Phenotypic variation in biting behavior associated with differences in expression of olfactory genes in the vector mosquito, Aedes albopictus (Diptera: Culicidae)

Abstract

We evaluated miRNA and mRNA expression differences in head tissues between avid-biting vs. reluctant-biting Aedes albopictus (Skuse) females from a single population over a 20-min timescale. We found no differences in miRNA expression between avid vs. reluctant biters, indicating that translational modulation of blood-feeding behavior occurs on a longer timescale than mRNA transcription. In contrast, we detected 19 differentially expressed mRNAs. Of the 19 differentially expressed genes at the mRNA level between avid-biting vs. reluctant-biting A. albopictus, 9 are implicated in olfaction, consistent with the well-documented role of olfaction in mosquito host-seeking. Additionally, several of the genes that we identified as differentially expressed in association with phenotypic variation in biting behavior share similar functions with or are inferred orthologues of, genes associated with evolutionary variation in biting behaviors of Wyeomyia smithii (Coq.) and Culex pipiens (Lin.). A future goal is to determine whether these genes are involved in the evolutionary transition from a biting to a non-biting life history.

Introduction

In many mosquito spp., blood feeding allows females to obtain protein- and lipid-rich blood that can be used for egg production. Blood feeding is comprised of a sequential series of behaviors. First, host-seeking involves locating a potential blood meal host, landing on the host’s skin, and evaluating the host skin chemistry using both olfactory and gustatory perception. These behaviors are followed by biting, when the mosquito fascicle penetrates the host skin, and then probing, when the fascicle searches for a blood vessel. Finally, the imbibing phase occurs when a blood vessel is lacerated and the blood is ingested through the fascicle (Clements 1992). In addition to providing resources for reproduction, blood feeding is also the process by which female mosquitoes transmit a wide variety of human and veterinary pathogens, including Plasmodium spp., viruses, and filarial worms (Lwande et al. 2020). Thus, there is a large body of work in mosquitoes focused on the molecular and physiological basis of host-seeking and biting behavior (Klowden 1997, DeGennaro et al. 2013, McBride et al. 2014, McMeniman et al. 2014, Ni et al. 2022). A major rationale for this body of work is to identify molecular targets to disrupt host-seeking and/or biting behavior as a mechanism to suppress disease transmission (Duvall 2019).

The molecular and physiological processes that occur during host-seeking have been elucidated in relatively specific detail in mosquitoes. However, the processes that mediate biting behavior once a mosquito has landed on its vertebrate host remain less well understood. Female mosquitoes integrate olfactory, visual, and thermal cues to locate a potential host; CO₂ and skin odors appear to be particularly important cues, but no single odor or sensory modality suffices to stimulate host-seeking (DeGennaro et al. 2013, McMeniman et al. 2014, Giraldo et al. 2023). Once a mosquito lands on the skin of its vertebrate host, biting can be stimulated by thermal cues alone (Corfas and Vosshall 2015), although gustatory reception may also be important. Moreover, in addition to external environmental stimuli, the internal physiological status of a female mosquito related to factors such as age, nutrition, hydration, and even pathogen infection can affect individual variation in host-seeking and biting behavior (Klowden 1990, Cator et al. 2013, Hagan et al. 2018).

One strategy for exploring the molecular physiology of host-seeking and biting behavior has been to identify genes whose transcription patterns shift in response to the manipulation of physiological or environmental factors that affect these behaviors. For example, several studies have examined the effect of age or nutrition on gene expression and host-seeking and/or biting behavior (Dittmer et al. 2019, Omondi et al. 2019, de Carvalho et al. 2021). Furthermore, multiple studies have demonstrated that imbibing a blood meal makes females refractory to host-seeking and biting for several days while the blood is digested and eggs are matured (Klowden and Lea 1979, Klowden 1990, Takken et al. 2001), and several studies have sought to identify the molecular basis of this blood-meal-induced suppression of host-seeking and biting (Matthews et al. 2016, Duvall et al. 2019, Ni et al. 2022). Together, these studies implicate the expression of genes involved in olfaction and other chemosensory processes as mediating plasticity in host-seeking and biting, although neuroendocrine (Duvall et al. 2019) and energy utilization processes (de Carvalho et al. 2021) can also be important.

