{"paper_id":"712af127-b49c-4f6d-9074-d74b1b39f07c","body_text":"1 \nMetabolic reprogramming and gut microbiota ecology drive divergent \nPlasmodium vivax infection outcomes in Anopheles darlingi \n \nBianca Cechetto Carlos1,2,3, Kamila Voges1,2,3, Pedro Henrique de Andrade Affonso1,2, Amie Jaye4, \nCarlos Tong Rios5,7, Bruno Tinoco-Nunes1,2, Diego Peres Alonso3, Robert M. MacCallum4, Marta \nMoreno6,8, Dina Vlachou4*, Jayme A. Souza-Neto1,2,3,9* & George K. Christophides3,4* \n \n1São Paulo State University, School of Agricultural Sciences, Department of Bioprocesses and \nBiotechnology, Botucatu, Brazil \n2São Paulo State University, School of Agricultural Sciences, Central Multiuser Laboratory, \nBotucatu, Brazil \n3São Paulo State University, Institute of Biotechnology, Botucatu, Brazil  \n4Department of Life Sciences, Imperial College London, London, United Kingdom  \n5Laboratorio ICEMR-Amazonia, Laboratorios de Investigacion y Desarrollo, Facultad de Ciencias y \nFilosofia, Universidad Peruana Cayetano Heredia, Lima, Peru \n6Division of Infectious Diseases, Department of Medicine, University of California San Diego, La \nJolla, California, USA \n7Current address: Unidad de Entomología del Laboratorio de Referencia Regional, Gerencia \nRegional de Loreto (GERESA), Perú \n8Current address: Department of Infection Biology; London School of Hygiene & Tropical \nMedicine, Keppel Street, WC1E 7HT, London, United Kingdom \n9Current addresses: Department of Diagnostic Medicine/Pathobiology, College of Veterinary \nMedicine, Kansas State University, Manhattan, KS 66506, USA; Kansas Veterinary Diagnostic \nLaboratory, College of Veterinary Medicine, Kansas State University, Manhattan, KS 66506, USA \n  \n*Correspondence: G.K.C. (g.christophides@imperial.ac.uk) or J.A.S.N. (jsouzaneto@vet.k-\nstate.edu) or D.V. (d.vlachou@imperial.ac.uk)  \n \nKey works: Anopheles darlingi, Plasmodium vivax, vector competence, malaria transmission, \nmidgut microbiota, redox regulation, metabolic reprogramming, Amazon basin \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 2 \nAbstract \nAnopheles darlingi is the principal malaria vector in the Amazon basin, where Plasmodium vivax \naccounts for the majority of cases. Despite its epidemiological importance, the molecular and \nmicrobial determinants of A. darlingi susceptibility to P. vivax remain poorly understood. Here, we \ninvestigated vector-parasite-microbiota interactions using experimental infections with field-derived \nP. vivax gametocytaemic blood, which produced two distinct infection phenotypes: low and high \noocyst burdens. Transcriptomic profiling of mosquito midguts across key parasite developmental \ntimepoints revealed that low-infection mosquitoes mounted an early and sustained response \ncharacterised by activation of detoxification pathways, redox regulation, aromatic amino acid \ncatabolism, and purine depletion, likely coordinated through neurophysiological cues, which \ncollectively create a metabolically restrictive environment for parasite development. These \nphysiological changes were accompanied by reduced bacterial diversity and enrichment of \nEnterobacteriales and Pseudomonadales, taxa previously linked to anti-Plasmodium activity. \nConversely, high-infection mosquitoes exhibited limited metabolic reprogramming, expansion of \nFlavobacteriales, and transcriptional signatures consistent with permissive physiological states, \npotentially associated with reproductive trade-offs. Importantly, low infection outcomes \nconsistently arose from bloodmeals with the lowest gametocyte densities, suggesting that host- and \nparasite-derived components of the bloodmeal act as early conditioning factors that prime the \nmosquito midgut for either resistance or susceptibility. These findings reframe A. darlingi vector \ncompetence to P. vivax not as a fixed immune trait but as a dynamic outcome of early redox, \nmetabolic, and microbial interactions. They also highlight ecological and physiological targets for \ntransmission-blocking strategies and reinforce the importance of studying vector-parasite \ninteractions in regionally relevant systems. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 3 \nIntroduction \nMalaria, a mosquito-borne disease caused by protozoan parasites of the genus Plasmodium, remains \na major global health challenge. In 2023, an estimated 263 million people were infected, resulting in \n597,000 deaths worldwide [1]. While Plasmodium falciparum accounts for most global malaria \ncases and fatalities, Plasmodium vivax continues to cause a substantial proportion of the burden, \nparticularly outside sub-Saharan Africa. In the Americas, P. vivax is the predominant species, \nresponsible for 68-72% of reported malaria cases, with the highest burden concentrated in countries \nof the Amazon basin. Transmission in this region is largely confined to rural and remote \ncommunities, where limited access to healthcare, diagnostics, and treatment impedes control efforts \n[2]. The ability of P. vivax to relapse via dormant liver-stage hypnozoites presents an additional \nchallenge, sustaining transmission even in areas of low endemicity and complicating elimination \nstrategies. \nAnopheles darlingi is the principal malaria vector in the Amazon basin. This neotropical mosquito \nspecies is highly anthropophilic, endophilic, aggressive, opportunistic and susceptible to \nPlasmodium infection [3-5]. Despite its critical epidemiological importance in the Americas, A. \ndarlingi remains largely understudied, and the molecular mechanisms underpinning its interactions \nwith Plasmodium are still poorly understood. \nAfter ingestion by female mosquitoes during a blood meal, Plasmodium gametocytes rapidly \ndifferentiate into gametes in the mosquito midgut. Fertilization leads to zygote formation, which \nsubsequently develops into motile ookinetes. These ookinetes traverse the peritrophic matrix and \nthe midgut epithelium, eventually anchoring beneath the basal lamina to form oocysts. The traversal \nof the midgut, typically between 18 and 26 hours post-bloodmeal, represents a major population \nbottleneck for the parasite [6-8]. Parasites that survive this barrier multiply within oocysts via \nsporogony, producing thousands of sporozoites. Between 9 and 12 days later, mature oocysts \nrupture, releasing sporozoites into the hemocoel, from where they migrate to the salivary glands for \ntransmission to a new vertebrate host. \nThe successful completion of this complex Plasmodium developmental journey is shaped by \nmultiple vector-derived factors, particularly the mosquito innate immune response and midgut \nmicrobiota composition [9-11]. Genome-wide transcriptomic studies and functional analyses in \nmodel systems have uncovered a diverse immune landscape that governs mosquito susceptibility to \ninfection [12-14]. These immune responses are not only responsive to Plasmodium but also to \nbacteria, indicating extensive crosstalk between antibacterial and antiparasitic defenses [15-18]. \nOur current understanding of these immune responses is almost entirely based on studies of A. \ngambiae infected with P. falciparum or rodent parasites such as P. berghei. In contrast, the A. \ndarlingi-P. vivax transmission system, which is central to malaria persistence in the Americas, has \nbeen largely neglected. This knowledge gap stems in part from the relatively late development of \ngenomic and experimental resources: the first draft genome of A. darlingi was published only in \n2013 [19], and stable laboratory colonies were established in Peru, Brazil and French Guiana within \nthe last decade [20-23]. A more fundamental barrier is the absence of a continuous in vitro culture \nsystem for P. vivax [24], which restricts experimental infection studies to field settings with \nsimultaneous access to both laboratory-reared mosquitoes and blood from naturally infected \ngametocyte carriers [3, 22, 25]. \nHere we investigate the tripartite interaction between A. darlingi, its midgut microbiota, and P. \nvivax using a semi-field experimental system in Iquitos, Peru. By infecting colony-reared A. \ndarlingi with blood from naturally infected P. vivax carriers and combining transcriptomic and \nmicrobiome analyses across multiple timepoints and infection intensities, we identify metabolic, \nand microbiota trajectories associated with vector competence. Our findings suggest that P. vivax \ndevelopmental success is shaped by orchestrated transcriptional reprogramming of mosquito \nmetabolism and shifts in microbiota diversity, likely triggered by host- and/or parasite-derived \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 4 \nfactors in the bloodmeal. These findings provide new insight into a neglected but epidemiologically \nimportant malaria transmission system. \nResults & Discussion \nExperimental infections reveal distinct oocyst load phenotypes \nWe