Mosquitoes, infections, and tissue dissections

Mosquitoes (Ae. albopictus Roma-RM strain mosquitoes, collected and reared in Rome, Italy) were reared under standard laboratory conditions (25 ± 1 °C, relative humidity 60 ± 10%, light: dark photoperiod 14:10 h) in the insectary of the Department of Infectious Diseases at the National Institute of Health, Rome (Istituto Superiore di Sanità, ISS). In vivo experimental infections of Ae. albopictus mosquitoes with CHIKV were performed in a biosafety level 3 laboratory (BSL3) at the ISS, obtaining three independent biological replicates. In each experiment, around 120–150 females 5–7 dpe (days post emergence) were collected to allow infected blood feeding (rabbit blood with viral particles of the CHIKV strain isolated from a human biological sample collected during the Italian 2007 CHIKV outbreak [49]; titer: 2.98x10⁷/ml) and control blood feeding (rabbit blood with MEM containing inactivated fetal serum, amino acids, penicillin, streptomycin). After blood feeding (1 hour lasting), mosquitoes were chilled on ice and engorged females were isolated. Mosquitoes were then dissected at two different timepoints (T), 1 day (T1) and 5 days (T5) post infected blood meal (dpi): for each replicate, around 20 midguts and 6–8 carcasses (whole body without the midgut) were collected for each T as well as 6–8 whole females (details are reported in the S1 Text and Table A in S1 Text). The samples were next collected in RNA later and then stored at -80°C until further use.

RNA extraction, library preparation and sequencing

Total RNA was extracted from midguts, carcasses, and whole bodies of infected and not-infected female mosquitoes at different T (T1 and T5), quantified by spectrophotometric reading (Take3 module of the Microplate Reader BioTek SynergyHT) and evaluated by agarose gel electrophoresis. Only samples showing appropriate quality and quantity parameters were selected for further treatments, i.e., DNase I treatment, library preparation and RNA-seq. Before proceeding to NGS (Next Generation Sequencing), CHIKV titer was also determined in representative samples by Real Time quantitative PCR (RT-qPCR). To evaluate CHIKV titer, 3–5 entire females and 8–10 carcasses (females after midguts dissections) were collected at day 5 and analysed by RT-qPCR (see below for technical details and Supplementary S1 Text): all the samples analysed confirmed the occurrence of viral infection in the mosquitoes. RNA samples were initially quantified using the Qubit 3.0 Fluorometer to ensure that the quantity of material submitted was adequate. Total RNA integrity was checked using the Fragment Analyzer (Bioanalyser) to measure the samples quality as RNA Quality Number (RQN). RNA-seq was performed at Polo d’Innovazione di Genomica Genetica e Biologia SCaRL, Siena – Italy, thanks to a grant awarded by the European Project Infravec2 (http://infravec.mdmdemo.ch). The libraries were prepared in accordance with the Illumina TruSeq Stranded mRNA Sample Preparation Guide for Illumina Paired-End Indexed Sequencing and then validated using the Fragment Analyzer to check the distribution. Finally, concentration of library samples was defined based on the Qubit 3.0 Fluorometer quantification. Indexed DNA libraries were normalized to 2 nM and then pooled in equal volumes. The pool was loaded at a concentration of 1.2 pM onto an Illumina NextSeq 550 Flowcell High Output, with 1% of Phix control. The samples were then sequenced using the Illumina chemistry V2, 2x75 bp paired end run.

Expression profiling and identification of differentially expressed transcripts

Raw sequencing reads were subjected to quality filtering using Trimmomatic-0.32 to eliminate low-quality sequences and adaptor contaminants [50]. The resulting high-quality read pairs were aligned to the Ae. albopictus reference transcriptome (Genome version: AalbF5, assembly: GCF_035046485.1, NCBI) using Bowtie [51]. Expression levels were quantified as Fragments Per Kilobase of transcript per Million fragments mapped (FPKM) using RSEM [52] (Supplementary S1 Data). Differential expression analysis was performed using edgeR and pairwise comparisons between the conditions [53]. Transcripts with a False Discovery Rate (FDR) less than 0.05 and a Fold Change (FC) greater than 2 were considered statistically significant and selected for subsequent analyses. DE contigs were identified for the following comparisons (Supplementary S2 Data): infected midguts (Mi-V) versus control midguts (Mi-C) at T1; both infected midguts (Mi-V) versus control midguts (Mi-C), and infected carcasses (Ca-V) versus control carcasses (Ca-C) at T5.