An additional approach to determining mechanisms underlying host-seeking and biting is to focus on evolved differences in blood feeding (Bradshaw et al. 2018, Siperstein et al. 2022, Marzec et al. 2023). In fact, the evolutionary transition from a blood-feeding to non-blood-feeding life history has occurred multiple times in mosquitoes; 3 entire genera of Culicidae never blood feed (Malaya, Topomyia, and Toxorhynchites), and several genera are comprised mostly of species that do blood feed, but also include obligately or facultatively non-blood-feeding species (Downes 1958, Mogi 1984, Foster 1995, Rattanarithikul et al. 2007, Wahid et al. 2007, Miyagi et al. 2012, Zhou et al. 2014). The repeated, independent evolution of a non-blood-feeding life history in mosquitoes has likely been driven at least in part by the significant physiological costs associated with blood feeding, including the energetic costs of host-seeking (Edman and Scott 1987); ingesting viscous, hot blood (Benoit et al. 2011, Lahondère and Lazzari 2012, de Carvalho et al. 2021, Marzec et al. 2023); and detoxifying heme and iron when blood is digested (Graça-Souza et al. 2006). Moreover, it is becoming increasingly clear that the physiological and metabolic costs of blood feeding can occur in anticipation of a blood meal, before blood is imbibed, rather than as a response to actually ingesting blood (Bradshaw et al. 2018, de Carvalho et al. 2021, Siperstein et al. 2022, Marzec et al. 2023).

Here, we investigate the transcriptional differences underlying phenotypic variation in biting behavior in the invasive vector mosquito, Aedes albopictus (Skuse). We quantified differences in messenger RNA (mRNA) and microRNA (miRNA) abundance in head tissues of “avid-biting” females that initiated biting during 5 min of exposure to a human host, and “reluctant-biting” females that never landed on or bit the human host during 20 min of exposure (Fig. 1). We used head tissues in an attempt to identify differences in gene expression related to sensory perception associated with host-seeking and biting. Our experimental design is distinct from previous studies on this topic because we did not manipulate physiological or environmental factors known to affect host-seeking and biting behavior (i.e., adult age, nutritional status, blood-feeding history, etc.). Instead, we reared the experimental cohorts of mosquitoes under near-optimal conditions in which all individuals had access to abundant nutritional resources at both the larval and adult stages, then used females of the same age (days post-eclosion) for testing. Our rationale is that by using rearing conditions in which experimental cohorts had access to the same near-optimal resources, we are reducing (but not eliminating) environmentally induced physiological variation, thereby increasing the opportunity to identify genetically based transcriptional variation. We then compare our results to previous studies of mRNA and miRNA expression between avid vs. reluctant-biting and non-biting females of mosquitoes from 2 other genera, Wyeomyia smithii (Coq.) (Bradshaw et al. 2018) and Culex pipiens (Lin.) (Siperstein et al. 2022, Marzec et al. 2023).

Materials and Methods

Insect Colony Maintenance

The stock population used for this study was established with pupae and larvae collected from an auto-salvage yard located in Manassas, Virginia in 2018 (MAN population in previous studies from this lab). The population was reared under standard laboratory conditions as described previously for 8 subsequent generations (Batz et al. 2020). All mosquitoes were reared in the same temperature-controlled walk-in incubator held at 21 °C, 80% relative humidity, and long-day (LD) photoperiod (16:8, light:dark), unless otherwise noted.

The parental generation of experimental mosquitoes was reared under standardized conditions as follows: F₈ eggs were stimulated to hatch in a 5.5-L Sterilite container containing 2 L of deionized (DI) water and approximately 3 mL of larval food slurry. The food slurry was made by combining 120 g of dog food (Nutro Ultra Small Breed Puppy, Nutro Products Inc., Franklin, TN) and 40 g of frozen brine shrimp (Sally’s Frozen Brine Shrimp, San Francisco Bay Brand, Newark, CA) in 1 L of DI water. Every Monday-Wednesday-Friday (M-W-F), larvae were moved to fresh DI water and given approximately 2 mL of food slurry. Pupae were collected and transferred to adult cages (2.5-gallon plastic buckets with mesh windows) and provisioned with organic raisins (Newman’s Own, Westport, CT) to allow ad libitum sugar feeding. A total of 3–5 days after all adults had eclosed, females were offered a human blood meal. The Georgetown University Institutional Review Board (IRB) has determined that mosquito blood feeding is not human research and thus does not require IRB approval; however, the blood-feeding protocol has been approved by the Georgetown University Occupational Health and Safety Office. Following the bloodmeal, eggs were collected 3 times weekly for 2 weeks as previously described (Batz et al. 2020).

