Virome dynamics of Aedes albopictus during the 2025 chikungunya outbreak in Foshan

Highlights

• •Field surveillance and genomic analysis characterized CHIKV circulation in Aedes albopictus during the 2025 Foshan outbreak.
• •CHIKV infection was associated with reduced viral diversity and restructuring of the mosquito virome.
• •CHIKV abundance showed distinct associations with co-occurring insect-associated viruses, suggesting potential virome-CHIKV interactions.

Abstract

Chikungunya virus (CHIKV) has re-emerged globally, yet little is known about how natural CHIKV transmission influences the resident virome of mosquito vectors under field conditions. Here, we combined continuous entomological surveillance, viral genomics and metagenomic sequencing to characterize CHIKV transmission and virome dynamics in Aedes albopictus during the 2025 chikungunya outbreak in Foshan, China. Mosquitoes were continuously collected throughout the outbreak using pooled and individual sampling strategies, and viral metagenomic sequencing was performed to compare virome composition between CHIKV-positive and CHIKV-negative mosquitoes. CHIKV-positive mosquitoes first appeared shortly after the onset of local human transmission and expanded geographically in parallel with the epidemic. Mosquito- and human-derived CHIKV genomes were nearly identical and belonged to a genetically homogeneous East/Central/South African (ECSA) lineage, supporting sustained local mosquito–human transmission. Despite widespread CHIKV circulation, the resident mosquito virome remained broadly stable, with sampling time contributing more to virome variation than CHIKV infection status. Only two resident viruses, Orthophasmavirus barstukorius and Guato virus, were significantly associated with CHIKV infection. Integration of pooled and individual sequencing further showed that pooled samples reliably represented the population-level core virome, whereas individual mosquitoes revealed substantial inter-individual heterogeneity. Five previously undescribed RNA viruses were also identified, expanding the known virosphere of Ae. albopictus. Together, our findings show that natural CHIKV emergence is accompanied by limited species-specific shifts rather than wholesale restructuring of the resident mosquito virome, and highlight the value of integrating genomic and metagenomic surveillance into vector-borne disease monitoring.

Graphical abstract

Introduction

Arboviral outbreaks are typically viewed through the lens of a single pathogen. In mosquito vectors, however, epidemic viruses circulate within complex viral communities composed of insect-specific viruses (ISVs) and other mosquito-associated viruses whose ecological responses to pathogen invasion remain largely unexplored. Rather than serving merely as passive background microorganisms, these resident viruses are increasingly recognized as integral components of the mosquito virome that may influence pathogen transmission (Vasilakis and Tesh, 2015), vector competence (Olmo et al., 2023), and viral evolution (Lewis et al., 2023; Shi et al., 2018). Understanding how an emerging arbovirus interacts with this pre-existing viral ecosystem has therefore become an important question in vector biology.

Advances in viral metagenomics have substantially expanded our understanding of mosquito viromes (Shi et al., 2016; Pan et al., 2024). Across mosquito species and geographical regions, diverse RNA and DNA viruses have been shown to persist as stable components of mosquito populations (Shi et al., 2019), many of which are maintained through vertical transmission and long-term host associations (Shi et al., 2020). Experimental studies further suggest that resident viruses actively interact with medically important arboviruses. For example, cell-fusing agent virus (CFAV) (Baidaliuk et al., 2019), Phasi Charoen-like virus (PCLV), and Humaita-Tubiacanga virus (HTV) (Olmo et al., 2023) have been shown to either suppress or facilitate dengue virus (DENV) and Zika virus (ZIKV) infection through mechanisms including superinfection exclusion, modulation of antiviral immunity, and competition for intracellular resources.

However, almost all current evidence supporting these virus interactions has been generated under highly simplified laboratory conditions (Shi et al., 2022). Most studies have relied on experimental infection of laboratory mosquito colonies, often focusing on interactions between a single ISV and a single arbovirus. Although these systems have provided valuable mechanistic insights, they cannot fully capture the ecological complexity of natural mosquito populations, where diverse viruses coexist and fluctuate simultaneously. Likewise, field-based virome studies have primarily emphasized viral discovery and diversity profiling, whereas relatively little is known about how naturally circulating arboviruses reshape mosquito viral communities during real outbreaks. Moreover, routine entomological surveillance generally relies on pooled mosquito samples, making it difficult to distinguish population-level community shifts from virome variation occurring within individual infected mosquitoes.