Pfam and GO enrichment

To identify enriched Gene Ontology (GO) and Protein Families (Pfam) terms, enrichment analyses were conducted for the following pairwise comparisons: at T1, infected midguts (Mi-V) versus control midguts (Mi-C); and at T5, both infected midguts (Mi-V) versus control midguts (Mi-C), and infected carcasses (Ca-V) versus control carcasses (Ca-C) (Supplementary S3 Data and S4 Data). The sma3s and pfamscan software [54–56] were utilized to assign, respectively, GO terms and Pfam term to transcripts of the AalbF5 reference transcriptome and to transcripts identified as differentially expressed by edgeR analysis in pairwise T1 and T5 comparisons. Subsequently, Fisher’s Exact Test was applied to assess the overrepresentation of specific GO terms and Pfam domains among the differentially expressed transcripts, with p-values adjusted via the Benjamini-Hochberg procedure to control for multiple testing (significance threshold: adjusted p-value < 0.05) [57]. All statistical analyses were conducted within the R programming environment [58].

Validation of candidate genes expression by quantitative real-time PCR

To validate the transcript abundance patterns identified through RNA-seq analysis, RT-qPCR was conducted on five selected candidates. Tissues and organs were dissected from uninfected adult females (4–7 days post-emergence, dpe), which were reared under standard laboratory conditions in the Insectary of the Department of Public Health and Infectious Diseases at Sapienza University and maintained on a 10% sucrose diet. Mosquitoes were initially anesthetized on ice for 2–3 minutes, and dissections were carried out to obtain three independent biological replicates for each of the following tissues/organs: hemolymph (collected using the proboscis clipping technique and containing only circulating hemocytes), ovaries, midguts, carcasses without hemolymph, carcasses without the midgut, carcasses without the ovaries and head, and whole females. The samples were then preserved in PBS solution at -80°C until RNA extraction.

Total RNA was extracted from the collected samples using Trizol reagent (Invitrogen), following the manufacturer’s instructions. The RNA quantity and quality were assessed through spectrophotometric measurements (using the Take3 module of the BioTek SynergyHT plate reader with GEN5 software) and 1% agarose gel electrophoresis. To eliminate genomic DNA contamination, DNase I treatment was performed using the Ambion DNA-free kit, following the manufacturer’s protocol. The efficiency of DNase I treatment was verified by endpoint PCR, comparing untreated and treated RNA samples using the ribosomal S5 (rpS5) gene as a target. Different amounts of DNase I-treated RNA (ranging from 200 ng to 1 µg) were used as templates to synthesize First-Strand cDNA with Superscript II RT (Invitrogen) and Oligo dT (Invitrogen), following the manufacturer’s guidelines. The resulting cDNA samples were diluted to a concentration of 10 ng/µL and used as templates for quantitative real-time PCR. Standard curves were generated for each target gene as well as for the endogenous reference gene (rpS5). Specifically, the cDNA samples used for standard curve preparation were derived from a pool of four whole female mosquitoes, following the previously described procedures. Serial dilutions of cDNA were prepared to establish five points for each standard curve (dilution factor: 1:5) at the following concentrations: [100 ng/µL], [20 ng/µL], [4 ng/µL], [0.8 ng/µL], and [0.16 ng/µL]. Each PCR reaction began with an initial holding stage of 2 minutes at 50 °C, followed by 2 minutes at 95 °C. This was followed by 40 cycles of amplification (95 °C for 15 seconds; 60 °C for 1 minute). A melt curve analysis was included to evaluate primer efficiency and confirm the specificity of the amplicon for each target gene. Relative quantification analysis was performed using both the standard curve method and the ΔΔCt method, where applicable. In the standard curve approach, the relative expression levels of selected genes across different tissues were determined by calculating the ratio of the target gene to the endogenous reference gene (rpS5), based on their Ct values interpolated from the standard curves. Statistical analysis was conducted using one-way ANOVA, followed by Tukey’s multiple comparisons test, applied to log10-transformed values from the three biological replicates. Statistical significance was denoted as follows: *p < 0.05; **p < 0.01; ***p < 0.001. Primers’ sequences are reported in Supplementary S1 Text (Table D in S1 Text).