All rearing of the experimental F₉ generation was done at 28 °C with LD photoperiod and 80% relative humidity in the same Percival I-36VL incubator (Percival Scientific, Perry, IA). We adjusted the larval-rearing conditions of the experimental animals in the F₉ generation relative to the conditions described above to more closely match previous studies of biting behavior in W. smithii (Bradshaw et al., 2018) and C. pipiens (Siperstein et al. 2022, Marzec et al. 2023). The F₉ eggs were stimulated to hatch in two 8.5-L rearing pans each containing 2 L of DI water and 100 mg of fish food (TetraMin Tropical Flakes, Spectrum Brands Pet LLC, Blacksburg, VA). Within 3 days of hatching, groups of 200 larvae were collected and each group was placed in a separate 5.5-L Sterilite container with 2 L of DI water and 100 mg of fish food. New food and water were provided to larvae every M-W-F. Pupae were collected 4 times weekly and sorted by sex based on the morphology of the genital lobe; males were transferred to rearing cages and females were placed into a separate experimental cage (bioQuip cages 30.48 cm × 30.48 cm × 30.48 cm). To ensure that females within each experimental cage were eclosed within the same 24 h, cups containing pupae were moved to a new cage each day between Zeitgeber time (ZT) 10 and 13.5. In order to provide females with an opportunity to mate before the biting trials (see below), adult males were added to each cage with experimental females to achieve an approximately 5:1 female:male ratio. All experimental cages were given organic raisins (Newman’s Own, Westport, CT), 8% sucrose solution, and a mesh-covered water cup to maintain humidity.

Female biting trials were conducted 72–96 h post-eclosion between ZT 12 and 13.5 of an 16:8 (light:dark) cycle, corresponding to 4–2.5 h before dark. This post-eclosion time range was chosen because most anautogenous mosquitoes develop the functional competence to blood feed between 24 and 78 h after eclosion (de Carvalho et al. 2021, Hill and Ignell 2021). As in our prior experiment that measured mosquito biting responses (Siperstein et al. 2022), sugar sources (raisins and sucrose solution) were removed the day before the trial. For each trial, a single adult cage was removed from the incubator and between 130 and 200 females were transferred to a 30.48 cm (height) × 30.48 cm (width) × 53.34 cm (length) plexiglass cage with two 11.43-cm-diameter holes to allow access to the cage. Females were allowed to acclimate to room temperature (20–23 °C) in the plexiglass cage for approximately 15 min. Human breathing into the cage was performed for approximately 1 min before a human blood meal source (a human hand) was introduced inside the cage. Females that landed and initiated biting of the host within 5 min of exposure to the hand were aspirated before they had imbibed any blood, transferred into a collection tube, snap-frozen in liquid nitrogen, and stored at −80 °C. These samples comprise the “avid biters” (Fig. 1A). Human breath was introduced into the cage at 5-min intervals for a 15-min period, and any mosquitoes that landed and probed the host during this time were aspirated out of the cage and discarded. Females that still had not landed and bit the host 20 min after the hand was first inserted into the cage were aspirated into a separate collection tube, snap-frozen in liquid nitrogen, and stored at −80 °C. These samples constitute the “reluctant biters” (Fig. 1A). Although reluctant biters did not attempt to bite at any point during the 20-min trial, we assume that they eventually would have because the incidence of autogeny in this population is very low (~1%, personal observation). We conducted a total of 5 trials over the course of 3 days, with a maximum of 2 trials per day. After all trials were complete, we placed the mosquitoes that had been collected from individual trials on dry ice and examined them for the presence of blood in their abdomens. Heads from female mosquitoes that did not contain blood were retained and stored at −80 °C (n = 33–97 heads per biological replicate).

RNA Extraction, Library Preparation, and Sequencing

After the heads were separated from the bodies as described above, 500 ul of TRIzol (Invitrogen) was added to each biological replicate. The samples were shipped to the University of Oregon Genomics and Cell Characterization Core Facility for tissue homogenization, RNA extraction, and integrity assessment. We selected 4 biological replicate samples of avid biters and 4 biological replicate samples of reluctant biters based on total RNA quantity and quality. cDNA library preparation and sequencing were then performed as in Siperstein et al. (2022) for mRNAs and Marzec et al. (2023) for miRNAs. Briefly, each sample was homogenized using Silica grinding beads in a Spex Genogrinder 2100 (2 × 1500 RPM for 2 min). RNA was extracted using a Zymo Direct-Zol kit (Zymo Research) followed by isolation of mRNA with Oligo dT beads according to the manufacturer’s instructions. The 4 biological replicate samples from avid-biting females and the 4 biological replicate samples from reluctant-biting females were assessed for integrity on an RNA chip (Bioanalyzer 2100). These samples were then used for paired-end, barcoded, and stranded library construction with the Universal Plus mRNA-Seq protocol (Tecan Genomics). These 8 libraries were then sequenced on a single lane of a NovaSeq 6000 instrument.