conducted experimental A. darlingi infections using blood samples from six P. vivax \ngametocytaemic patients (P1-P6) with varying gametocyte densities (Dataset 1A). Adult female A. \ndarlingi mosquitoes were divided into 12 groups: six were fed with untreated blood containing \npotentially infectious gametocytes (I groups), and six with heat-inactivated blood serving as non-\ninfectious controls (C groups) using direct membrane feeding assays (DMFAs). \nOocyst enumeration in dissected midguts 7 days post blood feeding (pbf) revealed a complete \nabsence of infection in all C groups, validating the heat inactivation protocol [26]. In contrast, all I \ngroups displayed varying degrees of infection (Fig. 1A and Dataset 1B). Two distinct infection \nphenotypes emerged: a low oocyst burden group (P1, P3, P4) and a high oocyst burden group (P2, \nP5, P6). Median and mean oocyst loads were largely concordant across groups, except for P6, \nsuggesting a near-normal distribution of infection intensities in most groups. This contrasts with the \ncommonly observed overdispersion in P. falciparum infections of A. gambiae [13, 26]. \nInterestingly, oocyst prevalence remained high and relatively uniform across infection groups \n(67%-100%), regardless of the oocyst burden (Fig. 1B). This decoupling of infection prevalence \nfrom intensity suggests that factors beyond simple gametocyte presence influence oocyst load. \nNotably, gametocyte density positively correlated with the abundance of asexual blood stages \n(ABS) (Fig. 1C), and high infection loads were only observed when gametocyte counts exceeded \n~700/µL of blood (Fig. 1D), suggesting a possible threshold effect in infectivity. \nTo capture key developmental stages of P. vivax within the mosquito, midguts were sampled at six \ntimepoints post blood feeding: 1h (hour), 6h, 18h, 22h, 26h, and 7d (days) pbf (Fig. 1E). Based on \nestablished timelines of parasite development, we expect that at 1h pbf the midguts would contain \nABS, gametocytes, and newly emerged gametes. By 6h pbf, zygote presence is anticipated, while \nookinetes are expected to appear between 18h and 26h pbf. By 7d pbf, only mature oocysts are \nexpected to remain, following the digestion of the blood bolus and progression of parasite \ndevelopment. \nTranscriptional profiling highlights infection-specific gene expression programmes \nA total of 72 RNA-seq libraries were generated from A. darlingi midguts across the six timepoints \npbf, spanning high and low P. vivax infection groups and their respective controls. Sequencing \nyielded an average of 15.5 million reads per library for low infections and 34.5 million reads for \nhigh infections. About 60% of reads mapped to the A. darlingi reference genome AdarC3, \nidentifying transcripts corresponding to 10,895 annotated genes: 10,423 with A. gambiae orthologs \nand 472 with no identifiable A. gambiae orthologs. \nUnsupervised hierarchical clustering based on global transcriptome similarity revealed consistent \ntemporal organization among samples (Fig. 2A). For high infection groups, samples from 18h, 22h, \nand 26h pbf clustered tightly. These timepoints were therefore grouped together and referred to as \n“1d pbf” in subsequent analyses. Notably, within this 1d pbf group, I and C samples separated into \ndistinct clusters, indicating that P. vivax infection elicits specific transcriptional responses. A \nsimilar pattern was observed in low infection groups, although the separation between I and C \nsamples at 1d pbf was less pronounced. This suggests that the mosquito transcriptional response is \ninfection-dependent and potentially modulated by the magnitude or quality of the infecting parasite \npopulation. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 5 \nDifferential gene expression analysis revealed 479 (Dataset 2) and 634 (Dataset 3) differentially \nexpressed genes (DEGs) in high and low infection groups, respectively (Fig. 2B). The majority of \nDEGs were detected at 1d pbf in both cases (359 in high and 463 in low infections), coinciding with \nthe critical window of midgut traversal. Whereas DEGs at this timepoint were moderately balanced \nbetween upregulated and downregulated genes in high infection groups, 96.3% of DEGs in low \ninfection groups were upregulated, indicating a strong transcriptional response that may play a role \nin limiting infection. Earlier in the infection course, high infection groups exhibited a muted \nresponse, with only 11 and 41 upregulated DEGs at 1h and 6h pbf, respectively, compared to 111 \nand 178 in the corresponding low infection groups. \nThis early and sustained activation in low infection groups suggests a more rapid and effective \nresponse that may hinder parasite infection. In support of this, Venn diagram analysis of DEGs \nacross timepoints revealed a substantial overlap in the low infection group between genes \nupregulated at 1h, 6h, and 1d pbf (Fig. 2C), consistent with a primed and coordinated early \nresponse. \nMetabolic deprivation and redox regulation define divergent infection trajectories \nA comparative analysis of gene ontology (GO) enrichment in highly versus lowly infected mosquito \ngroups revealed clear differences in the host response, reflecting divergent physiological \nengagements. A total of 45 significantly enriched GO terms were identified in the high and 35 in the \nlow infection groups, spanning molecular function (MF), biological process (BP), and cellular \ncomponent (CC) categories (Fig. 2D and Dataset 4). Notably, 11 GO terms were shared between \nthe two groups, encompassing terms such as transporter activity and alpha-amino acid metabolism. \nThis overlap likely represents core physiological responses to infection, involving solute exchange \nand protein turnover, irrespective of infection intensity. \nThe high infection group was uniquely enriched for GO terms related to proteolysis and its \nregulation, as well as pathways linked to oocyte development. As there is significant overlap \nbetween genes involved in immune responses and proteolysis, such as clip-domain serine proteases \nand their homologs, this pattern could suggest a balanced response to the parasite challenge coupled \nwith a trade-off in reproductive investment (Fig. 2D). In contrast, mosquitoes in the low infection \ngroup exhibited a distinct response dominated by genes involved in alpha and aromatic amino acid \ncatabolism, including phenylalanine and tyrosine, as well as purine biosynthesis. These metabolic \nprocesses are particularly relevant given that both aromatic amino acids and purines are essential for \nPlasmodium, which lacks the biosynthetic capacity to produce them [27, 28]. This suggests that the \nmosquito may limit parasite development through metabolic deprivation or sequestration of these \ncritical nutrients. \nAdditionally, the enrichment of monooxygenase and oxidoreductase activity in the low infection \ngroup suggests a shift toward redox regulation and detoxification pathways (Fig. 2D), possibly \nlinked to elevated mitochondrial activity or microbial dysbiosis. Together, these findings support \nthe view that midgut transcriptional responses related to the mosquito metabolic state may actively \ndetermine the trajectory of P. vivax infection. \nTime-course analysis identifies early divergence in infection trajectories \nTo further dissect the mechanisms underlying the divergent P. vivax infection outcomes, we \nextended our analysis beyond GO enrichment by examining the temporal dynamics of DEGs at \neach of the four timepoints pbf (Dataset 5). This high-resolution analysis allowed us to pinpoint \nspecific transcriptional changes that may influence the final infection outcome. \nInterestingly, temporal correlation analysis between high and low infection groups revealed striking \ndivergence at the earliest (1h) and most critical (1d) timepoints (Fig. 3). At 1h pbf, responses were \nlargely uncorrelated, reflecting highly distinct early physiological states between mosquitoes that \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 6 \nwould go on to develop either high or low parasite burdens (Pearson r = 0.03, R² = 0.0009, p = 0.67). \nA similar lack of correlation was observed at 1d pbf, coinciding with the peak of microbial load and \nactivity in the midgut and the critical window of establishment of infection, suggesting that \ntranscriptional trajectories diverge most sharply at times when parasite establishment is most \nvulnerable to vector responses (Pearson r = –0.07, R² = 0.0049, p = 0.063). In contrast, at 6h pbf, \ngene expression patterns between high and low infection groups showed greater convergence, \nsuggesting transient homeostatic realignment after the initial response to blood feeding. By 7d pbf, \nwhen the parasites have matured into oocysts and midgut activity has stabilised, expression profiles \nagain became more correlated, consistent with