Experimental outline

The main aim of this study was to better understand the Ae. albopictus responses to CHIKV infection, focusing specifically on two key stages of the viral lifecycle within the mosquito: the early phase, when the virus enters the mosquito midgut after an infectious blood meal, and the later stage, when the viral particles cross the midgut and spread into the mosquito haemolymph. In fact, shortly after ingestion with the blood, at 1 day post-infection (dpi), the virus is initially targeted by the “midgut infection barrier,” an immune response primarily driven by midgut epithelial cells and/or midgut-associated hemolymph cells [13,59]. During this early phase, several additional factors may influence the outcome of the infection, including abundance and composition of the mosquito microbiota, production of the peritrophic matrix and effectiveness of blood digestion. Viral particles eluding this first line of defense may replicate within midgut cells, cross the “midgut escape barrier,” and disseminate into the mosquito hemocoel [60]. This triggers a second wave of immune responses, which take place in the few days following midgut escape and primarily involves cellular components of the hemolymph (such as hemocytes, fat body cells, and other immune cells) working together to counteract viral replication, diffusion and invasion of salivary glands [14]. The balance between the mosquito immune responses and viral replication and dissemination is central in determining the mosquito competence for transmitting the virus to the next host. Consequently, our experimental design employed an RNA-seq approach to analyze the transcriptional responses of infected and uninfected mosquitoes at 1 and 5 dpi, with a special focus on midgut, as representative of the early responses to infection and on carcasses (whole body without midguts, i.e., containing hemolymph) as representative of later responses following viral dissemination (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic representation of the experimental design and biological samples used for RNA-seq.

The diagram illustrates key mosquito tissues sampled for RNA sequencing. Depicted within the mosquito body are: the gut (midgut shown in dark red at T1 and light red at T5), the crop (dark grey diverticulum connected to the foregut), the salivary glands (light blue structures in the thorax), hemocytes and fat body cells (represented by black dots in the grey hemolymph), and virions (blue hexagons). The term carcass refers to the entire body after the midgut has been removed.

https://doi.org/10.1371/journal.pntd.0013588.g001

Sequencing, transcriptome assembly and annotation, and differential expression (DE) analysis

Total RNA was extracted from the samples summarized in Fig 1 and used for library preparation and sequencing. A total of 1,170,016,748 Illumina raw reads was generated (Table 1) across 24 libraries (with two or three biological replicates for each time point). Raw RNA-seq sequences were deposited in the NCBI Sequence Read Archive (SRA) database with accession number PRJNA1260742. After trimming adapters and filtering out low-quality reads, cleaned reads were mapped against the reference AalbF5 transcriptome to calculate row counts and expression values for each of the 38,020 transcripts, using the FPKM metrics (Supplementary S1 Data).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of samples and RNA-Seq Data.

https://doi.org/10.1371/journal.pntd.0013588.t001

To identify differentially expressed (DE) contigs, datasets from infected and uninfected control samples—specifically, midguts at T1 and carcasses at T5—were compared using edgeR tool (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of Differential Expression (DE) Analysis. The table reports the number of upregulated (UP) and downregulated (DOWN) genes at three significance thresholds (FDR < 0.05, FDR < 0.01, FDR < 0.001). Comparisons shown are: Mi T1 V vs C (midguts at 1 dpi, CHIKV-infected vs control) and Ca T5 V vs C (carcasses at 5 dpi, CHIKV-infected vs control).

https://doi.org/10.1371/journal.pntd.0013588.t002

Differential expression analysis revealed that at day 1 post-infection (T1) 73 contigs were upregulated in CHIKV-infected midguts, while 72 were downregulated in CHIKV-infected midguts compared to the control group with FDR ≤ 0.05 (Fig 2A). Similarly, at day 5 post-infection (T5), 41 contigs were more abundant in CHIKV-infected carcasses and 85 were downregulated in CHIKV-infected carcasses compared to the control group (Fig 2B). The whole edgeR outputs, including the comparison between infected and control midgut at 5 dpi, are reported in Supplementary S2 Data.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Volcano plots displaying the results of the differential expression (DE) analysis.