For microRNA sequencing, the same 8 total RNA samples described above (4 avid biting, 4 reluctant biting) were used to create small-RNA sequencing libraries with Perkin Elmer’s NextFlex small RNA-seq kit v3 (NOVA-5132-05) according to the manufacturer’s instructions and as described in Marzec et al. (2023). Size selection was performed according to the gel-free method as described in the manual. All 8 libraries were combined in equimolar ratios and single-end, 160-bp reads were sequenced on a single lane of NovaSeq 6000. The raw reads from both mRNA and miRNA sequencing are available in NCBI’s short read archive under accession PRJNA1022310.

mRNA Bioinformatics Analyses

Details of the bioinformatics workflow can be found on the GitHub repository here: https://github.com/mch246/Ae.-albopictus-avid-vs-reluctant-biting-RNAseq-/blob/main/MasterNotes.md. Briefly, reads were filtered for quality and adapters using Trimmomatic (version 0.39) with an additional parameter of HEADCROP: 15 to remove the first 15 bases from each read. Reads were then mapped using a 2-pass method in STAR (version 2.7.1a) to the most recent A. albopictus reference genome, AalbF3 (Boyle et al. 2021), obtained from NCBI (GCA_018104305.1). Read counts were obtained from the resulting alignment (bam) files with HTSeq (version 0.13.5). We retained for analysis transcripts that had at least 10 reads across all 8 samples. Read counts were transformed to a log₂ scale using rlog and then a principal components analysis (PCA) was performed using plotPCA in DESeq2 (version 1.34.0; Love et al. 2014) to determine if biological replicate samples within treatments (avid biting, reluctant biting) exhibited similar overall transcriptional profiles. Next, differential expression analysis was performed using DESeq2 on normalized read counts in R (version 4.1.3). Differentially expressed genes (DEGs) were identified as genes with a Benjamini–Hochberg false discovery rate adjusted P-value less than 0.05 and a log₂-fold change >|1|. To determine if the DEGs from our study were also differentially expressed in association with differences in biting behavior in W. smithii (Bradshaw et al. 2018) and C. pipiens (Siperstein et al. 2022), we searched for putative orthologs by performing a reciprocal best Blastp hit search. Because DEGs associated with biting were identified using a microarray in W. smithii, we first converted the list of differentially expressed microarray sequences from Bradshaw et al. (2018) to a list of NCBI gene names (GCF_029784165.1) using Blastx before performing the Blastp search.

KEGG Pathway Analysis

All A. albopictus Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and genes within those pathways were downloaded from KEGG using KEGGREST (version 1.30.1) in R and the number of DEGs relative to the total number of genes in each pathway was determined. Enrichment was tested using a Wilcoxon Rank Sum test. We defined pathways as enriched if they had a significant P-value of < 0.05 and more than 4 DEGs.

miRNA Bioinformatics Analyses

Detailed notes and workflow used for the miRNA analyses can be found at this Github repository: https://github.com/srmarzec/albopictus_biting_miRNA/blob/main/MasterNotes.md. Briefly, reads were cleaned with Trimmomatic to remove Illumina small RNA NexFlex adapters and low-quality reads. Additionally, the first 4 bases were removed from the single-end reads. Next, reads were filtered to remove tRNA and rRNA sequences by aligning the reads to the tRNA and rRNA features from the A. albopictus reference genome using bowtie2 (v2.4.4) and retaining all unaligned reads. The reads were then size sorted with a custom python script, retaining only reads that fell within 18–24 base-pair size range, the expected size of mosquito miRNAs (Hong et al. 2014).

miRNA identification and mapping were done with miRDeep2 (miRDeep2.0.1.3; Friedlaender et al. 2012). Indexing for miRDeep2 was performed with Bowtie (v1.3.1). Known mature miRNA sequences were obtained for A. albopictus from previous studies (Skalsky et al. 2010, Gu et al. 2013, Chen et al. 2015, Batz et al. 2017, Palatini et al. 2020). Exact sequence duplicates were removed from the combined list resulting in 304 unique known mature miRNA sequences (Supplementary Table S1). Additionally, 31 novel miRNAs with unique precursor sequences were identified by miRDeep2 in our dataset; all of these miRNAs passed thresholds of having 10 reads for both mature and star miRNA sequences, a significant randfold P-value, and a miRDeep score greater than 4 (Supplementary Table S1). With the combination of known and newly identified miRNAs in this work, miRDeep2 generated a list of 201 unique matched precursor and mature miRNAs that had at least 10 reads across all samples (Supplementary Table S2).