a return to physiological equilibrium. \nThese findings highlight the 1h and 1d pbf in the low infection group as the most informative \ntimepoints for understanding the molecular drivers of infection outcome. The early divergence at 1h \nlikely reflects immediate transcriptional reprogramming triggered by bloodmeal composition and, \npossibly, microbiota response, while the divergence at 1d pbf coincides with the outcome-\ndetermining phase of parasite midgut traversal and immune engagement. We therefore focused \nsubsequent analyses on these two key timepoints to uncover mechanisms that underpin the \nestablishment or restriction of infection. \nLow infection is marked by early redox activation and metabolic reprogramming \nAt 1h pbf, mosquitoes in the low infection group displayed a striking upregulation of genes \nassociated with detoxification and redox regulation (Fig. 3A). Multiple cytochrome P450s (e.g., \nCYP6P5, CYP6P3, CYP4AA1, CYP9b2) and glutathione S-transferases (e.g., GSTE2 and GSTE7) \nwere upregulated, along with glutaminase, an enzyme that liberates glutamate and urea from \nglutamine. These changes indicate early oxidative stress responses and a detoxification state in the \nmidgut. \nMetabolic interference also emerged as a key theme. We observed upregulation of allantoicase \n(ALLC), a purine-degrading enzyme, along with concurrent downregulation of multiple genes \ninvolved in purine salvage and biosynthesis, such as purine nucleoside phosphorylase (PNP), \nphenylalanine-4-hydroxylase (PAH), 4-hydroxyphenylpyruvate dioxygenase (HPPD), and a \nbifunctional purine biosynthesis protein (BPBP). These shifts suggest programmed depletion of \npurine and nucleoside resources, which are essential for parasite DNA and RNA synthesis, \nespecially given parasite reliance on host purines [28]. \nFurther metabolic reprogramming was evident through the modulation of AMP-dependent CoA \nligases, glucosyl/glucuronosyl transferases (UGT), and various solute transporters. Transferrin was \nstrongly downregulated, hinting at reduced iron availability, while carbonic anhydrase suppression \nsuggests altered pH regulation and gas exchange. Collectively, these responses point to a \nnutritionally restrictive and immunologically hostile midgut environment. \nInterestingly, classical immune pathway components were not strongly activated at this stage, \nexcept for the fibrinogen-like recognition protein FBN9, which was significantly upregulated in low \ninfection mosquitoes but downregulated in the high infection group. Given the known roles of \nFBN9 in both microbial recognition and immune modulation [29], this suggests early microbiota-\ndriven immune priming may be linked to the low infection phenotype. \nIn contrast, the high infection group exhibited a muted transcriptional response at 1 h pbf, with only \na small number of upregulated genes and minimal immune or metabolic reprogramming. This lack \nof early transcriptional activation may contribute to the permissive environment that allows P. vivax \nto proceed through midgut traversal and establish oocysts. \nTogether, these early timepoint comparisons provide mechanistic insight into how differential \nregulation of redox homeostasis, purine metabolism, and microbial recognition may drive the \ndivergent P. vivax infection outcomes observed in A. darlingi. The early and sustained activation of \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 7 \nthese processes in low infection groups supports a model in which rapid physiological adaptation \ndisrupts parasite development before midgut invasion is complete. \nAt 6h pbf, a moderate positive correlation in gene expression was observed between high and low \ninfection groups (Pearson r = 0.33, R² = 0.11, p < 0.0001), indicating partial convergence of \ntranscriptional responses (Fig. 3B). This alignment may reflect a transient state of physiological \nhomeostasis following the initial perturbation induced by blood ingestion. However, consistent with \nthe patterns observed at 1h pbf, genes associated with redox regulation, including two GSTs and a \ncytochrome P450, remained strongly upregulated in the low infection group. In addition, \nhomogentisate 1,2-dioxygenase (HGD), an enzyme involved in the degradation of aromatic amino \nacids tyrosine and phenylalanine, was also strongly upregulated. Given the importance of these \namino acids for parasite survival and their role in the synthesis of folate and para-aminobenzoic \nacid (PABA) that are essential in nucleotide biosynthesis [30], HGD upregulation may reflect a \nhost-driven strategy to deprive the parasite of essential metabolites. Additional genes involved in \npurine metabolism, such as amidophosphoribosyltransferase, were also upregulated though more \nmoderately, further supporting the hypothesis that nutritional deprivation, particularly of purines \nand nucleotides, may underpin the low infection phenotype. \nAmong the most strongly upregulated genes (32-38-fold change) were two previously unannotated \ngenes (ADAC003727, ADAC003402) involved in chitin metabolism. These likely relate to \nperitrophic matrix dynamics, possibly in response to microbial dysbiosis [31]. The strong \nupregulation of ornithine decarboxylase (ODC), a key enzyme in the biosynthesis of polyamines \nthat are essential for the DNA/RNA stability, is also worth noting, as this may suggest disruption of \ngut epithelial homeostasis due to microbiota dysbiosis [32]. \nIntriguingly, several neuronal signalling components, including three synaptic vesicle protein (SVP) \ngenes and a ligand-gated ion channel (LGIC), were also among the top DEGs in the low infection \ngroup. SVPs are involved in neurotransmitter storage and release, while LGICs mediate neuronal \nsignal transduction. Their coordinated induction raises the possibility of a gut-neural axis response \nmodulating midgut physiology or homeostasis in reaction to the infectious bloodmeal [33], \npotentially contributing to a more restrictive environment for parasite development. \nSustained transcriptional activity and epithelial modulation during midgut invasion \nThe 1d (18-26h) pbf timepoint marked the peak of transcriptional activity in both high and low \ninfection groups, as well as the point of greatest divergence between them (Fig. 3C). No correlation \nwas observed in gene expression profiles (Pearson r = -0.07, R2 = 0.0049, p = 0.063), indicating that \ndistinct transcriptional programs were engaged as the parasite attempted midgut invasion. This \nwindow corresponds with the peak of microbial proliferation and the critical stage of ookinete \ntraversal through the midgut epithelium, when vector responses can decisively influence infection \nsuccess [34, 35]. While no single dominant transcriptional pathway stood out, the patterns observed \nat earlier timepoints were broadly reinforced. \nIn the low infection group, genes associated with detoxification and cellular stress, including UGT, \nGSTE7 and ODC, remained strongly upregulated. This sustained stress response likely maintains an \nunfavourable physiological state for parasite development. Consistent with earlier stages, multiple \nenzymes linked to purine and aromatic amino acid degradation were also strongly upregulated. \nThese included ALLC, HGD, tyrosine aminotransferase (TAT), and HPPD, enzymes that \nsequentially degrade tyrosine and phenylalanine. Notably, TAT and HPPD were concurrently \ndownregulated in the high infection group, suggesting that modulation of this pathway may be a key \ndeterminant of infection outcome. Suppression of tyrosine degradation in high infection mosquitoes \nmay inadvertently preserve critical precursors for parasite metabolism, thus facilitating parasite \ndevelopment. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 8 \nAmong the top DEGs in the low infection group was cathepsin B (CATHB), a lysosomal cysteine \nprotease involved in protein degradation, apoptosis, and immune modulation. Its strong \nupregulation may reflect midgut epithelial turnover in response to microbial imbalance and \ninfection or physical damage during parasite invasion. Another highly upregulated gene was \ncuticular protein 75 (CPR75), which is structurally associated with soft, flexible cuticle membranes \n[36]. While not previously linked to the peritrophic matrix, its expression in this context raises the \npossibility of a role in midgut structural remodelling. Finally, we observed a surprising \ndownregulation of two cecropins in the low infection group. Given prior studies linking \nantimicrobial peptide expression and peritrophic matrix dynamics to microbiota composition [31], \nthis pattern may signal a homeostatic shift in microbial control rather than direct immunological \nsuppression. \nLate-stage alignment of responses and residual expression differences \nBy 7d pbf, gene expression patterns between high and low infection groups showed strong \nconvergence (Pearson r = 0.45, R2 = 0.20, p < 0.0001; Fig. 3D), suggesting a return to shared \nphysiological processes as the mosquito recovers from early infection-driven perturbations. This \nalignment likely reflects the completion of midgut remodelling and digestion, as well as the reduced \nimmune and metabolic stress following parasite traversal. However, some transcriptional \ndifferences persist and may be linked to the continued presence and growth of oocysts particularly \nin the high infection group, where parasite burden and associated nutrient demands are greater. \nInterestingly, several genes that had been downregulated in the early response phases of low \ninfection mosquitoes, including transferrin and the two cecropins, were now upregulated. This late-\nstage induction of transferrin may reflect either recovery of iron homeostasis or continued attempts \nto restrict parasite access to this essential micronutrient [37]. Similarly, the delayed upregulation of \ncecropins may indicate a secondary wave of antimicrobial activity, possibly in response to shifts in \nmicrobial community structure or earlier epithelial damage. \nMicrobiota composition and structure correlate with infection phenotype \nTo understand how the midgut microbiota may contribute to the observed P. vivax infection \nphenotypes, we examined bacterial load, diversity, and taxonomic composition in A. darlingi \nmidguts across the two phenotypic groups and timepoints. Particular attention was paid to the low \ninfection group, where early transcriptional activation of redox regulation, aromatic amino acid \ndegradation, purine depletion, and epithelial remodelling were observed. \nTotal bacterial load, as quantified by 16S rRNA qRT-PCR, remained relatively stable across \ntimepoints in low infection mosquitoes (Fig. 4A, lower panel). This contrasted with high infection \ngroups, where a significant increase was detected at 1d pbf. The absence of bacterial expansion in \nthe low infection phenotype suggests that early transcriptional responses, particularly those \ninvolving oxidative stress and metabolic restriction, may restrict microbial growth and maintain \nepithelial stability during a critical phase of parasite development. \nSupporting this, alpha-diversity analysis revealed a significant reduction in microbiota diversity in \nthe low infection group at 1d pbf (Fig. 4B, right panel), coinciding with peak expression of \ndetoxification, redox, and nutrient-deprivation genes. The loss of bacterial diversity may reflect \nmicrobiota selection driven by redox or metabolic stress rather than canonical immune activity, \nconsistent with the observed downregulation of antimicrobial peptides like cecropins. \nBeta-diversity (Bray-Curtis PCA) analysis revealed consistent microbiota differences between high \nand low infection groups across all comparisons, including controls, suggesting that infection \nphenotype was associated with persistent microbiota structuring, independent of parasite exposure \n(Fig. 4C). Notably, groups producing low infection (P1, P3, P4) clustered separately from those \nproducing high infections (P2, P5, P6), indicating that differences in host bloodmeal composition or \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 9 \npre-existing microbiota may act as upstream determinants of mosquito-microbiota-parasite \ninteractions and, ultimately, infection outcome. \nSuppressive bacterial taxa and redox-linked dysbiosis in low infection mosquitoes \nTaxonomic profiling at the class and order levels revealed striking differences in bacterial \ncommunity composition between phenotypes (Fig. 5A-C). At 1d pbf, low infection mosquitoes \nwere significantly enriched for Enterobacteriales and Pseudomonadales, two Gammaproteobacterial \norders previously associated with antiplasmodial activity [38, 39]. In particular, Pantoea, Serratia \nand Thorsellia (Enterobacteriales), as well as Pseudomonas were more abundant in low infection \nmosquitoes (Fig. 5B). These genera are frequently found in wild Anopheles populations and include \nspecies capable of generating reactive oxygen species or producing metabolites that interfere with \nPlasmodium development, potentially linking them to the observed redox activation in low \ninfection mosquitoes [16, 34]. \nIn contrast, Flavobacteriales, particularly Elizabethkingia and Chryseobacterium, dominated high \ninfection mosquitoes but were depleted in the low infection group. These bacteria have been \npreviously associated with A. gambiae immune suppression and Plasmodium infection (Akhouayri \net al., 2013; Gimonneau et al., 2014), further supporting the idea that microbiota composition may \nshape midgut barrier integrity and redox tone. In particular, Elizabethkingia anophelis has been \nreported to persistently colonize mosquitoes and produce factors that modulate redox balance and \nvector immunity and protect Plasmodium from human complement-mediated killing [40-42]. \nTaken together, these findings suggest that selective enrichment of certain bacterial taxa and \nreduced bacterial diversity in the low infection group reflect a microbially moderated metabolic \nstate that is less permissive to parasite development. Rather than direct immune clearance, the low \ninfection phenotype appears to arise from early microbiota-mediated modulation of midgut \nhomeostasis, redox dynamics, and nutrient availability, conditions that challenge parasite survival \nduring midgut traversal. \nConclusion \nThis study provides a systems-level insight into the molecular and microbial factors shaping P. \nvivax infection outcomes in the major neotropical vector A. darlingi. By integrating time-resolved \ntranscriptomic and microbiota data from experimentally infected mosquitoes, we identified two \nsharply divergent phenotypes, high and low oocyst burdens, which emerged despite similar \ninfection prevalence. Importantly, these outcomes were not primarily governed by canonical \nimmune pathways but instead arose from early and sustained interactions between midgut \nmetabolism, microbial ecology, and bloodmeal composition. \nMosquitoes that developed low infection burdens mounted a rapid and sustained transcriptional \nresponse beginning as early as 1 hour post blood feeding, characterised by detoxification, redox \nregulation, aromatic amino acid catabolism, and purine depletion, likely coordinated through \nneurophysiological cues. These changes collectively create a nutritionally restrictive and hostile \nmidgut environment for parasite development. Concurrently, microbial diversity contracted and \nbecame enriched in taxa such as Pantoea, Serratia, and Pseudomonas, genera previously associated \nwith ROS production and anti-Plasmodium activity [43, 44]. In contrast, high-infection mosquitoes \nexhibited subdued metabolic reprogramming and expansion of Flavobacteriales such as \nElizabethkingia and Chryseobacterium, genera linked to immune suppression and compromised \nepithelial integrity. These microbial shifts, along with the distinct metabolic tone of the midgut, \nsuggest that early microbiota-mediated modulation of physiological state is a key determinant of \nvector competence. Our findings are consistent with and extend previous work in Anopheles \nstephensi, where P. vivax infection outcomes were shown to depend on specific bacterial taxa and \nmicrobiota-driven immune activation [45]. However, unlike A. stephensi, in which parasite \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 10 \nclearance was primarily linked to canonical immune pathways, our results in A. darlingi suggest a \ndistinct mode of vector competence regulation dominated by early metabolic and redox \nprogramming. This contrast underscores species-specific strategies of P. vivax–vector co-adaptation \nand highlights the importance of studying ecologically relevant vectors in different malaria-endemic \nsettings. \nAll low infection phenotypes originated from patients with the lowest gametocyte densities. While \nthis may at first suggest a quantitative threshold for infectivity, our findings point to a more \ncomplex model in which qualitative differences in blood composition, linked to host and parasite-\nderived factors, prime the mosquito midgut either for resistance or susceptibility. Such priming may \nbe mediated by labile host metabolites or parasite signals that influence microbiota structure and \nmidgut redox-metabolic programming, enabling the mosquito to sense and respond to the \ntransmission potential of the bloodmeal. \nTogether, our findings redefine A. darlingi vector competence as a dynamic, emergent property of \nearly physiological and microbial interactions rather than a fixed immune trait. This perspective opens \nnew avenues for transmission -blocking interventions based on modulating midgut metabolism or \npromoting protective microbial communities. More