(A) DE contigs in midguts at T1. (B) DE contigs in carcasses at T5. The x-axis represents Log10 fold change (FC), while the y-axis represents -Log10 false discovery rate (FDR). DE genes are highlighted in red. Transcripts differentially upregulated in the infected groups compared to control groups are shown on the left side of each graph, whereas those downregulated in the infected groups compared to control groups are displayed on the right side.

https://doi.org/10.1371/journal.pntd.0013588.g002

A minimal overlap was observed when comparing the different DE groups summarized in Table 2 (see Fig B and Table C in S1 Text). This low degree of overlap suggests that distinct, organ-specific responses are activated within the mosquito compartments (midguts and carcasses) during the different stages of the virus lifecycle: at 1 day post-infection, when the virus enters the enterocytes, and at 5 days post-infection, when the virus disseminates into the mosquito hemocoel. Such limited overlap in transcriptionally modulated Ae. albopictus genes upon CHIKV infection across different organs (midguts vs. carcasses) and time points (1 and 5 days post-infection) was also observed in previous studies by Vedururu and colleagues [46,47], and by Modahl and colleagues [48].

We also performed GO and Pfam enrichment analyses by comparing GO terms (encompassing the three main categories: biological process, BP; cellular component, CC; molecular function, MF) and Pfam terms for each sample against the general transcriptome. Differentially enriched GO and Pfam terms in infected and control midguts at day 1 (T1_Mi_C-UP and T1_Mi_V-UP), and infected and control carcasses at day 5 (T5_Ca_C-UP and T5_Ca_V-UP) are shown in the following figures and described in more detail below for each timepoint. Analysis of infected and control midguts at day 5 post-infection (T5_Mi_C-UP and T5_Mi_V-UP) are reported in Supplementary S1 Text (Table B and Fig A in S1 Text). To complement the information provided by GO and Pfam enrichment analyses, selected DE contigs for each group have also been included in the figures (for complete DE, GO and Pfam lists, refer to the Supplementary S2 Data, S3 Data and S4 Data and Supplementary S1 Text).

Gene expression profiles in midguts at day 1 post infection

As already mentioned, the midgut epithelium is the first physical barrier the virus encounters during its lifecycle within the mosquito. Various factors contribute to the complex interactions between the virus and the midgut lumen environment, including the microbiota, the formation of the peritrophic matrix, and the secretion of proteolytic enzymes responsible for blood meal digestion. Moreover, virus invasion and replication within midgut cells evokes specific defense responses. Among these responses, the RNAi pathway is known to be activated by CHIKV during the early stages of its lifecycle in the mosquito midgut.

Accordingly, several differentially expressed (DE) transcripts and Pfam domains linked to RNA interference (RNAi)-mediated defense mechanisms were significantly more abundant in infected midguts. These included RNA helicases and subunits of the PAN2-PAN3 deadenylase complexes [61–63]. Additionally, other upregulated GO terms, Pfam categories, and DE contigs in virus-infected midguts (Fig 3) were related to ubiquitination processes, such as E3 ubiquitin ligases and ubiquitin hydrolases [64]. These ubiquitin-related factors may represent an initial line of defense against viral replication in enterocytes or they may be part of a viral strategy to elude mosquito immune responses [65–67]. Also, laminin, a component of the basal lamina that interacts with invading pathogens [68], was significantly upregulated in infected midguts. Among classical immune responses, members of the scavenger receptor family, leucine-rich receptors, and STAT factors were activated by viral infection in the midgut. Finally, cadherins, which act as binding proteins for DENV in mosquito cells, may serve as critical receptors during viral infection in mosquito cells [69].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Selection of enriched terms in infected midguts at 1 dpi, i.e., upregulated in response to CHIKV infection.

(A) List of enriched Gene Ontology (GO) terms according to the three categories, Biological process (BP, red bars), Molecular Function (MF, blue bars) and Cellular Component (CC, green bars); (B) list of enriched Pfam domains, families and repeats: Pfam IDs and descriptions are reported as the adjusted p value (adjp); (C) list of differentially expressed (DE) contigs: transcript IDs, according to AalbF5 genome assembly, are reported with relative annotation; enrichment parameters are logFC (Fold Change) logCPM (Counts Per Million) and FDR, False Discovery Rate.