Chikungunya virus (CHIKV), a mosquito-borne alphavirus primarily transmitted by Aedes aegypti and Aedes albopictus (Ae. albopictus), has re-emerged as one of the most important arthropod-borne viruses worldwide. Since the emergence of the Indian Ocean lineage in the mid-2000s, CHIKV has caused repeated outbreaks across Africa, Asia, Europe, and the Americas, driven by globalization, climate change, and the continued expansion of competent mosquito vectors (Rocklov and Dubrow, 2020; Zhang et al., 2026). In 2025, a large local outbreak occurred in Foshan, Guangdong Province, representing one of the most substantial episodes of autochthonous CHIKV transmission documented in mainland China (Liang et al., 2026). Continuous mosquito surveillance conducted throughout this outbreak provided a rare opportunity to investigate mosquito virome dynamics under natural transmission conditions.

Here, we leveraged the 2025 CHIKV outbreak in Foshan as a natural ecological experiment to investigate how an emerging arbovirus is embedded within, and associated with, the broader mosquito virome. By integrating longitudinal entomological surveillance with complementary pooled and individual mosquito viral metagenomic sequencing, together with CHIKV whole-genome sequencing and phylogenetic analyses, we examined virome dynamics throughout the course of the outbreak at both population and individual mosquito levels. Rather than simply cataloguing mosquito-associated viruses, we sought to determine whether natural CHIKV transmission is associated with reproducible restructuring of mosquito viral communities and consistent virus associations under field conditions. Our findings provide a comprehensive characterization of mosquito virome ecology during a natural CHIKV outbreak and establish a field-based framework for understanding how resident mosquito viruses coexist with, and potentially respond to, the emergence of epidemic arboviruses.

Materials and methods

Mosquito collection

Adult mosquitoes were collected as part of routine mosquito surveillance conducted by the Foshan Center for Disease Control and Prevention (Foshan CDC) from June to September 2025 at multiple surveillance sites in Foshan, Guangdong Province, China (Supplementary Table S1). Adult mosquitoes were captured using BG-traps (Biogents, Germany) baited with BG-Lure according to the national mosquito surveillance protocol. Following collection, mosquitoes were transported to the laboratory on ice and morphologically identified to species using standard taxonomic keys. Only Ae. albopictus specimens were included in the subsequent analyses.

Sample processing and CHIKV detection

Ae. albopictus collected from the same BG-trap on the same date were separated by sex, with female and male mosquitoes processed as independent pools for CHIKV detection. Each pool was individually placed into 2 ml Micro tube (SARSTEDT, Nümbrecht, Germany) with about 10 pieces of 3.2 mm zirconia bead (Jing Xin, Shanghai, China) and 250 μl grinding medium (Minimum Essential Medium, Life Technologies, NY, USA). The samples were homogenized at 6000 rpm for 1 min using a MagNa Lyser (Roche, Mannheim, Germany) and centrifuged at 13,500 g for 3 min with a centrifuge 5430R (Eppendorf, Germany) to obtain supernatants. One aliquot of each supernatant was used for CHIKV detection by RT-qPCR, whereas another aliquot was reserved for viral metagenomic sequencing. Total RNA was extracted from 200 μl supernatant with a Nucleic Acid Extraction & Purification Kit (Tianlong Science & Technology, Xi’an, China) according to the manufacturer’s instructions. The sample RNA was screened for Chikungunya by an RT-qPCR assay (AODONG INSPECTION & TESTING, Shenzhen, China) according to the manufacturer’s instructions on a CFX96 Deepwell Dx Real-Time PCR Detection System (BIO-RAD, CA, USA), and positive result was denoted by cycle threshold (Ct) value.

During the peak outbreak period (7–15 August 2025), an additional individual-mosquito sampling strategy was implemented. Individual Ae. albopictus mosquitoes were dissected into heads and bodies under sterile conditions. Mosquito heads were subjected to CHIKV detection by RT-qPCR, while the corresponding bodies were preserved for viral metagenomic sequencing.

Viral metagenomics library prepration and sequencing

Samples selected for viral metagenomic sequencing included CHIKV-positive and matched CHIKV-negative pooled mosquito homogenates, as well as bodies of individual mosquitoes whose corresponding heads had been tested for CHIKV by RT-qPCR (Supplementary Table S2). A PBS blank control was processed alongside all samples throughout the experimental workflow to monitor potential contamination. 150 μl supernatant of selected samples were then used for filtration (0.8 μm pore size) to enrich for viral particles. The filtrate was then treated with a cocktail of Benzonase (Novagen) and Micrococcal Nuclease (New England Biolabs) in a homemade buffer (1 M Tris, 100 mM CaCl₂, and 30 mM MgCl₂) to digest free-floating nucleic acids. DNA and RNA were extracted (QIAGEN Viral RNA mini kit). Sequencing libraries were prepared using the VAHTS Universal V8 RNA-seq Library Prep Kit for Illumina (Vazyme, Nanjing, China) following the manufacturer’s protocol. Sequencing of the samples was performed on a NextSeq-1000CN Sequencing System (Illumina) for 300 cycles (2 × 150 bp paired ends).