broadly, it highlights the need to study vector –\nparasite–microbiota interactions in ecologically and regionally relevant systems, where distinct \nevolutionary and epidemiological forces shape the biology of malaria transmission. \nMethods \nMosquito rearing \nA. darlingi mosquitoes were obtained from a laboratory colony established at Universidad Peruana \nCayetano Heredia in Iquitos, Peru [22]. The colony was maintained under standard insectary \nconditions at 27°C, 80% relative humidity, and a 12:12 h light/dark photoperiod. Adult mosquitoes \nwere housed in mesh cages and provided with a 10% sucrose solution ad libitum. Female \nmosquitoes aged 3-5 days were used for colony propagation or experimental infections with P. \nvivax. Larvae were reared in mineral water at 28-30°C and fed Nutrafin fish food: L1/L2 instars \nwere fed twice daily, while L3/L4 instars received three feedings per day. \nParasitological survey  \nPeripheral blood samples were collected from six symptomatic malaria patients (P1-P6), aged 19 to \n47 years, who presented at a local health facility in Iquitos, Peru. Thick blood smears were \nprepared, stained with 10% Giemsa, and examined by light microscopy for parasite detection and \nquantification. All patients were found to be infected exclusively with P. vivax. Parasite densities, \nincluding both sexual and asexual stages, were estimated using a standard assumption of 6,000 \nwhite blood cells (WBC) per μL of blood, following national guidelines established by the Peruvian \nMinistry of Health. Parasite counts were based on 200 WBCs, or 500 WBCs when fewer than 10 \nparasites were observed. Blood collection was performed with informed consent from all \nparticipants. \nGametocytaemic blood processing and mosquito infections \nBlood samples from six P. vivax-infected patients (P1-P6), confirmed to carry gametocytes, were \nprocessed and used to infect A. darlingi mosquitoes via DMFAs as described previously for A. \ngambiae [13, 26]. Briefly, patient serum was replaced with an equal volume of non-immune human \nserum to remove potential transmission-blocking factors. Each processed sample was then split into \ntwo aliquots: one was used directly for mosquito infection (viable gametocytes), and the other was \nheat-inactivated by incubation at 42°C for 12 minutes to serve as a control. Adult female \nmosquitoes, starved for 12 hours prior to feeding, were allowed to feed for 20 minutes on either the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 11 \nviable or heat-inactivated bloodmeal using glass membrane feeders. This resulted in matched \ninfected (I) and control (C) mosquito groups for each of the six blood donors. Each infection event \nusing a different donor blood was treated as an independent biological replicate (six replicates in \ntotal). \nMosquito dissections and RNA isolation \nA. darlingi mosquitoes from both P. vivax-infected and control (heat-inactivated) groups were cold-\nanaesthetised, and their midguts were dissected in phosphate-buffered saline (PBS) under a \nstereomicroscope at six defined timepoints pbf: 1h, 6h, 18h, 22h, 26h, and 7d. For assessment of \ninfection intensity, an additional 10-15 midguts were dissected at 7-8 days pbf, stained with 2% \nmercurochrome, and examined under a light microscope for oocyst enumeration. For transcriptomic \nanalysis, dissected midguts were immediately transferred into microcentrifuge tubes containing \nTrizol® reagent (Invitrogen) in pools of 30-45 midguts per timepoint and stored at -80°C. In total, \n72 midgut pools were collected, representing six biological replicates (P1-P6) across two \nexperimental conditions (infected and control) and six timepoints. Total RNA was extracted \nfollowing the manufacturer’s instructions for Trizol reagent and quantified using a Qubit \nfluorometer (Invitrogen). \nRNAseq library preparation and sequencing \nStrand-specific RNA-Seq libraries were prepared individually for each of the 72 midgut pools \ndescribed above. For each sample, 0.8-2 μg of total RNA was used to generate libraries using the \nSureSelect Strand-Specific RNA Library Prep Kit for Illumina Multiplexed Sequencing (Agilent). \nLibrary quality was assessed using the QIAxcel Advanced System (QIAGEN) with the QIAxcel \nDNA High Resolution Kit (1200), and library concentrations were quantified via real-time PCR \nusing the KAPA Library Quantification Kit (Roche). Final libraries were normalised to 2 nM and \npooled for multiplexed paired-end sequencing on an Illumina NextSeq 500 platform. Raw RNA-seq \ndata have been deposited in the European Nucleotide Archive (ENA) under the study title: RNA-\nSeq of A. darlingi infected or uninfected with P. vivax at various time points with high and low \ninfections; Study ID: PRJEB29445 (ERP111747); Release Date: 03-Jan-2023. \nRNAseq data processing and analysis \nThe quality of paired-end sequencing reads was assessed using FastQC, and adapters and low-\nquality bases were trimmed using Trimmomatic v0.32.3 [46]. Cleaned reads were aligned to the A. \ndarlingi genome assembly AdarC3 [47] using TopHat2 v2.0.14 with default settings [48]. Read \nquantification and transcript abundance were calculated as fragments per kilobase of exon per \nmillion mapped reads (FPKM) using Cufflinks v2.2.1 [49]. Differential gene expression analysis \nwas conducted with Cuffdiff v2.2.1 [50], comparing P. vivax-infected mosquito midguts to their \ncorresponding heat-inactivated controls at each dissection timepoint. To assess the consistency \namong replicates and overall transcriptional patterns, unsupervised hierarchical clustering was \nperformed using the ward.D2 method on global transcript profiles. Clustering and heatmap \nvisualization were performed with Cluster 3.0 and Java TreeView [51], respectively. Pearson \ncorrelation analyses of log2 fold-change profiles between high and low infection groups were \nperformed to evaluate the temporal similarity in transcriptional responses across timepoints. \nGene Ontology term assignment and enrichment \nGO terms were assigned to A. darlingi transcripts (AdarC3.5 gene set, release date: June 2017) by \northology-based mapping to the A. gambiae gene set (AgamP4.6) using BioMart. Functional \nenrichment analysis of differentially expressed genes (DEGs) was conducted using the PANTHER \nclassification system available through VectorBase. Enrichment was assessed across the three GO \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 12 \ncategories: Biological Process (BP), Molecular Function (MF), and Cellular Component (CC), with \nsignificance determined by Fisher’s exact test and false discovery rate (FDR) correction \n(Padj < 0.05). \nQuantification of A. darlingi gut microbiota by real-time qPCR  \nQuantitative real-time PCR (RT-qPCR) was used to assess bacterial load in the midguts of P. vivax-\ninfected and control A. darlingi mosquitoes by targeting the bacterial 16S ribosomal RNA gene \n(16S rRNA) as described [52]. Total RNA (1 µg) from each of the 72 midgut pools was treated with \nDNase I (Invitrogen™) at 37 °C for 2 hours and reverse-transcribed using the RevertAid First \nStrand cDNA Synthesis Kit (Thermo Scientific™), following the manufacturer’s protocol. RT-\nqPCR reactions were performed in triplicate using SYBR Green PCR Master Mix (Applied \nBiosystems™) in a final volume of 20 µL. Expression of the 16S rRNA gene was normalised \nagainst the A. darlingi reference gene Rp49 (ADAC007403) using the 2^–ΔCT method [53]. RT-\nqPCR primers for A. darlingi Rp49 gene expression analysis were designed using Primer 3 online \ntool and their efficiency was determined by RT-qPCR: Rp49-F, ACAGTACCTGATGCCGAACA; \nRp49-R: TTCTGCATCATCAGCACCTC  Statistical significance of microbiota load differences \nbetween infected and control groups at the 1d pbf timepoint of midgut invasion was assessed using \nthe Mann-Whitney U test. \nMultiplex bacterial 16S rRNA amplicon Illumina sequencing  \nThe gut bacterial communities of P. vivax-infected and control A. darlingi mosquitoes were profiled \nusing multiplex 16S rRNA amplicon sequencing as described [54]. Amplicon libraries targeting the \nV4 hypervariable region of the bacterial 16S rRNA gene were generated via PCR using cDNA \nsynthesised from previously DNase-treated midgut RNA as template. Each reaction used a single \nuniversal forward primer and a set of reverse primers tagged with unique barcodes (indexes) to \nenable sample multiplexing. PCRs were performed in triplicate per sample, and resulting amplicons \nwere pooled and purified using AMPure XP magnetic beads (Beckman Coulter). Purified libraries \nwere quantified with the KAPA Library Quantification Kit (Roche®), normalised to 2 nM, and \npooled equimolarly. Sequencing was performed using a V2 MiSeq Reagent Kit on the Illumina \nMiSeq platform. \n16S rRNA amplicon Illumina sequencing analysis \nRaw sequencing data were first processed using