https://doi.org/10.1371/journal.pntd.0013588.g003

Several GO terms downregulated by the virus (i.e., upregulated in control midguts, Fig 4) were associated with lipid and lipoprotein metabolism, a pathway already known to be affected by arboviral infection of mosquitoes [70]. Other digestive processes also appeared to be modulated or suppressed by CHIKV, including triglyceride lipase activity and trypsin/carboxypeptidase activity. Among contigs overexpressed in control midguts were many peptidases, particularly carboxypeptidases, suggesting that blood meal digestion in infected midguts may be impaired to support viral survival and maintenance. The presence of DE contigs (Fig 4C) encoding lipase and peritrophin aligns well with findings from GO and Pfam enrichment analyses. One of the mosquito defense mechanisms against viral infections may involve targeting factors and pathways required for viral invasion of midgut cells through endosome formation and maturation [71]. Accordingly, an enrichment of GO and Pfam terms related to phagocytosis, recognition, vacuolar acidification, and endosomal vesicle fusion was observed in uninfected midguts (Fig 4A and 4B). Pathogen entry via clathrin-mediated endocytosis, which relies on receptors associated with clathrin, is another critical pathway [72]. This process requires specific chemical and physiological conditions, such as a pH below 6 and a particular lipid composition. Cholesterol and a pH below 6–6.5 are essential for the fusion of the alphavirus envelope with endosomes. Additional GO and Pfam terms more abundant in control midguts compared to infected midguts are related to the peritrophic matrix, chitin binding, and chitin metabolic processes. A specific repression of E3 ubiquitin ligases occurs in infected midguts, indicating that ubiquitination and proteasome formation may be altered by both the virus and the mosquito immune responses, as mentioned above. Recent studies have shown that the downregulation of genes involved in N-glycosylation, such as Alg9 and Alg3, is associated with the activation of the JNK pathway as a stress response [73]. Significant transcriptional changes in genes involved in the mitochondrial respiratory chain were also observed: while DENV and ZIKV infections were previously reported to affect mitochondrial metabolism, the transcriptional modulation of mitochondria-linked targets in CHIKV infected cells is less understood [74,75]. Furthermore, viral infection also appeared to affect factors involved in epigenetic modifications, which is not surprising since DNA/RNA methylation and histone acetylation (e.g., histone H2-H3 acetylation) have been previously associated to regulation of insect immune responses to pathogens [76,77].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Selection of enriched terms in control midguts at 1 dpi, i.e., downregulated in response to CHIKV infection.

(A) List of enriched Gene Ontology (GO) terms according to the three categories, Biological process (BP, red bars), Molecular Function (MF, blue bars) and Cellular Component (CC, green bars); (B) list of enriched Pfam domains, families and repeats: Pfam IDs and descriptions are reported as the adjusted p value (adjp); (C) list of differentially expressed (DE) contigs: transcript IDs, according to AalbF5 genome assembly, are reported with relative annotation; enrichment parameters are logFC (Fold Change) logCPM (Counts Per Million) and FDR, False Discovery Rate. The asterisk indicates XM_019688536.3, whose expression is analysed through RT-qPCR validation (Fig 9).

https://doi.org/10.1371/journal.pntd.0013588.g004

Gene expression profiles in carcasses at day 5 post infection

After midgut invasion, between 2 and 6 days post-infection, viral particles escape from the midgut to disseminate into the hemocoel and a general systemic transcriptional modulation of mosquito genes occurs in several mosquito organs [59,78]. To specifically focus on the systemic immune responses of mosquito organs, such as the fat body, and immune cells such as hemocytes, we dissected and removed the midguts from mosquitoes 5 days after an infectious or non-infectious (control) blood meal (Fig 1). RNA-seq datasets from infected and uninfected midguts and carcasses (whole females after midgut removal) were analyzed to evaluate CHIKV-induced transcriptional modulation.

Mosquito carcasses exhibited upregulated GO terms associated with peroxisomal organization and metabolism [79,80]. Key immune-related biological processes, such as calcium homeostasis, positive regulation of antimicrobial peptides, response to exogenous dsRNA, hemocyte proliferation and immune system development were also enriched in infected carcasses. Among enriched GO or Pfam terms and differentially expressed (DE) transcripts were also several processes related to lipid metabolism and fatty acid homeostasis (Fig 5). Also, the ubiquitin carboxyl-terminal hydrolase 20, a deubiquitinating enzyme involved in autophagy and cellular antiviral responses was strongly upregulated in infected carcasses compared to controls, and immunoglobulin domains —key components of immune-reactive molecules in invertebrates— were significantly upregulated upon pathogen challenge [81]. The 37 kDa salivary gland allergen Aed a 2 is highly similar (84% identity) to the juvenile hormone-binding protein (mJHBP) found in Ae. aegypti, a mosquito-specific protein that binds juvenile hormone III and is abundant in pupae and adults [82]. Disrupting the mJHBP gene using CRISPR-Cas9 led to impaired immune responses, including delayed antimicrobial peptide production, reduced immune gene expression, defective phagocytosis, and higher vulnerability to bacterial infections like Serratia marcescens. Additionally, mutant mosquitoes showed altered hemocyte populations and decreased hemocyte phagocytic activity [83]. Members of ABC transporter family, which are involved in translocation of various molecules across biological membranes, were highly enriched in CHIKV-infected Ae. albopictus carcasses suggesting their involvement in antiviral immunity. Among DEG, Talin, an integrin-like transmembrane receptor, may mediate immune responses in insect hemocyte-like cells [84,85]. Finally, a leucine-rich repeat-containing protein and a toll-like receptor, both upregulated in infected carcasses, are potentially involved in the defense pathways described above.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Selection of enriched terms in infected carcasses at 5 dpi, indicating upregulation in response to CHIKV infection.