Viral metagenomic sequence analysis

High-throughput sequencing data were processed using the bioinformatics pipeline ViPER (ViPER). Paired-end reads were first processed using Trimmomatic (Bolger et al., 2014) to remove sequencing adapters and trim low quality bases (Supplementary Table S2). Viral taxa detected in the PBS blank control were regarded as potential contaminants and excluded from downstream analyses. Quality filtered reads from each sample were independently subjected to de novo assembly using metaSPAdes (Bankevich et al., 2012), with k-mer sizes of 21, 33, 55, and 77 and a minimum contig length of 500 bp. To minimize fragmentation of highly abundant viral genomes caused by excessive sequencing depth, each sample was assembled using three input datasets: the complete read set, a randomly subsampled set containing 10% of the reads, and a second subsampled set containing 1% of the reads. Contigs generated from the three assemblies were combined and clustered to remove redundancy. Contigs were grouped using thresholds of 95% average nucleotide identity and at least 85% alignment coverage, and the longest contig within each cluster was selected as the representative sequence.

Putative eukaryotic viral contigs were identified by aligning predicted protein sequences against the NCBI non-redundant protein database using DIAMOND (Buchfink et al., 2015) in sensitive mode. Taxonomic assignments were generated using KronaTools (Ondov et al., 2011) according to a lowest common ancestor approach (Supplementary Table S3). Viral abundance (Supplementary Table S4) was quantified by mapping the quality-filtered reads from each sample to the non-redundant contig catalogue using BWA-MEM2 (Li and Durbin, 2009). SAMtools (Li et al., 2009) was then used to calculate the breadth of coverage of each representative contig and to generate a consensus sequence for each classified viral contig.

Bayesian Markov chain Monte Carlo (MCMC) analysis

To reconstruct the evolutionary history of CHIKV circulating in Foshan, a Bayesian time-scaled phylogenetic analysis was performed using the Bayesian Markov chain Monte Carlo (MCMC) framework implemented in BEAST v2.7.7 (Baele et al., 2025). This analysis was used to estimate viral evolutionary rates and reconstruct the evolutionary history of CHIKV. The BEAST input file was generated using Bayesian Evolutionary Analysis Utility (BEAUti). Phylogenetic analyses were conducted under the HKY nucleotide substitution model with four gamma-distributed rate categories (HKY+Γ4) to account for among-site variation in evolutionary rates. A strict molecular clock model was applied, assuming a constant substitution rate across all branches. The demographic history of CHIKV populations was modeled using the Bayesian Skyline coalescent model. MCMC chains were run for 100 million generations, with parameter states sampled every 10,000 generations. The convergence of MCMC analyses and the reliability of parameter estimates were assessed using Tracer v1.7.2 by calculating the effective sample size (ESS) for each parameter. All parameters achieved ESS values of ≥200, indicating adequate sampling and convergence.Based on the posterior distribution of trees generated from the MCMC analysis, a maximum clade credibility (MCC) tree was reconstructed using TreeAnnotator v2.7.7 implemented in the BEAST package after discarding 10% burn-in samples. The resulting phylogenetic tree was visualized and further annotated using FigTree v1.5.0, with different viral lineages highlighted by branch colors to illustrate their evolutionary relationships.

Phylogenetic analysis of novel viruses

To determine the phylogenetic placement of the viruses identified in this study within the known viral diversity, phylogenetic analyses were performed based on the amino acid sequences of the RNA-dependent RNA polymerase (RdRp) protein-coding regions of each virus. Reference viral sequences were initially retrieved from Genbank, and representative homologous protein sequences were further identified from multiple public databases through BLASTP searches (Camacho et al., 2009). Multiple sequence alignments of the RdRp amino acid sequences were performed using MAFFT (Katoh et al., 2002), followed by the removal of poorly aligned and highly variable regions using Gblocks (Gblocks). The resulting trimmed alignments were converted into PHYLIP format and subsequently analyzed using ProtTest (Darriba et al., 2011) to determine the optimal amino acid substitution model. The best fit model was selected based on the Akaike Information Criterion, which evaluates the relative goodness of fit among alternative substitution models. Phylogenetic trees were then reconstructed using the maximum-likelihood (ML) method implemented in PhyML 3.0 (Guindon et al., 2010) under the optimal substitution model determined by ProtTest. The resulting phylogenetic trees were further edited, annotated, and visualized using the Interactive Tree of Life (iTOL) (Letunic and Bork, 2024).

Statistical analysis

All statistical analyses and data visualization were performed in R using the phyloseq (McMurdie and Holmes, 2013), vegan (Dixon, 2003), and ggplot2 (Wickham, 2016) packages. Alpha diversity was assessed by observed richness using the phyloseq package. Differences in continuous variables, including alpha diversity indices and viral abundances, between groups were evaluated using the Wilcoxon rank-sum test (or pairwise Wilcoxon test where appropriate). P values from multiple comparisons were adjusted using the Benjamini-Hochberg (BH) procedure, with FDR < 0.05 considered statistically significant.