After obtaining read counts from miRDeep2, all downstream analysis was performed in R (version 4.0.2). Principal component analysis and differential expression were performed using the same process as for the mRNA data with DESeq2 (version 1.30.1) as described above. The sample NB9 was removed from the miRNA dataset as an outlier since it contained a small number of reads (8,087) compared to the other replicate samples (each of which had > 12 million reads); excluding this sample had no effect on the results.

Quantitative Reverse Transcription (qRT)-PCR

Quantitative real-time PCR (qPCR) of the same biting and non-biting A. albopictus samples that were used for RNAseq analyses was used to confirm differential expression of 10 genes from the mRNA results. cDNAs for each sample were synthesized using 200 ng of total RNA and the Maxima First Strand cDNA Synthesis Kit with dsDNase (Thermo Fisher) according to the manufacturer’s instructions. Primers (Supplementary Table S3) were designed using Primer3 (Untergasser et al. 2012). Prior to qPCR analyses, melt and standard curves were run to ensure that each primer pair met MIQE specificity and efficiency guidelines (Bustin et al. 2010; Supplementary Table S3 contains information about the standard curve efficiency and R² values). Quantitative reverse transcription-PCR (qRT-PCR) was performed in a 96-well plate using a CFX Connect qPCR detection system (Bio-Rad). All reactions were performed in triplicate in a total volume of 10 μL containing 5-μL iTaq Universal SYBR green PCR Master Mix (Bio-Rad), 400 nM of each primer, and 1-μL sample cDNA.

The qPCR data were analyzed by first averaging the relative cycle threshold (CT) of 3 technical replicates. The relative mRNA abundance was calculated by subtracting the average of the CT values of 2 reference genes, Rp49 and RpL34 (Urbanski et al. 2010), from the CT value for the gene of interest within the same biological replicate (2^−ΔCT method; Schmittgen and Livak 2008). A Student’s t-test was then used to compare the average relative expression of the gene of interest between the head tissues of avid-biting vs. reluctant-biting A. albopictus with 4 biological replicates per treatment (α = 0.05).

Results

Biting Trials

Averaging across the 6 biting trials, approximately 50% of females landed on the human blood meal host and initiated biting within 5 min of exposure. These females were classified as “avid biters.” Approximately 30% of females did not land on the human host and attempt to bite at any time during the 20-min trial period and were classified as “reluctant biters.” Of the remaining females, approximately 20% landed on the human host and initiated biting between 5 and 20 min after the start of the trial; these females were categorized as “intermediate biters” and were discarded (Fig. 1B).

Messenger RNAs (mRNAs)

RNA sequencing for 4 biological replicate samples of avid-biting and 4 biological replicate samples of reluctant-biting mosquito-head tissue produced between 27,236,164 and 31,025,831 read pairs per sample. After quality filtering with Trimmomatic, between 81% and 85% of reads were retained. The STAR 2-pass alignment produced alignment rates between 71% and 74.5% per sample (Supplementary Table S4). A total of 14,746 genes had at least 10 mapped read pairs from summed across all 8 samples. Principal component analysis showed that the transcriptional profiles of avid-biting samples clustered separately from the reluctant-biting samples, with distinct separation on the second principal component axis, which explained 25% of the variance in gene expression (Figure 2). Nineteen genes were significantly differentially expressed between avid-biting and reluctant-biting samples (adjusted P-value < 0.05, log₂FC > |1|; Fig. 2, Table 1). No putative orthologs associated with biting were identified between A. albopictus and W. smithii. However, 2 genes that were upregulated in avid biters of A. albopictus in the current study (LOC115268560, LOC109418392) were identified as putative orthologs of 2 genes that were upregulated in biters of C. pipiens (LOC6049616, LOC6047495; Siperstein et al. 2022). No KEGG pathways had more than 1 DEG and thus, no pathways were significantly enriched.