(A) List of enriched Gene Ontology (GO) terms according to the three categories, Biological process (BP, red bars), Molecular Function (MF, blue bars) and Cellular Component (CC, green bars); (B) list of enriched Pfam domains, families and repeats: Pfam IDs and descriptions are reported as the adjusted p value (adjp); (C) list of differentially expressed (DE) contigs: transcript IDs, according to AalbF5 genome assembly, are reported with relative annotation; enrichment parameters are logFC (Fold Change) logCPM (Counts Per Million) and FDR, False Discovery Rate.

https://doi.org/10.1371/journal.pntd.0013588.g005

Among GO terms significantly downregulated during viral dissemination were factors involved in the ubiquitin pathway as the GID complex, which is associated with E3 ligase assembly and may play a role in the ubiquitination process [64]. Moreover, several protein families and domains, including RING domain-containing proteins, E3 ubiquitin protein ligases, and autophagy-related domains, were also downregulated by the presence of the virus (Fig 6). Certain Major Facilitator Superfamily (MFS) proteins play critical roles in immune processes, including viral invasion and pathogen resistance [86]. Additionally, a mucin protein in Ae. aegypti has been shown to interact with DENV, influencing viral infection dynamics [87]. Among other terms differentially expressed in carcasses at day 5, an LDL receptor was found downregulated in infected carcasses: notably, a reduction in LDL levels is linked to severe viral infections in Ae. aegypti mosquitoes infected with DENV [88]. Overall, the involvement of fatty acid and lipid metabolism in viral infection development has been previously discussed and plays a key role in the viral dissemination in the hemocoel.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Selection of enriched terms in control carcasses at 5 dpi, indicating downregulation in response to CHIKV infection.

(A) List of enriched Gene Ontology (GO) terms according to the three categories, Biological process (BP, red bars), Molecular Function (MF, blue bars) and Cellular Component (CC, green bars); (B) list of enriched Pfam domains, families and repeats: Pfam IDs and descriptions are reported as the adjusted p value (adjp); (C) list of differentially expressed (DE) contigs: transcript IDs, according to AalbF5 genome assembly, are reported with relative annotation; enrichment parameters are logFC (Fold Change) logCPM (Counts Per Million) and FDR, False Discovery Rate.

https://doi.org/10.1371/journal.pntd.0013588.g006

Distinct temporal and spatial immune responses: local and systemic alternatives

To get a wider overview on the modulation of immune-related pathways and gene families at local and systemic level following CHIKV infection, we first compiled a comprehensive list of Ae. albopictus immune-related gene families and then analyzed the transcriptional patterns governing mosquito immune responses to arboviral infection across different time points (T1 and T5) and body compartments (i.e., at 1 dpi in the midgut and at 5 dpi in the carcasses/hemolymph). To this end, immune-related genes from Aedes mosquitoes identified and/or catalogued in previous studies [48,70,89,90] were used to identify Ae. albopictus orthologs by tblastn searches and then further classified/refined using available annotations [91]. In addition, putative novel immune-related contigs were identified by searching specific annotation terms in our transcriptome. This way we obtained a final list of Ae. albopictus immune-related genes that included transcript isoforms with FPKM values ≥ 1 in at least one sample. The presence of multiple isoforms may result from differences in transcript length at the 5′ and/or 3′ UTRs, alternative splicing events, or occasional errors in transcript annotation. This becomes particularly relevant when truncated or spliced variants encode distinct domains with potential functional implications. In cases where discrepancies are observed, gene-specific analyses are required before drawing conclusions. Nonetheless, for most of the genes analyzed, both the expression profiles and the absolute FPKM values of the different isoforms exhibited a high degree of consistency. This catalogue is composed of 891 contigs grouped into the following immune-related families: ubiquitination, RNA interference, IMD, JAK-STAT, TOLL pathways, PGRP, scavenger receptors, LRR receptors, AMPs, PPO, CTL, CLIP, TEP, serpin, autophagy, and apoptosis (Supplementary S5 Data).