the MiSeq Reporter software (Illumina) to remove \nadapter sequences and generate demultiplexed FASTQ files. Subsequent analysis was carried out \nusing the Microbial Genomics Module of CLC Genomics Workbench v10.1.2 (QIAGEN), \nfollowing the manufacturer’s standard pipeline. Paired-end reads were merged using the following \nparameters: minimum score of 8, gap cost of 3, mismatch cost of 2, and no maximum unaligned \nbases. Merged reads were trimmed to 264 bp, and samples with fewer than 100 reads or below 50% \nof the median read count were excluded. Operational Taxonomic Unit (OTU) classification was \nperformed using the SILVA v119 16S rRNA reference database with a minimum sequence identity \nthreshold of 97%. Low-abundance OTUs representing <0.05% of total reads within a sample were \nfiltered out. Microbial diversity was assessed using both alpha and beta diversity metrics. Alpha \ndiversity was calculated using Simpson’s diversity index, which incorporates both OTU richness \nand evenness, placing greater weight on dominant taxa. Beta diversity was assessed using Bray-\nCurtis dissimilarity distances and visualised via principal coordinates analysis (PCoA) using the top \nthree principal components. Clustering patterns were statistically evaluated using PERMANOVA, \nbased on sample metadata. \nEthics statement \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 13 \nAll procedures involving A. darlingi infections with naturally circulating P. vivax gametocytes from \nhuman carriers were conducted at Universidad Peruana Cayetano Heredia under approved ethical \nguidelines. Ethical approval was granted by the institutional review board of Universidad Peruana \nCayetano Heredia (Protocol No. R157-13-14). Informed consent was obtained from all participating \nindividuals prior to blood collection, in accordance with national and institutional ethical standards. \nAcknowledgements \nWe thank Lutecio Torres, Gerson Guedez, Cristian Rodriguez, Juan Michi and Zaira Villa for their \nassistance with the DMFAs, and Daniel Lawson for help with genome alignments and de novo gene \nmodel generation. We are grateful to the six volunteer patients who donated blood for this study. \nThis work was supported by a Young Investigator Award to J.A.S.N. from the São Paulo Research \nFoundation (FAPESP; grant number 2013/11343-6), a mobility grant to J.A.S.N. and G.K.C. under \nthe FAPESP-Imperial College London cooperative agreement (grant number 2014/50454-0), and a \nSpecial Visiting Fellowship from the National Council for Scientific and Technological \nDevelopment (CNPq; grant number 401433/2014-5) to J.A.S.N. and G.K.C. under the Science \nwithout Borders Program. B.C.C. was supported by a CNPq postdoctoral fellowship (grant number \n154727/2016-4). G.C.K. and D.V. were also supported by a Wellcome Trust Investigator Award \n(grant number 107983/Z/15/Z) and a Medical Research Council (MRC) project grant \n(MR/T000929/1). C.T.R. and M.M. were supported by the National Institutes of Health-National \nInstitute of Allergy and Infectious Diseases (NIH-NIAID) ICEMR-Amazonia program \nU19AI089681. We are grateful to Jan E. Conn and Joseph M. Vinetz for facilitating the \ncollaboration with ICEMR-Amazonia and enabling this study. \nAuthor Contributions \nConceptualization: D.V., G.K.C. and J.A.S.N.; Methodology: M.M., D.V., G.K.C. and J.A.S.N.; \nFormal analysis: B.C.C., P.H.A.A., A.J., D.P.A., B.M, D.V., G.K.C. and J.A.S.N.; Investigation: \nB.C.C., K.V., P.H.A.A., B.T.N., D.P.A., D.V., G.K.C. and J.A.S.N.; Resources: M.M., D.V., \nG.K.C. and J.A.S.N.; Data Curation: A.J. and B.M.; Writing - Original Draft: B.C.C., D.V., G.K.C. \nand J.A.S.N.; Writing - Review & Editing: D.V. and G.K.C.; Visualization: G.K.C.; Supervision: \nD.V., G.K.C. and J.A.S.N.; Project administration: M.M., D.V., G.K.C. and J.A.S.N.; Funding \nacquisition: D.V., G.K.C. and J.A.S.N. \nCompeting Interests \nThe authors declare no competing interests. \nMaterials & Correspondence \nAll correspondence and materials requests should be addressed to G.K.C. \n(g.christophides@imperial.ac.uk) or J.A.S.N. (jsouzaneto@vet.k-state.edu) or D.V. \n(d.vlachou@imperial.ac.uk) \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 14 \nReferences \n1. World malaria report 2024: Geneva: World Health Organization. 2024. \n2. Laporta GZ, Grillet ME, Rodovalho SR, Massad E, Sallum MAM. Reaching the malaria \nelimination goal in Brazil: a spatial analysis and time -series study. Infect Dis Poverty. \n2022;11(1):39. Epub 20220405. doi: 10.1186/s40249-022-00945-5. PubMed PMID: 35382896; \nPubMed Central PMCID: PMCPMC8981179. \n3. Araujo MDS, Andrade AO, Dos Santos NAC, Castro RB, Pereira DB, Rodrigues MMS, et al. \nFirst Observation of Experimental Plasmodium vivax Infection of Three Malaria Vectors from \nthe Brazilian Amazon. Vector Borne Zoonotic Dis. 2020;20(7):517 -23. Epub 202004 07. doi: \n10.1089/vbz.2019.2527. PubMed PMID: 32255759. \n4. Carlos BC, Rona LDP, Christophides GK, Souza-Neto JA. A comprehensive analysis of malaria \ntransmission in Brazil. Pathog Glob Health. 2019;113(1):1 -13. Epub 20190304. doi: \n10.1080/20477724.2019.1581463. PubMed PMID: 30829565; PubMed Central PMCID: \nPMCPMC6425916. \n5. Chu VM, Sallum MAM, Moore TE, Emerson KJ, Schlichting CD, Conn JE. Evidence for family-\nlevel variation of phenotypic traits in response to temperature of Brazilian Nyssorhynchus \ndarlingi. Parasit Vectors. 2020;13(1):55. Epub 20200210. doi: 10.1186/s13071 -020-3924-7. \nPubMed PMID: 32041663; PubMed Central PMCID: PMCPMC7011564. \n6. Graumans W, Jacobs E, Bousema T, Sinnis P. When Is a Plasmodium -Infected Mosquito an \nInfectious Mosquito? Trends Parasitol. 2020;36(8):705 -16. Epub 20200701. doi: \n10.1016/j.pt.2020.05.011. PubMed PMID: 32620501; PubMed Central PMCID: \nPMCPMC7386819. \n7. Povelones M, Osta MA, Christophides GK. Chapter Eight - The Complement System of Malaria \nVector Mosquitoes. In: Raikhel AS, editor. Advances in Insect Physiology. Progress in Mosquito \nResearch: Elsevier; 2016. p. 223-42. \n8. Smith RC, Vega-Rodriguez J, Jacobs-Lorena M. The Plasmodium bottleneck: malaria parasite \nlosses in the mosquito vector. Mem Inst Oswaldo Cruz. 2014;109(5):644-61. doi: 10.1590/0074-\n0276130597. PubMed PMID: 25185005; PubMed Central PMCID: PMCPMC4156458. \n9. Cai JA, Christophides GK. Immune interactions between mosquitoes and microbes during \nmidgut colonization. Curr Opin Insect Sci. 2024;63:101195. Epub 20240327. doi: \n10.1016/j.cois.2024.101195. PubMed PMID: 38552792. \n10. Gendrin M, Favia G, Herren JK. Editorial: Microbiota: A Consequential Third Wheel in the \nMosquito-Pathogen Relationship. Front Microbiol. 2021;12:811880. Epub 20220118. doi: \n10.3389/fmicb.2021.811880. PubMed PMID: 35116014; PubMed Central PMCID: \nPMCPMC8804289. \n11. Saab SA, Cardoso-Jaime V, Kefi M, Dimopoulos G. Advances in the dissection of Anopheles -\nPlasmodium interactions. PLoS Pathog. 2025;21(3):e1012965. Epub 20250331. doi: \n10.1371/journal.ppat.1012965. PubMed PMID: 40163471; PubMed Central PMCID: \nPMCPMC11957333. \n12. Dimopoulos G, Christophides GK, Meister S, Schultz J, White KP, Barillas -Mury C, et al. \nGenome expression analysis of Anopheles gambiae: responses to injury, bacterial challenge, and \nmalaria infection. Proc Natl Acad Sci U S A. 2002;99(13):8814 -9. Epub 20 020619. doi: \n10.1073/pnas.092274999. PubMed PMID: 12077297; PubMed Central PMCID: \nPMCPMC124381. \n13. Mendes AM, Awono -Ambene PH, Nsango SE, Cohuet A, Fontenille D, Kafatos FC, et al. \nInfection intensity-dependent responses of Anopheles gambiae to the African malaria parasite \nPlasmodium falciparum. Infect Immun. 2011;79(11):4708 -15. Epub 20110815. doi: \n10.1128/IAI.05647-11. PubMed PMID: 21844236; PubMed Central PMCID: \nPMCPMC3257935. \n14. Vlachou D, Schlegelmilch T, Christophides GK, Kafatos FC. Functional genomic analysis of \nmidgut epithelial responses in Anopheles during Plasmodium invasion. Curr Biol. \n2005;15(13):1185-95. doi: 10.1016/j.cub.2005.06.044. PubMed PMID: 16005290. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 15 \n15. Bahia AC, Dong Y, Blumberg BJ, Mlambo G, Tripathi A, BenMarzouk -Hidalgo OJ, et al. \nExploring Anopheles gut bacteria for Plasmodium blocking activity. Environ Microbiol. \n2014;16(9):2980-94. Epub 20140224. doi: 10.1111/1462 -2920.12381. PubMed PMID: \n24428613; PubMed Central PMCID: PMCPMC4099322. \n16. Dong Y, Manfredini F, Dimopoulos G. Implication of the mosquito midgut microbiota in the \ndefense against malaria parasites. PLoS Pathog. 2009;5(5):e1000423. Epub 20090508. doi: \n10.1371/journal.ppat.1000423. PubMed PMID: 19424427; PubMed Central PMCID: \nPMCPMC2673032. \n17. Meister S, Agianian B, Turlure F, Relogio A, Morlais I, Kafatos FC, et al. Anopheles gambiae \nPGRPLC-mediated defense against bacteria modulates infections with malaria parasites. PLoS \nPathog. 