, Tauxe et al. 2013). While it is well-known that female host-seeking is activated by CO₂ detection of the maxillary palps (Omer and Gillies 1971, Healy and Copland 1995, Lu et al. 2007), we know little about the ecological function of the male mosquito palp. We know that males of at least some mosquito species detect CO₂ with their palps because palp ablation largely removes their aversion for high CO₂ concentrations (Bässler 1958).

The cpB neuron co-expresses Or8 and Orco, and is highly sensitive to 1-octen-3-ol. Finally, the cpC neuron co-expresses Or28 and Orco and responds to a wider range of odorants (Lu et al. 2007). Besides the highly expressed Or8, Or28 and the Grs that encode the CO₂ receptor, other chemosensory receptors are expressed in female palps at lower level, such as IR100a, IR7w, IR75k, and Gr52 (Pitts et al. 2011, Athrey et al. 2017). Several of these receptors are differentially expressed in female palps between the anthropophilic An. coluzzii and its zoophilic sibling species An. quadriannulatus. Several odorant binding proteins (Obps) are also expressed at a higher level in the female palps of An. coluzzii than in those of An. quadriannulatus: Obp25, Obp26, and Obp57 (Athrey et al. 2017). Whether these differences in chemosensory gene expression show that the palps have functionally diverged to some extent between these species remains an open question.

Understandably female mosquitoes are studied extensively compared to males, and much remains unknown about male biology. However, as sterile-insect techniques and transgenic insect releases are being explored as vector control tools, interest in male behavior has increased, as mating competitiveness could be crucial to the success of such programs. Furthermore, malaria vector control might be improved by targeting anopheline males using a lure-and-kill strategy based on sound, visual or chemical cues (Diabaté and Tripet 2015). Currently, the role of the maxillary palps in anopheline males remains largely understudied. To facilitate work in this area, we compared chemosensory gene expression in the maxillary palps of males from the anthropophilic An. coluzzii and the zoophilic An. quadriannulatus. These results may provide an impetus for future work that may be able to put the observed differences in an ecological and behavioral context.

Methods

Results

The sequencing generated between 50.6 and 59.3 million reads for each of the four libraries. Between 88.9 and 89.2% of reads mapped uniquely for each library, resulting in 45.1 to 52.7 million mapped reads for each library ( Supp Table S1 [online only]). Of the 13,796 annotated genes, 9,843 were detected at >1 TPM (Fig. 1 and  Supp Fig. S1 [online only]) in at least one species. Of these, 3,371 were differentially expressed between species, with 1,817 and 1,554 genes expressed at higher levels in An. coluzzii and An. quadriannulatus, respectively. Of the differentially expressed genes, 839 were unique to An. coluzzii, and 167 were unique to An. quadriannulatus ( Supp Data S1 [online only]).

Overall Or expression was low compared to female palps (Pitts et al. 2011, Athrey et al. 2017), but showed a high correlation between the two species (R² = 0.718, Fig. 2A). Overall gene family expression levels were not subjected to a statistical test, but total expression of tuning Ors was 57.5 ± 6.5 TPM in An. coluzzii versus 79.1 ± 1.2 TPM in An. quadriannulatus. The expression of the co-receptor Orco was considerably lower than that of the tuning Ors combined (14.7 ± 6.6 vs. 42.8 ± 0.1 in An. coluzzii and 9.4 ± 4.4 vs.69.7 ± 5.5 in An. quadriannulatus). We detected only three Ors in An. coluzzii: Orco, Or8, and Or28. These genes are also the only Ors expressed in female palps of this species (Lu et al. 2007), but their expression in males is at a much lower level (Pitts et al. 2011). Interestingly, Or9 is expressed at 14.2 ± 0.0 TPM in An. quadriannulatus, while absent in An. coluzzii (q < 0.0001). In addition, we detected Or5 and Or29 just above our threshold in An. quadriannulatus (Fig. 2A,  Supp Data S1 [online only]).

Fig. 1.

Venn diagram indicating differential gene expression between the maxillary palps of male Anopheles coluzzii and Anopheles quadriannulatus. Genes are considered differentially expressed if q > 0.05 and fold change > 2.0.