Beta diversity was calculated using Bray-Curtis dissimilarity and compared among groups using permutational multivariate analysis of variance (PERMANOVA) with 999 permutations implemented in the vegan package. To investigate the contribution of environmental and epidemiological variables to virome composition, distance-based redundancy analysis (dbRDA) based on Bray-Curtis dissimilarities was performed using the capscale function in vegan. The significance of the overall dbRDA model, individual explanatory variables (marginal effects), and sequentially added terms was assessed by permutation test.

Results

Entomological surveillance captured the establishment and persistence of CHIKV in local Ae. albopictus populations

Following the introduction of CHIKV into Foshan in early July 2025, the outbreak rapidly expanded from Shunde District to neighboring districts, becoming the largest recorded chikungunya outbreak in Guangdong Province. Previous epidemiological investigations identified Lecong and Beijiao as the earliest outbreak foci, prompting intensified entomological surveillance in these areas.

A total of 1778 Ae. albopictus pools, comprising 6862 mosquitoes, were collected between 1st June and 14th September 2025 from surveillance sites mainly distributed across the three administrative districts of Foshan (Fig. 1A and Supplementary Table S1). Overall, 50 pools (2.8%) tested positive for CHIKV on 25 independent surveillance days (Fig. 1B, C), with pool sizes ranging from 1 to 47 mosquitoes (median = 4). Surveillance intensity was not spatially uniform but reflected the epidemiological situation, with Shunde accounting for the majority of collections (1170 pools), followed by Chancheng (328 pools) and Nanhai (279 pools), whereas only limited sampling was required in Sanshui where no substantial local transmission occurred (Fig. 1A).

Fig. 1.

Entomological surveillance and viral metagenomic sampling during the 2025 CHIKV outbreak in Foshan, China. A) Geographic distribution of Ae. albopictus mosquito surveillance sites across the five administrative districts of Foshan from June to September 2025. Circle size represents the number of mosquito pools collected at each sampling site. Red circles indicate sampling sites where CHIKV positive mosquito pools were detected. B) Workflow for pooled mosquito viral metagenomic analysis. Ae. albopictus mosquitoes collected from the same sampling site, date, species and sex were combined into individual pools, homogenized, screened for CHIKV by RT-qPCR, and classified as CHIKV positive or negative. Supernatants from selected pools were subjected to viral metagenomic sequencing. C) Temporal dynamics of mosquito surveillance and CHIKV detection. The upper panel shows the daily number of mosquito pools tested from June to September 2025, with CHIKV-positive pools highlighted in red. The lower panel shows the temporal distribution of CHIKV-positive mosquito pools at the town level. Circle size represents the number of mosquito pools collected on each sampling day, and red circles indicate CHIKV-positive pools. Female (♀) and male (♂) positive pools are indicated where applicable. D) Workflow for individual mosquito viral metagenomic analysis. Individual Ae. albopictus collected during the outbreak were dissected into heads and bodies. Mosquito heads were screened for CHIKV by RT-qPCR, whereas the corresponding bodies were subjected to viral metagenomic sequencing. CHIKV positive and negative mosquitoes were included for comparative virome analyses.

Throughout June, all mosquito pools remained CHIKV-negative despite routine surveillance (Fig. 1C). Continuous surveillance was subsequently intensified in Lecong and Beijiao, the earliest affected communities, during the three days preceding the first positive detection. On 11 July, the first CHIKV-positive mosquito pool was identified from a female pool containing 28 Ae. albopictus collected in Lecong, representing the earliest entomological evidence of local CHIKV circulation. Notably, a CHIKV-positive male pool was detected in Lecong on the following day (12 July), with additional positive male pools detected on 16 July and 19 August. Because male mosquitoes do not acquire CHIKV through blood feeding, repeated detection of infected male pools provides indirect evidence that CHIKV was persistently maintained within local mosquito populations.

CHIKV-positive mosquito pools increased rapidly during July and remained detectable throughout August and into September (Fig. 1C). Repeated positive detections were continuously observed in both Lecong and Beijiao, the two earliest outbreak foci, whereas positive mosquito pools gradually appeared in additional towns, including Zumiao, Shiwan, and Zhangcha (Chancheng) and Guicheng, Dali, and Xiqiao (Nanhai), paralleling the geographic expansion of the human epidemic from Shunde to neighboring districts. Importantly, the first mosquito-positive pool was detected only after several consecutive days of negative surveillance in the same outbreak focus, indicating that mosquito infection emerged shortly after, rather than preceding, the onset of local human transmission. Together, mosquito surveillance captured the complete temporal trajectory of CHIKV circulation, from the absence of detectable mosquito infection before the outbreak, through rapid establishment and geographic expansion in July and August, to continued but declining detection during September.