MicroRNAs (miRNAs)

Small RNA sequencing of 4 biological replicate samples of avid-biting and 3 biological replicates of reluctant-biting head tissue produced between 12,114,191 and 30,093,158 single-end reads per biological replicate (Supplementary Table S5). As noted above, 1 biological replicate sample of reluctant-biting heads was excluded as an outlier due to very low raw sequence reads (8,087 raw reads); but this had no effect on downstream analyses and results. After quality and contaminant filtering of reads, between 25% and 67% of reads were retained. After size filtering, a total of 1,774,936–8,703,025 reads per biological replicate sample remained. Between 30% and 34% of the size-sorted reads mapped to 201 mature and precursor miRNA combinations with at least 10 reads summed across all samples (Supplementary Table S5).

Principal component analysis showed that the miRNA expression profiles of the reluctant-biting and avid-biting mosquitoes did not separate on the first 2 principal components which explained > 70% of the variation (Fig. 3). No miRNAs were significantly differentially expressed based on the criteria of an adjusted P-value < 0.05 and log₂FC > |1| (Fig. 3).

qRT-PCR of Differentially Expressed mRNA Transcripts

Six genes that were upregulated in avid-biting A. albopictus via RNAseq were also found to be upregulated via qRT-PCR, and 2 genes that were found to be upregulated in reluctant-biting A. albopictus via RNAseq were also found to be upregulated via qRT-PCR (Supplementary Table S6). Two genes that were found to be upregulated in avid-biting A. albopictus via RNAseq (5-hydroxytryptamine receptor 1-like and cytochrome P450 4C1-like; LOC109423594 and LOC109405722, respectively) were not found to be differentially expressed via qRT-PCR (Supplementary Table S6).

Discussion

Transmission of the many devastating pathogens vectored by mosquitoes cannot occur without biting. While biting and imbibing blood from a vertebrate host are necessary for reproduction in many mosquitoes, there are genera, species, populations, and individuals within populations that are capable of reproducing without a blood meal (Downes 1958, Mogi 1984, Foster 1995, Rattanarithikul et al. 2007, Wahid et al. 2007, Miyagi et al. 2012, Zhou et al. 2014). Thus, there is a growing body of research focused on determining the molecular and physiological basis of host-seeking and biting behavior (Duvall 2019, Giraldo et al. 2023). A major motivation for these studies is to identify potential molecular targets that would disrupt host-seeking and/or biting behavior, thereby suppressing disease transmission. The current study expands on this work by identifying differentially expressed transcripts in head tissue that distinguish phenotypically “avid-biting” versus phenotypically “reluctant-biting” individuals of A. albopictus from a single laboratory population raised under the same near-optimal conditions. We focused on head tissues to identify genes involved in sensory perception related to host-seeking and biting, but note that gene expression differences in other tissues such as ovaries and fat-body may also differ between reluctant and avid biters.

No Differences in MicroRNA Expression Between Avid and Reluctant Biters

We found that microRNA (miRNA) expression profiles did not differ between avid and reluctant biters (Fig. 3). A large body of literature has established a role for miRNAs in regulating the reproductive physiology of mosquitoes after blood has been imbibed (Feng et al. 2018), but only one other study has quantified differences in miRNA expression during biting but before blood is imbibed (Marzec et al. 2023). Marzec et al. (2023) found that 7 miRNAs regulating energy utilization, reproduction, and immunity differ in expression in head tissues between obligate biting C. p. pipiens and non-biting C p. molestus. One notable difference between the current study and that of Marzec et al. (2023) is that we compared phenotypic differences between individuals of a single, panmictic laboratory colony over 20 min, while Marzec et al. (2023) compared 2 subspecies that vary in ecology, life-history, and genetic background. This difference between taxonomic levels (i.e., individuals vs. subspecies) indicates that miRNA modulation of host-seeking and/or blood-feeding behavior occurs on a longer evolutionary timescale.

Differences in Messenger RNA Expression Between Avid and Reluctant Biters

In the current study, we also determined transcriptional differences at the mRNA level between “avid-biting” and “reluctant-biting” A. albopictus. The transcriptional differences we identified distinguish phenotypic variation in biting propensity on a rapid time scale of 5–20 min. One explanation for these differences is that they represent an immediate transcriptional response following host exposure. Another explanation is that these differences occurred before host exposure due to genetic or environmentally induced physiological variation among individuals (i.e., differences in body size, teneral reserves, and history of adult sugar feeding). While we cannot distinguish between changes that occurred before vs. immediately after host exposure, the transcriptional differences we identified nevertheless reveal mRNAs correlated with the initiation of blood feeding.