Overall Ir expression is low in both species (24.8 ± 5.8 vs. 16.9 ± 6.2 TPM in An. coluzzii vs. An. quadriannulatus, respectively,  Supp Data S1 [online only]). There is a moderate correlation between the two species in Ir expression (R² = 0.53, Fig. 2B). Only the co-receptor Ir25a is differently expressed between species, with a 3.5-fold higher expression in An. quadriannulatus. In addition, five Irs (Ir64a, 41t.1, 41t.2, Ir76b, Ir75k) were detected at very low levels in An. coluzzii and/or An. quadriannulatus (Fig. 2B,  Supp Data S1 [online only]). Recently, an additional set of Irs was identified in the An. gambiae genome (Matthews et al. 2018); however, none were expressed in the male palps of either species.

The total level of Gr expression was 74.7 ± 5.0 in An. coluzzii versus 83.3 ± 8.0 TPM in An. quadriannulatus ( Supp Data S1 [online only]), but there is little correlation between the two species (R² = 0.20, Fig. 2C). Males of both species express the CO₂ receptor genes (Gr22-24), but with much higher expression of Gr22 than Gr23 and Gr24 in both species. When considering only those Grs that are expressed >5 TPM, Gr52 and Gr37 were expressed only in the male palps of An. coluzzii (25.8 ± 3.9 and 7.3 ± 3.9 TPM, respectively, Table 1). Gr60 was expressed only in the male palps of An. quadriannulatus (16.8 ± 3.5 TPM, q < 0.0001, Table 1). In addition, seven Grs (Gr10, Gr21, Gr26, Gr40, Gr42, Gr51, Gr59) were detected at low level in one or the other species (Fig. 2C,  Supp Data S1 [online only]).

Fig. 2.

A comparison of the expression in transcripts per million (TPM) of four chemosensory gene families in male maxillary palps of Anopheles coluzzii vs. Anopheles quadriannulatus. A: Odorant receptors (Ors), (B): Ionotropic receptors (Irs), C: Gustatory receptors (Grs), D: Odorant binding proteins (Obps). The line indicates equal expression between the two species. Genes with differential expression (q < 0.05, fold-change >2) are indicated by grey dots. Only genes expressed at >1 TPM are shown.

Overall Obp expression was 6,895 ± 1,127.1 TPM in An. coluzzii versus 6,585 ± 306.7 TPM in An. quadriannulatus ( Supp Data S1 [online only]). The correlation of Obp expression between the species is relatively high (R² = 0.72, Fig. 2D). When considering Obps expressed above >50 TPM, only three Obps showed differential expression between the two species: Obp72 and Obp19 are expressed at higher level in An. coluzzii, although still at low levels relative to overall Obp expression (91.7 ± 2.2 and 54.0 ± 10.3 TPM, respectively, Table 1). The expression of Obp48 is significantly higher in An. quadriannulatus. This gene is the most highly expressed gene among the four chemosensory gene families considered here and is the 76^th most highly expressed gene overall in the male palps of An. quadriannulatus. Finally, an additional 23 detected Obps are expressed at low levels in one or both of the two species (Fig. 2B,  Supp Data S1 [online only]).

Table 1.

Chemosensory genes expressed at >5TPM (chemoreceptors) and > 50TPM (odorant binding proteins) in the male maxillary palps of Anopheles coluzzii versus Anopheles quadriannulatus

Gene ontology analyses of molecular function showed that among the 1,423 genes that were differentially expressed between species and that showed functional hits, 525 were classified as having ‘binding' (GO:0005488) as a molecular function, while 573 were classified as having ‘catalytic activity' (GO0003824). Genes classified under molecular function categories of ‘structural constituent ribosome' and ‘protein binding’ were significantly over-represented among differentially expressed genes (Fisher exact test; q = 0.019 and 0.015, respectively).

Discussion

Previously, we and others have shown species-, sex- and organ-specific patterns of chemosensory gene expression in the anthropophilic An. coluzzii and zoophilic An. quadriannulatus (Pitts et al. 2011, Rinker et al. 2013, Athrey et al. 2017, Saveer et al. 2018, Athrey et al. 2020). Here, we extended that work by comparing chemosensory gene expression in the male maxillary palps of these two species. As expected, the repertoire of chemosensory genes expressed in male palps is small. While much of it is shared between the two species, several chemosensory genes are highly species-biased or specific; specifically, Gr37 and Gr52 in An. coluzzii, and Or9 and Gr60 in An. quadriannulatus.