To characterize the mosquito virome associated with CHIKV infection, viral metagenomic sequencing was primarily performed on individual Ae. albopictus collected between 7 and 15 August 2025, when mosquito activity and human transmission were both high. Individual mosquitoes were dissected into heads and bodies, with heads used for CHIKV detection by RT-qPCR and the corresponding bodies reserved for viral metagenomic sequencing (Fig. 1D). To increase representation of CHIKV-positive mosquitoes across the outbreak, additional CHIKV-positive pooled mosquito homogenates collected at different time points were included. Corresponding CHIKV-negative individual mosquitoes or pooled samples were selected as matched controls for comparative virome analyses (Fig. 1B, D). The resulting datasets enabled characterization of CHIKV-associated virome dynamics at both the population and individual mosquito levels.

Genomic sequencing links mosquito-derived CHIKV to the contemporary Foshan outbreak lineage

Metagenomic sequencing detected substantial variation in CHIKV abundance among RT-qPCR positive Ae. albopictus samples, with viral read counts spanning several orders of magnitude. Despite this variation, genome coverage remained high across most samples, allowing recovery of near complete CHIKV genomes directly from field-collected mosquitoes (12021 bp, GenBank accession no. PZ746565) (Fig. 2A). CHIKV read abundance was inversely correlated with RT-qPCR Ct values (Spearman’s R = −0.59, P < 0.001), supporting concordance between molecular viral load estimates and metagenomic sequencing output (Fig. 2B and Supplementary Table S2). Representative high and low titer samples further demonstrated continuous genome-wide coverage, although sequencing depth became more heterogeneous in mosquitoes with lower viral loads (Fig. 2C). These results demonstrate that viral metagenomic sequencing provided robust genome recovery across a broad range of infection levels.

Fig. 2.

Genomic characterization of mosquito-derived CHIKV recovered during the 2025 Foshan outbreak. A) CHIKV sequencing reads and genome coverage recovered from RT-qPCR positive Ae. albopictus mosquitoes. Samples are ordered according to CHIKV read abundance. Bars indicate CHIKV reads (left y-axis, log10 scale), and the orange line indicates genome coverage (right y-axis). B) Correlation between CHIKV sequencing reads and RT-qPCR Ct values. The solid line represents the linear regression fit with the 95% confidence interval (gray shading). Spearman’s correlation coefficient (R) and P value are indicated. C) Representative genome-wide sequencing depth profiles of two mosquito-derived CHIKV genomes with high (upper panel) and relatively low (lower panel) viral loads. Genome coverage and corresponding Ct values are indicated for each sample. D) MCC phylogeny of representative global CHIKV genomes. Branches are colored according to CHIKV genotype: ECSA, Asian, Indian Ocean lineage (IOL), and West African. The mosquito-derived genome generated in this study is highlighted by a yellow star. E) Enlarged view of the ECSA clade from the MCC phylogeny showing the phylogenetic relationship between the mosquito-derived CHIKV genome and representative global ECSA strains. Human-derived CHIKV genomes from the 2025 Foshan outbreak and representative sequences from Réunion Island are included for comparison. The mosquito-derived genome obtained in this study is indicated by a yellow star.

Phylogenetic analysis based on complete genome sequences assigned the mosquito-derived CHIKV genome to the East/Central/South African (ECSA) genotype, specifically the ECSA-2 sublineage (Fig. 2D). In the time-resolved MCC phylogeny, the mosquito-derived genome clustered tightly with contemporaneous CHIKV genomes recovered from human cases during the 2025 Foshan outbreak, forming a well-supported outbreak clade within the ECSA-2 sublineage (Fig. 2E). This outbreak clade was most closely related to recent CHIKV strains from Réunion Island. Consistent with previously characterized Foshan outbreak strains (Sun et al., 2025), the mosquito-derived genome carried the characteristic substitutions C:T143A, E1:A226V, E2:L210Q, and E2:I211T.Comparison with the published Foshan genome further revealed complete amino acid identity across the coding region, with only minor nucleotide differences confined to the terminal untranslated regions, indicating that no additional coding mutations were detected in the mosquito-derived genome. These findings demonstrate that mosquitoes and human cases shared a genetically homogeneous CHIKV lineage throughout the outbreak.