We identified 19 mRNA transcripts that were differentially expressed between avid-biting and reluctant-biting mosquitoes (Table 1). Consistent with the established role of chemosensory function in host-seeking and biting (Matthews et al. 2016, Duvall et al. 2019, Omondi et al. 2019, Tallon et al. 2019, Ni et al. 2022), 9 of the 19 differentially expressed transcripts distinguishing avid vs. reluctant biters have been implicated in mosquito olfactory function. It is possible that the reluctant biters failed to recognize the presence of a human blood-meal host due to an inability to upregulate the relevant olfactory genes and/or a preference for a non-human host. However, the latter interpretation is unlikely because the stock colony was maintained by blood feeding on a human host for 9 generations prior to these experiments. Furthermore, several of the genes we identified as associated with phenotypic variation in biting behavior share similar functions with, or are inferred orthologues of, genes associated with evolutionary differences between biting and non-biting behaviors in W. smithii (Bradshaw et al. 2018) and C. pipiens (Siperstein et al. 2022). Below, we discuss the inferred function of 9 of 19 genes from Table 1 with a likely role in host-seeking in the current study; all of these genes are implicated in olfactory processes.

Odorant Binding Proteins (OBPs)

We identified 3 putative odorant binding protein (OBP) transcripts that were upregulated in avid relative to reluctant-biting mosquitoes (Table 1). Specifically, avid biters exhibited increased expression of general odorant-binding protein 56a-like (Table 1c) and general odorant-binding protein 83a-like (Table 1a). Avid biters also exhibited increased expression of LOC109409797, which is annotated as an uncharacterized protein in the A. albopictus (Table 1b) but is likely an OBP based on BLASTp results indicating strong similarity to an annotated OBP in W. smithii (Supplementary Table S7). The olfactory function of OBPs has been well characterized; OBPs bind hydrophobic odor molecules in the aqueous sensillar lymph of insect olfactory tissues and transport the odor molecules to receptors present on the dendritic membranes of the olfactory sensory neurons (Leal 2013). Although OBPs have likely evolved a diverse range of functions as carriers of hydrophobic ligands in non-olfactory tissues (Sun et al. 2018), the fact that these 3 OBPs were differentially expressed in head tissue during a biting assay supports the interpretation that they have an olfactory function related to host-seeking.

Our results are also consistent with previous comparative transcriptomic studies indicating a role for OBPs in host attraction in other mosquito genera. Bradshaw et al. (2018) show that, out of 21 contigs/singletons associated with odorant reception in W. smithii, 8 were annotated as OBPs and were expressed at higher levels in avid-biting relative to reluctant-biting females. Similarly, Siperstein et al. (2022) found that 9 OBPs exhibited greater expression in biting C. p. pipiens relative to non-biting C. p. molestus. Thus, OBPs are implicated in: (i) intrapopulation phenotypic variation of biting behavior in the current study; (ii) evolutionary variation of biting behavior both between laboratory-selected lines and between naturally evolved populations of W. smithii (Bradshaw et al. 2018); (iii) evolutionary variation of biting behavior between the subspecies C. p. pipiens and C. p. molesutus (Noreuil and Fritz 2021, Siperstein et al. 2022).

Cellular Receptors and Co-Receptors Implicated in Olfaction

Avid biters exhibited upregulation of 3 genes encoding putative or established cellular co-receptor or receptor proteins (Table 1). These 3 genes are the odorant co-receptor orco (Table 1l), sensory neuron membrane protein 1 (snmp1, Table 1i), and the ionotropic receptor glutamate receptor ionotropic, kainate 2-like (Table 1n). The odorant co-receptor ORCO has a well-established, highly conserved, and unique functional role in insect olfaction; individual odorant receptor proteins (ORs) must join with ORCO to form a functional heterodimeric receptor capable of responding to odorants (Krieger et al. 2003, Leal 2013). Consistent with our finding of greater orco expression in avid biters, disruption of orco expression resulted in lower blood-feeding rates in female A. albopictus (Liu et al. 2016) and reduced attraction to human scent (in the absence of CO₂) in female A. aegypti (DeGennaro et al. 2013). SNMP1 is involved in insect pheromone detection, potentially acting as a coreceptor in conjunction with ORs and ORCO (Cassau and Krieger 2021). Evidence of female antennae-biased expression patterns in several mosquito species further support the role of SNMP1 in host-seeking behavior (Pelletier and Leal 2011, Lombardo et al. 2017, Tallon et al. 2019). Ionotropic receptors (IRs) are a divergent family of ionotropic glutamate receptors that are expressed in olfactory sensory neurons (Benton et al. 2007, Croset et al. 2010). Unlike ORs, IRs have a strong binding affinity for carboxylic acids, aldehydes, and amines (Rytz et al. 2013, Pitts et al. 2017, Ray et al. 2023), which are important components of human-skin emanations (Bernier et al. 2000). Consistent with our current results in A. albopictus, we previously found 3 ionotropic receptors that were more abundantly expressed in biting C. p. pipiens relative to non-biting C. p. molestus (Siperstein et al. 2022).