Variations of transcript abundance were evaluated comparing FPKM values of infected samples (V) versus control samples (C) and taking into consideration only transcripts with log2-transformed V/C ratio ≥ 1 or ≤ -1, that is at least twice as abundant in infected versus control samples, or viceversa (Supplementary S5 Data and Supplementary S1 Text). The mirrored bar graph in Fig 7 illustrates the percentage of contigs, within each immune group, with at least 2x FPKM values in infected samples compared to control samples in the midgut at T1 and in the carcass/hemolymph at T5. These variations in transcript abundance provide an overview of local and systemic Ae. albopictus immune defense responses to CHIKV infection offering the opportunity to focus on gene family and pathway levels rather than on individual genes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Transcriptional modulation of immune gene family members following viral challenge.

The mirrored bar chart displays on the x-axis the percentage of transcripts that were at least twice as abundant in infected versus control samples from each immune group as listed on the y-axis (with the total number of family members in brackets). Red bars refer to data from midguts at 1 dpi, while grey bars represent data from carcasses at 5 dpi.

https://doi.org/10.1371/journal.pntd.0013588.g007

Several immune gene families and pathways exhibited an increased transcriptional activity in the infected midgut at 1 dpi. RNA interference (RNAi) is recognized as a key antiviral mechanism in invertebrates and known to be involved in the response of Aedes mosquitoes to viruses such as DENV, ZIKV, and CHIKV [48,92] although significant upregulation of RNAi transcripts during the early stages of midgut infection was not consistently reported [46,78,93]. We observed upregulation of the RNAi pathway at 1 dpi in infected midguts, where it likely contributes to restricting early viral replication [94], and also found a limited, yet detectable, activation in infected mosquito carcasses at 5 dpi (Figs 7 and 8B). The evolutionarily conserved Toll, Imd, and JAK-STAT pathways appeared to play a role in limiting arbovirus replication, with the transcriptional activation of Imd and JAK-STAT components being especially evident in midguts at 1 day post-infection (dpi), whereas transcriptional upregulation of Toll family members was more pronounced in carcasses at 5 dpi (Figs 7 and 8C). It is important to note that within each family, both activators and repressors may be either up- or downregulated. Thus, their transcriptional modulation in response to viral challenge provides only a broad overview of the host response. A more comprehensive understanding of the underlying biological processes and functional implications will require further investigation at the level of individual genes and specific pathways.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Heatmaps illustrating the transcriptional profiles of members of the (A) Autophagy, (B) RNAi and (C) JAK/STAT families.

Each heatmap includes: the identifier of each contig (including possible isoforms) in the TRANCRIPT_ID column; the log2 values of the ratio between the FPKM values (+1 FPKM) of infected samples (V) and control samples (C), presented for midgut samples at 1 dpi (T1-Mi V/C) and carcass samples at 5 dpi (T5-Ca V/C), with a corresponding color legend below; the gene description or annotation provided in the GENE_NAME column.

https://doi.org/10.1371/journal.pntd.0013588.g008

Members of gene families involved in molecular and cellular homeostasis and turnover exhibited higher transcriptional abundance in infected versus uninfected midguts at T1 and in comparison, to carcasses at T5 (Fig 7). Among them, the upregulation of ubiquitination-related components resulted evident. This observation appears particularly relevant considering that post-translational protein modification regulates numerous biological processes [66], including immune responses of insects infected by various pathogens [67,95,96]. Additionally, growing evidence suggests that microbial pathogens may exploit the ubiquitin pathway to evade the host immune system [95]. Several genes appeared instead downregulated in the midgut at T1 in response to viral infection. Remarkably, significant downregulation of transcripts encoding digestive enzymes was observed in the early stages of CHIKV replication, when the virus crosses the epithelial barrier in the mosquito midgut. Searching our transcriptome using the keywords “trypsin”, “peptidase”, “carboxypeptidase”, “metalloprotease”, and setting as cutoff for inclusion an expression level ≥ 1 FPKM in at least one sample, we selected a total of 117 contigs. Among these, 79% were more abundant in control than in infected midguts, an unbalanced pattern that was not observed in carcasses at 5 dpi (Fig 9A and 9B). This downregulation of trypsins and other digestive enzymes may be induced by the virus and favor the infection as previously shown in DENV-infected Ae. aegypti [97]. Members of the AMPs, FREP and LRR families were also notably downregulated in midguts at 1 dpi, whereas they were clearly upregulated in carcasses at 5 dpi as discussed in more detail below (Fig C in S1 Text, Fig 10A and 10B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Transcriptional Modulation of Digestive Factors Upon CHIKV Infection.