The function of the maxillary palps in male anopheline biology has remained largely unstudied. Nonetheless, a comparison of the species-specific or species-biased gene expression in the male maxillary palps provides a starting point for further inquiry. For example, male-specific chemosensory genes could play a role in mating biology (e.g., mate recognition), whereas female-specific chemosensory genes are more likely to play roles in behaviors such as oviposition and host-seeking.

The primary role of the maxillary palps in females is the detection of CO₂ (Lu et al. 2007), which is an important host kairomone for mosquitoes (Gillies 1980). The expression of CO₂ receptor genes in males suggests they detect this chemical, although the relatively low expression level suggests they do so with less sensitivity than females. This was not found in Ae. aegypti, in which CO₂ sensitivity appears to be similar in females and males (Grant et al. 1995). However, to our knowledge, the expression level of CO₂ receptors has not been compared between the two sexes in this species.

One possibility is that CO₂ detection in males may be useful during nectar seeking. Flowers emit more CO₂ when producing nectar, and hence this could signal an abundant food source (Thom et al. 2004). However, at higher concentrations, CO₂ likely repels male anophelines, as CO₂-baited traps rarely catch any male mosquitoes (e.g., Kweka et al. 2013, McPhatter and Gerry 2017). Carbon dioxide avoidance has been demonstrated in males of Ae. aegypti, who avoid airstreams with enhanced with CO₂ in an olfactometer (Bässler 1958).

The capitate pegs on the maxillary palps of An. coluzzii females are innervated by three olfactory neurons, designated the cpA, cpB, and cpC neurons (Lu et al. 2007). The cpA neuron expresses the CO₂ receptors, the cpB neuron co-expresses Orco and Or8, and the cpC neuron co-expresses Orco and Or28 (Lu et al. 2007). These genes are among the most highly expressed chemoreceptors in the male palps as well, and presumably the capitate pegs on the male palps contain the same three neurons. In An. coluzzii, Or8 is narrowly tuned to detect 1-octen-3-ol (Lu et al. 2007), an odorant found in both the breath and sweat of large herbivores and humans (Hall et al. 1984, Cork and Park 1996) and that serves as an attractant to Anopheles and Aedes mosquitoes (Takken and Kline 1989). Our results indicate that male Anopheles are capable of detecting 1-octen-3-ol as well. Nonetheless, the importance of octenol to host seeking is not entirely clear. For example, Ae. aegypti and An. quadrimaculatus Say females with removed palps readily locate a human arm (Roth 1951). Additionally, some plants also produce octenol and attract mosquitoes (Syed and Guerin 2004, Impoinvil et al. 2004). Work by Dekel et al. (2016) suggests that 1-octen-3-ol fulfills additional ecological roles beside host detection, as Or8 in the female palps of non-blood feeding Toxorhynchites amboinensis (Doleschall, Diptera: Culicidae) also responds to 1-octen-3-ol.

Or28 is more broadly tuned (Lu et al. 2007). Out of a panel of 82 odorants tested, it shows its strongest response to 2,4,5-trimethylthiazole and acetophenone, two odorants not associated with humans (Lu et al. 2007). Acetophenone is a plant volatile which also elicits responses in other insect species (Wright et al. 2005). This could suggest a role for Or28 in nectar seeking (Lu et al. 2007). However, Or28 also responds to 6-methyl-5-hepten-2-one (sulcatone) (Lu et al. 2007). Sulcatone was recently identified as one of five aggregation pheromones produced by male An. gambiae s.s. and An. arabiensis, although it is also a constituent of human sweat (Meijerink et al. 2000). A blend of these pheromones attracts males and females from a range of Anopheles species, including An. coluzzii (Mozūraitis et al. 2020). An. quadriannulatus was not included in this study, but given that the effect of the pheromones extends beyond the An. gambiae species complex to An. funestus, it is reasonable to assume that An. quadriannulatus is attracted to these as well. Certainly, the high expression of Or28 in the male and female palps of both species indicates that its role is conserved both between sexes and species. This is consistent with a role either in nectar seeking or in mating behavior. Interestingly, the expression of Or8, Or28, and Gr22 in females increases several-fold between 1 vs. 4 d post-emergence, which correlates with the onset of both host-seeking and mating (Omondi et al. 2015).