2009;5(8):e1000542. Epub 20090807. doi: 10.1371/journal.ppat.10 00542. PubMed \nPMID: 19662170; PubMed Central PMCID: PMCPMC2715215. \n18. Saraiva RG, Kang S, Simoes ML, Anglero -Rodriguez YI, Dimopoulos G. Mosquito gut \nantiparasitic and antiviral immunity. Dev Comp Immunol. 2016;64:53-64. Epub 20160129. doi: \n10.1016/j.dci.2016.01.015. PubMed PMID: 26827888. \n19. Marinotti O, Cerqueira GC, de Almeida LG, Ferro MI, Loreto EL, Zaha A, et al. The genome of \nAnopheles darlingi, the main neotropical malaria vector. Nucleic Acids Res. 2013;41(15):7387-\n400. Epub 20130612. doi: 10.1093/nar/gkt484. PubMed PMID: 23761445; Pu bMed Central \nPMCID: PMCPMC3753621. \n20. Villarreal-Trevino C, Vasquez GM, Lopez -Sifuentes VM, Escobedo -Vargas K, Huayanay -\nRepetto A, Linton YM, et al. Establishment of a free-mating, long-standing and highly productive \nlaboratory colony of Anopheles darlingi from the Peruvian Amazon. Malar J. 2015;14:227. Epub \n20150530. doi: 10.1186/s12936 -015-0733-0. PubMed PMID: 26024853; PubMed Central \nPMCID: PMCPMC4465318. \n21. Puchot N, Lecoq MT, Carinci R, Duchemin JB, Gendrin M, Bourgouin C. Establishment of a \ncolony of Anopheles darlingi from French Guiana for vector competence studies on malaria \ntransmission. Front Trop Dis. 2022;3. doi: https://doi.org/10.3389/fitd.2022.949300. \n22. Moreno M, Tong C, Guzman M, Chuquiyauri R, Llanos-Cuentas A, Rodriguez H, et al. Infection \nof laboratory-colonized Anopheles darlingi mosquitoes by Plasmodium vivax. Am J Trop Med \nHyg. 2014;90(4):612 -6. Epub 20140217. doi: 10.4269/ajtmh.13 -0708. PubMed PM ID: \n24534811; PubMed Central PMCID: PMCPMC3973502. \n23. Araujo MDS, Andrade AO, Santos N, Pereira DB, Costa GDS, Paulo PFM, et al. Brazil's first \nfree-mating laboratory colony of Nyssorhynchus darlingi. Rev Soc Bras Med Trop. \n2019;52:e20190159. Epub 20190718. doi: 10.1590/0037 -8682-0159-2019. PubMed PMID: \n31340377. \n24. Gunalan K, Rowley EH, Miller LH. A Way Forward for Culturing Plasmodium vivax. Trends \nParasitol. 2020;36(6):512 -9. Epub 20200429. doi: 10.1016/j.pt.2020.04.002. PubMed PMID: \n32360314; PubMed Central PMCID: PMCPMC7301965. \n25. Santos NAC, Andrade AO, Santos TC, Martinez LN, Ferreira AS, Bastos AS, et al. Evaluation \nof sustainable susceptibility to Plasmodium vivax infection among colonized Anopheles darlingi \nand Anopheles deaneorum. Malar J. 2022;21(1):163. Epub 20220603. doi: 10.1186/s12936-022-\n04204-8. PubMed PMID: 35658964; PubMed Central PMCID: PMCPMC9164182. \n26. Mendes AM, Schlegelmilch T, Cohuet A, Awono -Ambene P, De Iorio M, Fontenille D, et al. \nConserved mosquito/parasite interactions affect development of Plasmodium falciparum in \nAfrica. PLoS Pathog. 2008;4(5):e1000069. Epub 20080516. doi: 10.1371/journal.ppat.1000069. \nPubMed PMID: 18483558; PubMed Central PMCID: PMCPMC2373770. \n27. Krishnan A, Soldati-Favre D. Amino Acid Metabolism in Apicomplexan Parasites. Metabolites. \n2021;11(2). Epub 20210120. doi: 10.3390/metabo11020061. PubMed PMID: 33498308; \nPubMed Central PMCID: PMCPMC7909243. \n28. Cheviet T, Lefebvre -Tournier I, Wein S, Peyrottes S. Plasmodium Purine Metabolism and Its \nInhibition by Nucleoside and Nucleotide Analogues. J Med Chem. 2019;62(18):8365 -91. Epub \n20190425. doi: 10.1021/acs.jmedchem.9b00182. PubMed PMID: 30964283. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 16 \n29. Simoes ML, Dong Y, Hammond A, Hall A, Crisanti A, Nolan T, et al. The Anopheles FBN9 \nimmune factor mediates Plasmodium species -specific defense through transgenic fat body \nexpression. Dev Comp Immunol. 2017;67:257 -65. Epub 20160922. doi: \n10.1016/j.dci.2016.09.012. PubMed PMID: 27667688. \n30. Matz JM, Watanabe M, Falade M, Tohge T, Hoefgen R, Matuschewski K. Plasmodium Para -\nAminobenzoate Synthesis and Salvage Resolve Avoidance of Folate Competition and \nAdaptation to Host Diet. Cell Rep. 2019;26(2):356 -63 e4. doi: 10.1016/j.celrep.2018.12.062. \nPubMed PMID: 30625318. \n31. Rodgers FH, Gendrin M, Wyer CAS, Christophides GK. Microbiota-induced peritrophic matrix \nregulates midgut homeostasis and prevents systemic infection of malaria vector mosquitoes. \nPLoS Pathog. 2017;13(5):e1006391. Epub 20170517. doi: 10.1371/journal.ppat. 1006391. \nPubMed PMID: 28545061; PubMed Central PMCID: PMCPMC5448818. \n32. Anglero-Rodriguez YI, Blumberg BJ, Dong Y, Sandiford SL, Pike A, Clayton AM, et al. A \nnatural Anopheles -associated Penicillium chrysogenum enhances mosquito susceptibility to \nPlasmodium infection. Sci Rep. 2016;6:34084. Epub 20160928. doi: 10.1038/srep340 84. \nPubMed PMID: 27678168; PubMed Central PMCID: PMCPMC5039729. \n33. Das De T, Sharma P, Tevatiya S, Chauhan C, Kumari S, Yadav P, et al. Bidirectional \nMicrobiome-Gut-Brain-Axis Communication Influences Metabolic Switch -Associated \nResponses in the Mosquito Anopheles culicifacies. Cells. 2022;11(11). Epub 20220531. doi: \n10.3390/cells11111798. PubMed PMID: 35681493; PubMed Central PMCID: \nPMCPMC9180301. \n34. Gendrin M, Rodgers FH, Yerbanga RS, Ouedraogo JB, Basanez MG, Cohuet A, et al. Antibiotics \nin ingested human blood affect the mosquito microbiota and capacity to transmit malaria. Nat \nCommun. 2015;6:5921. Epub 20150106. doi: 10.1038/ncomms6921. PubMed PMID: 25562286; \nPubMed Central PMCID: PMCPMC4338536. \n35. Kumar S, Molina -Cruz A, Gupta L, Rodrigues J, Barillas -Mury C. A peroxidase/dual oxidase \nsystem modulates midgut epithelial immunity in Anopheles gambiae. Science. \n2010;327(5973):1644-8. Epub 20100311. doi: 10.1126/science.1184008. PubMed PMID: \n20223948; PubMed Central PMCID: PMCPMC3510679. \n36. Vannini L, Willis JH. Localization of RR -1 and RR -2 cuticular proteins within the cuticle of \nAnopheles gambiae. Arthropod Struct Dev. 2017;46(1):13 -29. Epub 20161030. doi: \n10.1016/j.asd.2016.10.002. PubMed PMID: 27717796; PubMed Central PMCID: \nPMCPMC5292290. \n37. Gorman MJ. Iron Homeostasis in Insects. Annu Rev Entomol. 2023;68:51 -67. Epub 20220928. \ndoi: 10.1146/annurev -ento-040622-092836. PubMed PMID: 36170642; PubMed Central \nPMCID: PMCPMC10829936. \n38. Cirimotich CM, Dong Y, Clayton AM, Sandiford SL, Souza-Neto JA, Mulenga M, et al. Natural \nmicrobe-mediated refractoriness to Plasmodium infection in Anopheles gambiae. Science. \n2011;332(6031):855-8. doi: 10.1126/science.1201618. PubMed PMID: 21566196; Pub Med \nCentral PMCID: PMCPMC4154605. \n39. Gonzalez-Ceron L, Santillan F, Rodriguez MH, Mendez D, Hernandez -Avila JE. Bacteria in \nmidguts of field -collected Anopheles albimanus block Plasmodium vivax sporogonic \ndevelopment. J Med Entomol. 2003;40(3):371 -4. doi: 10.1603/0022 -2585-40.3.371. PubMed \nPMID: 12943119. \n40. He B, Li M, Guo S, Zhu F, Jiao Z, Li J, et al. Host complement C3 promotes malaria transmission \nby killing symbiotic bacteria in the mosquito midgut. Proc Natl Acad Sci U S A. \n2025;122(22):e2424570122. Epub 20250528. doi: 10.1073/pnas.2424570122. PubMed P MID: \n40434644; PubMed Central PMCID: PMCPMC12146734. \n41. Chen S, Johnson BK, Yu T, Nelson BN, Walker ED. Elizabethkingia anophelis: Physiologic and \nTranscriptomic Responses to Iron Stress. Front Microbiol. 2020;11:804. Epub 20200507. doi: \n10.3389/fmicb.2020.00804. PubMed PMID: 32457715; PubMed Central PMCID: \nPMCPMC7221216. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 17 \n42. Kampfer P, Matthews H, Glaeser SP, Martin K, Lodders N, Faye I. Elizabethkingia anophelis \nsp. nov., isolated from the midgut of the mosquito Anopheles gambiae. Int J Syst Evol Microbiol. \n2011;61(Pt 11):2670 -5. Epub 20101217. doi: 10.1099/ijs.0.026393 -0. P ubMed PMID: \n21169462. \n43. Gao H, Cui C, Wang L, Jacobs -Lorena M, Wang S. Mosquito Microbiota and Implications for \nDisease Control. Trends Parasitol. 2020;36(2):98 -111. Epub 20191219. doi: \n10.1016/j.pt.2019.12.001. PubMed PMID: 31866183; PubMed Central PMCID: \nPMCPMC9827750. \n44. Song X, Zhong Z, Gao L, Weiss BL, Wang J. Metabolic interactions between disease -\ntransmitting vectors and their microbiota. Trends Parasitol. 2022;38(8):697 -708. Epub \n20220526. doi: 10.1016/j.pt.2022.05.002. PubMed PMID: 35643853. \n45. Sharma P, Rani J, Chauhan C, Kumari S, Tevatiya S, Das De T, et al. Altered Gut Microbiota \nand Immunity Defines Plasmodium vivax Survival in Anopheles stephensi. Front Immunol. \n2020;11:609. Epub 20200514. doi: 10.3389/fimmu.2020.00609. PubMed PMID: 324773 20; \nPubMed Central PMCID: PMCPMC7240202. \n46. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. \nBioinformatics. 2014;30(15):2114 -20. Epub 20140401. doi: 10.1093/bioinformatics/btu170. \nPubMed PMID: 24695404; PubMed Central PMCID: PMCPMC4103590. \n47. Giraldo-Calderon GI, Emrich SJ, MacCallum RM, Maslen G, Dialynas E, Topalis P, et al. \nVectorBase: an updated bioinformatics resource for invertebrate vectors and other organisms \nrelated with human diseases. Nucleic Acids Res. 2015;43(Database issue):D707 -13. Epub \n20141215. doi: 10.1093/nar/gku1117. PubMed PMID: 25510499; PubMed Central PMCID: \nPMCPMC4383932. \n48. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2: accurate alignment \nof transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. \n2013;14(4):R36. Epub 20130425. doi: 10.1186/gb -2013-14-4-r36. PubMed PMID : 23618408; \nPubMed Central PMCID: PMCPMC4053844. \n49. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, et al. Transcript \nassembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching \nduring cell differentiation. Nature Biotechnology. 2010;28(5):511-5. doi: 10.1038/nbt.1621. \n50. Trapnell C, Hendrickson DG, Sauvageau M, Goff L, Rinn JL, Pachter L. Differential analysis of \ngene regulation at transcript resolution with RNA-seq. Nature Biotechnology. 2013;31(1):46-53. \ndoi: 10.1038/nbt.2450. \n51. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome -wide \nexpression patterns. Proc Natl Acad Sci U S A. 1998;95(25):14863 -8. doi: \n10.1073/pnas.95.25.14863. PubMed PMID: 9843981; PubMed Central PMCID: \nPMCPMC24541. \n52. Nadkarni MA, Martin FE, Jacques NA, Hunter N. Determination of bacterial load by real -time \nPCR using a broad -range (universal) probe and primers set. Microbiology (Reading). \n2002;148(Pt 1):257-66. doi: 10.1099/00221287-148-1-257. PubMed PMID: 11782518. \n53. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative \nPCR and the 2( -Delta Delta C(T)) Method. Methods. 2001;25(4):402 -8. doi: \n10.1006/meth.2001.1262. PubMed PMID: 11846609. \n54. Caporaso JG, Lauber CL, Walters WA, Berg -Lyons D, Huntley J, Fierer N, et al. Ultra -high-\nthroughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. The \nISME Journal. 2012;6(8):1621-4. doi: 10.1038/ismej.2012.8. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 18 \nFigures \n \n \n \nFigure 1. Experimental infection of A. darlingi with P. vivax field isolates. (A) Oocyst load per \nmidgut in mosquitoes fed on blood from six gametocytaemic patients (P1-P6). Each red dot \nrepresents a single midgut from an infected mosquito; black squares represent uninfected \nindividuals. Horizontal lines indicate median oocyst load (black); the mean is also shown for P6 \n(blue). Numbers above bars indicate median values. Bar plots below show infection prevalence (% \ninfected, red; % uninfected, black) in each group. (B) Positive correlation between oocyst load and \ninfection prevalence across experimental groups; red line indicates linear regression. (C) \nCorrelation between ABS counts and gametocyte densities in patient blood samples. (D) Positive \ncorrelation between gametocyte density and oocyst load. A gametocyte threshold of ~700/µL \nappears necessary for high infection intensities. (E) Expected presence of parasite developmental \nstages in the mosquito midgut over time post blood feeding (pbf). Sampling timepoints: 1h, 6h, 18h, \n22h, 26h, and 7d pbf. Dot size represents relative abundance; black dots indicate stages expected \nacross all groups, grey dots those expected only in infected mosquitoes. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 19 \n \n \nFigure 2. Differential transcriptional responses in A. darlingi midguts during high and low P. \nvivax infections. (A) Hierarchical clustering of RNA-seq libraries from A. darlingi midguts \nsampled at multiple timepoints (1h, 6h, 18h, 22h, 26h, 7d) post blood feeding. Samples from high \n(top panel) and low (bottom panel) infection groups are shown separately. Grey boxes indicate \nclusters by timepoint; black dots on branches denote bootstrap support levels: >90% (large), >80% \n(medium), >70% (small). (B) Number of genes differentially regulated (infected vs. control) across \ntimepoints in high (left) and low (right) infection groups. Orange bars: upregulated genes; blue bars: \ndownregulated genes. (C) Venn diagrams showing overlap of differentially regulated genes across \nfour timepoints (1h, 6h, 1d, 7d) in high (left) and low (right) infection groups. (D) Gene ontology \n(GO) enrichment analysis of DEGs from the same four timepoints. The Venn diagram indicates 11 \nGO terms shared between infection groups, with 34 and 24 terms uniquely enriched in high and low \ninfection groups, respectively. Bubble plots show significantly enriched GO terms (Padj < 0.05) in \nthree GO categories: molecular function (MF), biological process (BP), and cellular component \n(CC). Bubble size reflects the number of associated genes, ranging from 3 (smallest) to 177 \n(largest). \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 20 \n \n \nFigure 3. Temporal divergence of A. darlingi transcriptional responses resulting in high and \nlow P. vivax infection outcomes. Scatter plots (left panels) compare log2 fold changes in gene \nexpression (infected vs. control) between high and low infection groups at the four timepoints pbf: \n1h (A), 6h (B), 1d (C), and 7d (D). Each point represents a DEG. Pearson correlation coefficient, R2 \nvalues, and p-values are shown within each panel. The least squares regression line in each scatter \nplot is shown in grey. Select DEGs discussed in the main text are highlighted in colour to reflect \nfunctional groupings: red, detoxification and redox regulation; blue, purine and amino acid \nmetabolism; green, immune-related genes; yellow, stress or cell turnover-related genes; violet, \nothers including neuronal and structural components. Heatmaps (right panels) display the \ncorresponding expression profiles across conditions, with yellow indicating upregulation and blue \nindicating downregulation. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 21 \n \n \nFigure 4. Midgut bacterial load and diversity in A. darlingi following P. vivax infection. (A) \nTotal bacterial load, measured by qRT-PCR targeting the 16S rRNA gene, at 1 day post blood \nfeeding (pbf) in high and low infection phenotypes. A significant increase in bacterial load was \nobserved in highly infected mosquitoes (High: I) compared to controls (High: C; p = 0.04). No \nsignificant difference was observed in the low infection group (p = 0.73). (B) Alpha diversity \n(Simpson’s index) across the four timepoints pbf (1h, 6h, 1d, 7d) in high (left) and low (right) \ninfection groups. A significant reduction in bacterial diversity was detected at 1d and 7d pbf in the \nlow infection group (p < 0.05), while no significant differences were observed in the high infection \ngroup. (C) Beta diversity analyses using Bray-Curtis Principal Component Analysis (PCA). Top \nleft: clustering of all samples by infection phenotype (high vs. low), showing distinct microbial \ncommunity structures (p = 0.00003). Top right: separation of control mosquitoes only (C high vs. C \nlow; p = 0.00301). Bottom left: clustering of infected mosquitoes only (I high vs. I low; p = \n0.00145). Bottom right: strongest divergence observed at 1d pbf between high and low infected \nmosquitoes (p = 0.00004). Percent variation explained by each principal component axis is \nindicated. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 22 \n \n \nFigure 5. Taxonomic composition and differential abundance of midgut bacterial \ncommunities in A. darlingi across P. vivax infection phenotypes. (A) Relative abundance of \nbacterial orders across timepoints (1h, 6h, 1d, 7d post blood feeding) in high and low infection \ngroups. Dominant orders include Enterobacteriales, Flavobacteriales, and Pseudomonadales (all \nGammaproteobacteria), along with Rhodospirillales (Alphaproteobacteria) and others. (B) Relative \nabundance of representative bacterial genera within dominant orders for high and low infection \ngroups at the same timepoints. Key differences include higher representation of Elizabethkingia and \nChryseobacterium (Flavobacteriales) in high infection groups, and increased abundance of Pantoea, \nSerratia, Thorsellia (Enterobacteriales), and Pseudomonas and Acinetobacter (Pseudomonadales) in \nlow infection groups. (C) Quantification of the relative abundance of key bacterial orders at all \ntimepoints (left) and specifically at 1d pbf (right) in high vs. low infection groups. Enterobacteriales \nand Pseudomonadales were significantly more abundant in low infection groups (p = 0.003 and p = \n0.0001, respectively), while Flavobacteriales were significantly enriched in high infection groups at \n1d pbf (p = 0.0114). Points represent individual mosquito samples; horizontal bars indicate group \nmedians. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint \n\n 23 \nSupplementary Datasets \nDataset 1. Parasitological study and oocyst load and prevalence data. \nDataset 2. Differential gene expression data in high infection groups. \nDataset 3. Differential gene expression data in low infection groups. \nDataset 4. Gene Ontology (GO) enrichment data in high and low infection groups. \nDataset 5. Differentially expressed genes at each timepoint. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 14, 2025. ; https://doi.org/10.1101/2025.08.13.670040doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}