Putative Odorant-Degrading Enzymes

Avid biters also exhibited greater expression of 2 genes annotated as cytochrome p450 4C1-like (Table 1m, s), though the former gene was not validated by qRT-PCR. Another gene, annotated as 2-hydroxyacylsphingosine 1-beta-galactosyltransferase-like (Table 1f), was also upregulated by avid biters and encodes a protein belonging to a conserved domain family of UDP-glucoronosyl and UDP-glucosyl transferases. All 3 of these genes have annotations that identify them as possible odorant-degrading enzymes (ODEs), a group of biotransformation enzymes (including cytochrome p450s and UDP glucosyltransferases) that are expressed in insect antennae (Wang et al. 1999, Steiner et al. 2019). Relatively few studies have investigated the expression and function of ODEs in insects, but one proposed activity is that ODEs inactivate odorants in the sensillar lymph, which would facilitate odorant turnover, prevent overstimulation of ORs, and preserve olfactory sensitivity (Steiner et al. 2019). Intriguingly, 2 of the genes noted above (Table 1f, s) are inferred orthologues (i.e., reciprocal best BLASTp hits) of 2 genes identified as upregulated in biting C. pipiens pipiens relative to non-biting C. pipiens molestus (Sipperstein et al. 2022: LOC6049616, LOC6047495).

Summary Conclusions

Previous studies have investigated the roles of age-, adult diet-, and bloodmeal-induced plasticity in mosquito host-seeking and biting behaviors that are associated with shifts in the expression of olfactory genes (Matthews et al. 2016, Omondi et al. 2019, Tallon et al. 2019, Ni et al. 2022). Our study similarly finds clear differences in the expression levels of olfactory-related genes associated with these behaviors. Additionally, whereas previous studies manipulated the physiological state of experimental mosquitoes to induce differences in biting behavior, in our study design, mosquitoes were reared under the same near-optimal conditions during both the larval and adult stages to reduce environmental variation affecting host-seeking and biting behavior. By reducing (but not eliminating) environmental variation, we maximized the opportunity to identify genetically based transcriptional variation between avid and reluctant biters. It is notable that we identified functional overlap between the olfactory-related DEGs in this study and evolved differences in gene expression between biting and non-biting populations of W. smithii (Bradshaw et al. 2018) and subspecies of C. pipiens (Siperstein et al. 2022). Future studies should continue to explore whether these genes are involved in the evolutionary transition to a non-biting life history in order to improve our understanding of the molecular physiology of biting behavior and, ultimately, to implement new avenues for vector control.

Acknowledgments

We thank members of the Armbruster lab and 2 anonymous reviewers for helpful suggestions on previous versions of this manuscript.

Author contributions

Mara Heilig (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Software [equal], Writing—original draft [equal], Writing—review & editing [equal]), Samantha Sturiale (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Software [equal], Validation [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Sarah Marzec (Formal analysis [equal], Software [equal], Validation [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Christina Holzapfel (Conceptualization [equal], Funding acquisition [equal], Methodology [equal], Project administration [equal], Supervision [equal], Writing—original draft [equal], Writing—review & editing [equal]), William Bradshaw (Conceptualization [equal], Funding acquisition [equal], Project administration [equal], Supervision [equal], Writing—original draft [equal], Writing—review & editing [equal]), Megan Meuti (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal], Supervision [equal], Writing—original draft [equal], Writing—review & editing [equal]), and Peter Armbruster (Conceptualization [equal], Data curation [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal], Resources [equal], Supervision [equal], Validation [equal], Writing—original draft [equal], Writing—review & editing [equal])

Funding

This research was funded by the National Institutes of Health (R21-AI144266), funds from the Davis Family Endowment, and the National Science Foundation (IOS-1944324).

References

Author notes