In female palps, several neurons express Orco without either Or8 or Or28 (Lu et al. 2007). It is possible that these neurons co-express Orco with the other chemoreceptors that are expressed at low levels (Pitts et al. 2011, Athrey et al. 2017). In Drosophila, there is some precedence for the co-expression of Orco and Irs, as DmOr35a, DmOrco, and DmIr76b are co-expressed in a subpopulation of coeloconic olfactory sensory neurons. However, this is exceptional, as Irs do not generally co-express with Orco (Benton et al. 2009). Additional chemoreceptors are expressed in males as well, and a notable difference between males and females is the relatively high expression level of these. At the moment, no information is available on which neurons express these genes in the male palps.

In contrast to the male antennae (Athrey et al. 2020), no chemosensory genes in the palps were male-specific in these species, i.e., all the genes expressed in males are expressed in females as well (Pitts et al. 2011, Athrey et al. 2017). However, several chemosensory genes were species-specific when comparing An. coluzzii and An. quadriannulatus male palps, which could indicate divergent odor responses between the males of these species. The gustatory receptors Gr37 and Gr52 are expressed only in An. coluzzii. Intriguingly, the expression of Gr52 in An. coluzzii exceeds that of all other chemoreceptors and the expression of this gene is also exclusive or highly biased towards An. coluzzii in male and female antennae, as well as female palps (Athrey et al. 2020). Furthermore, the labella of both sexes of An. coluzzii express Gr52 as well (Saveer et al. 2018). The expression pattern of Gr37 is less consistent in other olfactory organs, as it is more highly expressed in male An. quadriannulatus versus An. coluzzii antennae, although at low levels (Athrey et al. 2020). It was not detected in the female antennae or palps of either species (Athrey et al. 2017).

The species-specific expression of particularly Gr52, and to lesser extend Gr37, suggests they could play a role in divergent behaviors between these two closely related species. On the other hand, both Gr52 and Gr37 were also detected in the whole male body by Pitts et al. (2011), suggesting a role beyond olfaction or gustation, such as internal nutrient sensing (Miyamoto et al. 2012). Given these expression patterns, Gr52 and Gr37 may perform multiple functions in An. coluzzii. Neither gene has an identified ortholog in Drosophila, but whereas Or37 has orthologues throughout the Culicidae, Gr52 appears to be specific to Anopheles.

Both Or9 and Gr60 are exclusive to An. quadriannulatus in male palps. However, Or9 is more highly expressed in the female antennae of An. coluzzii versus An. quadriannulatus (Pitts et al. 2011, Athrey et al. 2017) but at similar levels in male antennae of the two species (Athrey et al. 2020). This gene is not expressed in the female palps of either species (Athrey et al. 2017, Pitts et al. 2017). A de-orphanization study showed that Or9 has an intermediate tuning breadth and that it is highly sensitive to 3-methylphenol, 4-methylphenol (major component of pig odor and cattle urine), 2-ethylphenol, 4-ethylphenol (yeast), and 2,4,5-trimethylthiazole (fox urine, wheat flour, cooked beef) (Carey et al. 2010). The significance of this receptor in male biology is unclear, but one possibility might be that male An. quadriannulatus use host odors to position swarming sites near locations where females may be more abundant.

We know nothing about the function of Gr60. It does not have a known ortholog in Drosophila, and like Gr52, appears to be present only in Anopheles. This gene is specific to the maxillary palps of An. quadriannulatus in both sexes (Athrey et al. 2017, Athrey et al. 2020), but is absent from their antennae (Athrey et al. 2017, Athrey et al. 2020) and the An. coluzzii labellum (Saveer et al. 2018).