(A) Heatmap displaying transcriptional changes. (B) Pie chart illustrating the percentage of genes with a control/infected ratio >1 (blue) or <1 (red). (C) Validation of the trypsin gene XM_019688536.3 by RT-qPCR across different tissues, including whole females, dissected heads, ovaries, midguts, and carcasses (whole body excluding head, ovaries, and midgut). Statistical analysis was conducted using ordinary one-way ANOVA followed by Tukey’s multiple comparison test. (D) Summary of differentially expressed (DE) contigs within the digestive factor list.

https://doi.org/10.1371/journal.pntd.0013588.g009

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Heatmaps of Transcriptional Profiles Across Immune Categories.

(A) AMPs (Antimicrobial Peptides); (B) LRR (Leucine-Rich Repeat proteins); (C) CTL-PAFs (C-type Lectins and Pathogen-Associated Factors); (D) TEP (Thioester-Containing Proteins); (E) PPO (Prophenoloxidase). Validation by RT-qPCR (contigs marked with asterisk): (F) Holotricin gene (XM_029879658.2); (G) LRR gene (XM_019686715.3); H. PAF2 gene (XM_029857423.2). RT-qPCR was performed across different tissues, including whole females, dissected heads, ovaries, midguts, and carcasses (whole body excluding head, ovaries, and midgut). Statistical analysis was conducted using ordinary one-way ANOVA followed by Tukey’s multiple comparison test.

https://doi.org/10.1371/journal.pntd.0013588.g010

As previously discussed, comparison of infected and uninfected transcriptional profiles at 5 dpi, corresponding to the viral diffusion stage in the mosquito hemolymph, revealed a limited but specific list of both immune and non-immune DEGs (Figs 5 and 6). The expression analysis of immune gene families’ members in carcasses at 5 dpi revealed a specialized systemic response, allowing for the identification of specific trends with transcriptional activation of immune-related gene families in infected carcasses as compared to uninfected controls (Figs 7 and 10). Mosquito hemocytes, the only immune cells capable of mounting both humoral and cellular immune responses, play a central role in defense against bacteria, fungi, and parasites. However, their involvement in antiviral responses is not yet fully understood [98]. The classic antiviral TOLL immune pathway appears to be activated in the hemolymph but to a lesser extent than in the early midgut stage (Fig 7). Notably, several pattern recognition receptors (PRRs) were upregulated in the hemolymph; more specifically, members of the FREP, LRR, CTL and TEP families were transcriptionally activated in infected carcasses at 5 dpi both in comparison to uninfected carcasses and in comparison to infected midguts at 1 dpi. Thioester-containing protein (TEP) family members play a key role in antibacterial and anti-plasmodium defense and lectins, which can recognize carbohydrate components of viral envelopes, promote opsonization activating immune mechanisms such as the PPOs (Phenoloxidases) cascade and phagocytosis. Melanization is indeed another hemolymph-specific immune process and, accordingly, we found that several members of the PPOs, CLIP proteases and PAF (Phenoloxidase-activating factors) families were more abundant in infected carcasses than in uninfected controls. Antimicrobial peptides (AMPs) were also found abundant in the hemolymph and specifically activated at this time point. Also, in addition to classical AMPs known to be activated by viral infection, we observed differentially expressed (DE) genes encoding glycine-rich peptides and a putative holotricin, which may represent a novel AMP involved in the immune response to invading pathogens [99–101]. To verify the tissue-specific expression profiles of a few selected candidate genes (marked with asterisks in the heat maps of Figs 9 and 10), we dissected Ae. albopictus female mosquitoes and isolated hemocytes using the proboscis clipping technique. Notably, two of the analyzed candidates—a member of the Leucine-Rich Repeat (LRR) protein family and a Pro-phenoloxidase Activating Factor 2 (PAF2)—were found to be specifically expressed in circulating hemocytes. Additionally, the putative AMP holotricin displayed a highly enriched expression profile in the hemocytes (Fig 10F–10H).