The resident virome of Ae. albopictus remained largely stable during the CHIKV outbreak despite substantial inter individual variation

To investigate whether CHIKV infection was associated with alterations of the mosquito virome, we characterized the eukaryotic viromes of 51 Ae. albopictus samples, including both pooled mosquito homogenates and individually sequenced mosquitoes collected during the outbreak. A total of 27 eukaryotic viral taxa were identified, comprising predominantly insect-associated RNA viruses together with CHIKV (Fig. 3A, Supplementary Table S3 and S4). The virome was consistently dominated by a limited number of insect associated viruses, including Wenzhou sobemo-like virus 4, Permutotetraviridae sp., Culivirus dunyae, Orthophasmavirus barstukorius, Yongsan bunyavirus 1 (Foshan bunyavirus) and Longgang virus, which were detected across most CHIKV-positive and CHIKV-negative mosquitoes (Fig. 3A). In contrast, the remaining viral taxa occurred sporadically and were generally restricted to a subset of samples, indicating that variation among mosquitoes was primarily attributable to low-abundance viruses rather than replacement of the dominant viral community. Notably, individually sequenced mosquitoes displayed considerably greater variation in the relative abundance of individual viruses than pooled samples, but exhibited the same dominant viral taxa, indicating that mosquito pooling reduced stochastic inter-individual variation without substantially altering the overall virome composition. Consistent with the heatmap, stacked relative abundance profiles showed that the dominant virome composition was broadly comparable between CHIKV-positive and CHIKV-negative mosquitoes (Fig. 3B). CHIKV-positive samples were characterized by abundant CHIKV reads, whereas the relative abundance of the major insect-associated viruses remained largely preserved, providing little evidence for extensive restructuring of the mosquito virome following CHIKV infection.

Fig. 3.

Composition and ecological characteristics of the Ae. albopictus virome during the 2025 CHIKV outbreak in Foshan. A) Heatmap showing the abundance of eukaryotic viruses detected in Ae. albopictus metagenomes. Columns represent individual metagenomic libraries and rows represent viral taxa. Samples are grouped according to CHIKV infection status. Abundance is displayed as log10-transformed viral read counts. The color bar above the heatmap indicates the number of mosquitoes included in each library: grey denotes individually sequenced mosquitoes (one mosquito per library), whereas red shades denote pooled mosquito samples. Samples marked with an asterisk (*) represent individually processed mosquitoes in which the head was used for CHIKV RT-qPCR screening and the corresponding body was subjected to viral metagenomic sequencing. The right annotation indicates the highest BLASTx amino acid identity of each viral contig to reference sequences. B) Relative abundance of viral taxa in each mosquito metagenomic library. Bars represent individual libraries and are colored according to viral taxa. Samples are grouped by CHIKV infection status. C) Alpha diversity of the mosquito virome measured by Observed richness and Shannon diversity. Samples are grouped by sampling month and CHIKV infection status. Statistical significance was assessed using the Wilcoxon rank-sum test with pairwise comparisons. D) dbRDA of Bray-Curtis dissimilarities showing variation in virome composition among mosquito samples. Samples are colored according to sampling month and shaped by CHIKV infection status. Shaded ellipses represent the 60% confidence interval for each sampling month. The arrow indicates the fitted environmental vector for mosquito number. E) Relative abundances of representative viruses showing significant differences between CHIKV-positive and CHIKV-negative mosquitoes. Statistical significance was determined using the Wilcoxon rank-sum test (*p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001).

Analysis of alpha diversity revealed limited effects of CHIKV infection on overall virome diversity (Fig. 3C). Observed viral richness differed primarily among sampling months, with July samples exhibiting significantly higher richness than August samples (P < 0.01) and a similar decreasing trend relative to September (P = 0.052), whereas Shannon diversity showed no consistent differences between CHIKV-positive and CHIKV-negative mosquitoes within each month. These results suggest that temporal variation contributed more strongly to changes in virome richness than CHIKV infection itself.

To further evaluate factors associated with community composition, we performed dbRDA using Bray-Curtis dissimilarities (Fig. 3D). Samples exhibited partial clustering according to sampling month, whereas CHIKV-positive and CHIKV-negative mosquitoes largely overlapped, indicating limited separation attributable to infection status. The overall constrained model was significant (permutation test, P = 0.009), demonstrating that ecological variables jointly explained variation in mosquito virome composition (Supplementary Table S5). However, marginal permutation tests indicated that no individual variable retained an independent effect after accounting for the others, suggesting substantial shared variation among sampling month, sampling location and pool size, while CHIKV infection status contributed minimally to overall community structure.

Although the global resident virome remained stable, differential abundance analysis identified a limited number of viruses associated with CHIKV infection status (Fig. 3E). Orthophasmavirus barstukorius and Guato virus were both significantly more abundant in CHIKV-negative mosquitoes than in CHIKV-positive mosquitoes, whereas the dominant resident viruses showed no consistent differences between groups. These observations indicate that CHIKV infection was associated with changes in the abundance of a small subset of resident viruses rather than broad disruption of the mosquito virome.