Besides the chemosensory genes expressed only in males of one of the two species, several showed differential expression. However, none of these genes stand out as particularly noteworthy. The An. coluzzii-biased Obp19 and Obp72 are expressed at very low level relative to overall Obp expression. Two chemosensory genes, Ir25a and Obp48 were expressed at higher level in An. quadriannulatus male palps. Ir25a is a co-receptor, whose function has been linked to the detection of amines (Ai et al. 2010, Abuin et al. 2011, Silbering et al. 2011, Pitts et al. 2017), and cool sensing (Ni et al. 2016). However, no other Irs were expressed in the male palps for Ir25a to form complexes with other than one of the other co-receptors Ir76b. Obp48 is among the most highly expressed Obp in the antennae of both sexes in both species (Pitts et al. 2011, Athrey et al. 2017), and is highly expressed in the labella of both sexes of An. coluzzii as well. (Saveer et al. 2018).

Our comparison between the male palps of An. coluzzii and An. quadriannulatus also provides some context for previous work on females. In particular, the expression of Ir7w, Ir41n, and Ir100a is highly biased towards An. coluzzii in female palps, whereas Ir41a, Gr21, and Gr26 are exclusive to An. quadriannulatus in female palps (Athrey et al. 2017). Except for Gr21, these genes are absent from the male palps of both species, suggesting they may play a role in divergent female odor responses between species. Gr21 is exclusive to An. quadriannulatus in male palps as well, although at a much lower level. Interestingly, Gr21 is an ortholog of sugar receptors in Drosophila and is also expressed in An. coluzzii labella in both sexes (Saveer et al. 2018). Its exclusive expression in the palps of An. quadriannulatus is somewhat puzzling. In Drosophila, sugar receptors are multimeric heteromers (Fujii et al. 2015). No other sugar receptor genes were detected in the palps, so it is not clear that Gr21 encodes a sugar receptor in An. quadriannulatus male palps. In Drosophila some sugar receptors are also expressed in olfactory neurons, and it has been speculated that these may dimerize with Orco to form a receptor tuned to new ligands (Fujii et al. 2015). This could also be the case in Anopheles.

Despite the recent exciting work on aggregation pheromones by Mozūraitis et al. (2020), important questions surrounding the role of chemosensation in anopheline male biology remain outstanding. For example, it is not known if anophelines use a sex pheromone for species recognition, although a close contact pheromone has been postulated (Pitts et al. 2014). The fact that the aggregation pheromones attract various species of the An. gambiae complex equally, but that very low levels of hybridization are observed is consistent with the presence of a contact or low volatility pheromone. Chemoreceptors with a species-specific expression profile could play a role in these. That being said, species recognition within swarms can also be explained at least partially by acoustic cues (Cator et al. 2010, Pennetier et al. 2010). Furthermore, we suspect that males use host odor in positioning mating swarms, although the data supporting this is circumstantial at present (Foster and Takken 2004, Diabaté et al. 2011). Therefore, many of the odor cues that could play an important role in male anopheline biology are expected to be species-specific, and the chemosensory genes that were shown to have species- and/or sex-specific expression in the maxillary palps of these species may be promising candidate genes underlying various components of male mating behavior.

In conclusion, we characterized differences in chemosensory gene expression between the male maxillary palps of An. coluzzii and An. quadriannulatus. The expression of several chemoreceptors is specific to either An. coluzzii or An. quadriannulatus. We postulate that these genes may underly species-specific behavior, such as close contact mate recognition. Regardless, the results present here may be useful for future work on the functional divergence of the maxillary palps between the malaria vector An. coluzzii and one of its sibling species.

Supplementary Data

Supplementary data are available at Journal of Medical Entomology online.

Figure S1. Plot of Log₂FC (Fold Change) against -log¹⁰(q-values) between the maxillary palps of Anopheles. coluzzii vs. Anopheles quadriannulatus males.

Table S1: Mapping statistics for male maxillary palps samples.

Supplemental Data S1: Gene expression (TPM) and differential gene expression data for all genes in the male maxillary palps of Anopheles coluzzii vs. Anopheles quadriannulatus.

Data Availability

The sequence data used in this study is deposited on the NCBI Sequence Read Archive, under BioSample accessions SAMN15539281–SAMN15539284. The full gene expression data are available in  Supp data S1 [online only].