Discovery of novel mosquito-associated RNA viruses

To investigate the evolutionary relationships of previously unclassified viral contigs identified in the mosquito virome, representative RdRp or polyprotein amino acid sequences were used for phylogenetic reconstruction. Five highly divergent viral genomes clustered within distinct lineages of Riboviria, including members of the Nodaviridae, Orthophasmavirus, Partitiviridae, and Endornaviridae (Fig. 4A–E).

Fig. 4.

Phylogenetic characterization of putative novel RNA viruses identified in Ae. albopictus collected during the 2025 CHIKV outbreak in Foshan. Maximum-likelihood phylogenetic trees were reconstructed using amino acid sequences of the RdRp or viral polyprotein from representative members of the corresponding viral groups together with the viral sequences identified in this study. The viruses identified from Foshan mosquitoes are highlighted in red. Branch support values indicate ultrafast bootstrap percentages, and branch lengths represent the estimated number of amino acid substitutions per site. Collapsed clades are shown as grey triangles, with triangle size proportional to the number of sequences contained within each clade. Scale bars indicate amino acid substitutions per site. Maximum-likelihood phylogeny of A) Foshan insect virus 21, B) Foshan bunyavirus 1, C) Foshan partiti-like virus 1, D) Foshan partiti-like virus 2 and E) Foshan endornavirus 1.

One virus, provisionally designated Foshan insect virus 21, formed an independent branch within Riboviria and showed the highest similarity to Qinjiang noda-like virus 72 while remaining clearly separated from currently recognized members of Nodaviridae, suggesting that it represents a divergent noda-like virus lineage (Fig. 4A). A second virus, Foshan bunyavirus 1, clustered within the genus Orthophasmavirus, where it grouped most closely with Aedes chasmavirus and Yunnan bunyavirus 1 but occupied a distinct evolutionary position from previously described mosquito-associated orthophasmaviruses (Fig. 4B). Phylogenetic analysis further identified a novel partiti-like virus that grouped within the Betapartitivirus lineage together with several mosquito-derived partiti-like viruses, including Hubei partiti-like virus 22 and Wuhan insect virus 22, indicating that this virus belongs to the expanding diversity of mosquito-associated partitiviruses (Fig. 4C). Another partiti-like virus, Foshan partiti-like virus 2, clustered within the broader Partitiviridae radiation but remained distinct from established genera, suggesting substantial evolutionary divergence (Fig. 4D). In addition, a highly divergent virus, tentatively named Foshan endornavirus 1, was positioned within the Alphaendornavirus clade of the family Endornaviridae. Although it shared ancestry with several fungal and plant associated endornaviruses, it formed an independent lineage separated from currently recognized species, supporting its classification as a putative novel endornavirus (Fig. 4E).

Collectively, these phylogenetic analyses demonstrate that mosquitoes collected during the Foshan CHIKV outbreak harbored multiple previously undescribed RNA viruses spanning diverse evolutionary lineages, substantially expanding the known virome diversity associated with Ae. albopictus.

Discussion

Continuous entomological surveillance during the 2025 Foshan outbreak captured the establishment and expansion of CHIKV within local Ae. albopictus populations. CHIKV positive mosquitoes first appeared shortly after local human transmission (Zhang et al., 2025) was recognized and subsequently spread in parallel with the geographic progression of the outbreak, indicating rapid viral amplification following introduction (Fig. 1A and C). Detection of CHIKV in male mosquitoes further suggests that virus circulation may not have been restricted to blood fed females and is compatible with vertical or venereal transmission (Mavale et al., 2010). Mosquito and human derived CHIKV genomes formed a highly homogeneous ECSA-2 sublineage (Fig. 2), consistent with sustained local mosquito-human transmission rather than repeated independent introductions (Sun et al., 2025; Zhao et al., 2025). These findings demonstrate that combining longitudinal mosquito surveillance with viral genomics provides a powerful framework for reconstructing arbovirus transmission dynamics, complementing conventional case-based surveillance and strengthening One Health approaches for outbreak investigation.

Against the background of extensive CHIKV circulation, the resident Ae. albopictus virome exhibited remarkably limited community wide change (Fig. 3). The dominant ISVs remained prevalent regardless of CHIKV infection status, whereas temporal variation exerted a stronger influence on virome composition than CHIKV itself. Previous surveys of Ae. albopictus in China have similarly revealed substantial virome heterogeneity across seasons, geographic locations, urban environments and mosquito populations (Xu et al., 2023; Shen et al., 2026; Li et al., 2025; He et al., 2021). Our findings extend these observations to an active CHIKV outbreak, showing that such background variation remained more prominent than the community-wide changes associated with CHIKV infection. This apparent stability may reflect several factors. The study captured only a single epidemic wave, which may have been insufficient to reveal slower ecological succession within the mosquito virome. In addition, natural CHIKV infections are highly heterogeneous with respect to infectious dose, mosquito age and environmental exposure, potentially obscuring subtle community level effects (Lambrechts et al., 2010). Finally, many dominant mosquito-associated viruses are persistent infections that may be vertically maintained across mosquito generations, forming a stable core virome that is relatively resistant to short term perturbation (Lee et al., 2019). Whether this stability reflects an intrinsic ecological resilience of the mosquito virome or simply the limited observation period remains an open question requiring longer term surveillance.

Despite the overall stability of the virome, Orthophasmavirus barstukorius and Guato virus differed consistently between CHIKV-positive and CHIKV-negative mosquitoes, suggesting that specific responses may occur without altering overall community structure. These associations may result from changes in mosquito physiology, competition for cellular resources, shared antiviral immune responses or other ecological factors linked to CHIKV infection (Olmo et al., 2023; Guegan et al., 2018). Given the observational nature of this study, causal relationships cannot be inferred.

A major strength of this study was the integration of pooled and individual mosquito metagenomic approaches (Fig. 1B and D). Pooled samples efficiently characterized the population level core virome, whereas individual mosquitoes revealed substantial heterogeneity associated with natural infection (Fig. 3A and B) (Atoni et al., 2019). The consistent detection of dominant resident viruses in both datasets indicates that pooled sequencing reliably captures the core virome while individual sequencing provides additional ecological resolution (Shi et al., 2019). Together, these complementary strategies offer a practical framework for mosquito virome surveillance during arbovirus outbreaks. We also identified five previously undescribed RNA viruses associated with Ae. albopictus (Fig. 4), further highlighting the largely unexplored diversity of mosquito-associated viruses (Shi et al., 2016) and providing genomic resources for future studies of virus evolution and arbovirus-ISVs interactions (Gutierrez-Bugallo et al., 2019).

Several limitations should be acknowledged. First, this study captured a single CHIKV outbreak within one geographic region and a relatively short observation period, which may not fully represent longer-term virome dynamics or other transmission settings. Second, as an observational metagenomic study, it cannot distinguish active viral replication from the presence of viral nucleic acids or establish causal interactions between CHIKV and resident viruses. These questions will require longitudinal surveillance across multiple transmission seasons together with experimental validation.

Despite these limitations, this study provides a field-based characterization of mosquito virome dynamics during a natural CHIKV outbreak. By integrating continuous entomological surveillance, viral genomics and metagenomic sequencing, we demonstrate that CHIKV rapidly established within local Ae. albopictus populations while the resident virome remained broadly stable at the community level, accompanied by only limited changes. These findings suggest that early CHIKV emergence may be accommodated within an established mosquito viral ecosystem rather than immediately restructuring it, providing new insights into mosquito virome ecology during natural arbovirus transmission.

CRediT authorship contribution statement

Tong Wang: Formal analysis, Investigation, Methodology, Data curation, Writing – original draft, Writing – review & editing. Yi Xie: Resources, Investigation, Methodology, Data curation, Writing – review & editing, Writing – original draft. Shizhen Liu: Formal analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing. Yuanpeng Xin: Resources, Data curation, Writing – review & editing. Jianwen You: Resources, Data curation, Writing – review & editing. Hui Zeng: Resources, Data curation, Writing – review & editing. Huaiyuan Wang: Resources, Data curation, Writing – review & editing. Yanrong Li: Resources, Data curation, Writing – review & editing. Baisheng Li: Conceptualization, Resources, Writing – review and editing. Xinchun Chen: Conceptualization, Writing – review and editing. Xing Liu: Funding acquisition, Investigation, Supervision, Writing – review and editing. Jintao Liang: Funding acquisition, Resources, Investigation, Supervision, Project administration, Writing – review and editing. Chenyan Shi: Conceptualization, Funding acquisition, Supervision, Project administration, Investigation, Visualization, Writing – original draft, Writing – review & editing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

This project was funded by grants from the Natural Science Foundation of Guangdong Province (No. 2025A1515011647), Guangdong Medical Science and Technology Research Fund (B2026307) and Shenzhen Science and Technology Program (ZDCYKCX20250901093300001).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. Supplementary materials

Supplementary Table S1. Mosquito survillence information

Supplementary Table S2. Metadata of samples used for viral metagenomic sequencing

Supplementary Table S3. Assembly and annotation statistics of viral contigs.

Supplementary Table S4. Abundance of viral taxa identified in mosquito samples.

Supplementary Table S5. Permutation test results for the dbRDA model.

Data availability

The raw metagenomic sequencing data generated in this study have been deposited in the Genome Sequence Archive (GSA) under BioProject accession PRJCA069978. The genome sequences of the newly identified viruses have been deposited in GenBank under accession numbers PZ746565-PZ746570. All other data supporting the findings of this study are available within the article and its Supplementary Information files or are available from the corresponding author upon reasonable request.