nature communications Article https://doi.org/10.1038/s41467-022-32234-y Zika vector competence data reveals risks of outbreaks: the contribution of the European ZIKAlliance project Received: 5 April 2022 Thomas Obadia1,2, Gladys Gutierrez-Bugallo 3,4, Veasna Duong 5, Ana I. Nuñez6, Rosilainy S. Fernandes 7, Basile Kamgang 8, Liza Hery4, Accepted: 18 July 2022 Yann Gomard 9, Sandra R. Abbo 10, Davy Jiolle11, Uros Glavinic12, Myrielle Dupont-Rouzeyrol13, Célestine M. Atyame9, Nicolas Pocquet14, Sébastien Boyer 15, Catherine Dauga16, Marie Vazeille 16, André Yébakima17, Check for updates Michael T. White 2, Constantianus J. M. Koenraadt18, Patrick Mavingui9, 1234567890():,; 1234567890():,; Anubis Vega-Rua4, Eva Veronesi12, Gorben P. Pijlman 10, Christophe Paupy11, Núria Busquets6, Ricardo Lourenço-de-Oliveira 7, Xavier De Lamballerie 19 & Anna-Bella Failloux 16 First identified in 1947, Zika virus took roughly 70 years to cause a pandemic unusually associated with virus-induced brain damage in newborns. Zika virus is transmitted by mosquitoes, mainly Aedes aegypti, and secondarily, Aedes albopictus, both colonizing a large strip encompassing tropical and temperate regions. As part of the international project ZIKAlliance initiated in 2016, 50 mosquito populations from six species collected in 12 countries were experi- mentally infected with different Zika viruses. Here, we show that Ae. aegypti is mainly responsible for Zika virus transmission having the highest susceptibility to viral infections. Other species play a secondary role in transmission while Culex mosquitoes are largely non-susceptible. Zika strain is expected to sig- nificantly modulate transmission efficiency with African strains being more likely to cause an outbreak. As the distribution of Ae. aegypti will doubtless expand with climate change and without new marketed vaccines, all the ingredients are in place to relive a new pandemic of Zika. 1 Institut Pasteur, Université Paris Cité, Bioinformatics and Biostatistics Hub, F-75015 Paris, France. 2Institut Pasteur, Université Paris Cité, G5 Infectious Disease Epidemiology and Analytics, F-75015 Paris, France. 3Department of Vector Control, Center for Research, Diagnostic, and Reference, Institute of Tropical Medicine Pedro Kouri, Havana, Cuba. 4Institut Pasteur of Guadeloupe, Laboratory of Vector Control Research, Unit Transmission Reservoir and Pathogens Diversity, Les Abymes, Guadeloupe. 5Institut Pasteur du Cambodge, Virology Unit, Phnom Penh, Cambodia. 6IRTA, Centre de Recerca en Sanitat Animal (CReSA, IRTA-UAB), Campus de la Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. 7Laboratorio de Mosquitos Transmissores de Hematozoarios, Instituto Oswaldo Cruz, Fiocruz, Rio de Janeiro, RJ, Brazil. 8Centre for Research in Infectious Diseases, Department of Medical Entomology, Yaoundé, Cameroon. 9UMR PIMIT (Processus Infectieux en Milieu Insulaire Tropical), Sainte-Clotilde, La Réunion, France. 10Laboratory of Virology, Wagen- ingen University, Wageningen, The Netherlands. 11IRD, MIVEGEC, University of Montpellier, IRD, CNRS, Montpellier, France. 12National Centre for Vector Entomology, Institute of Parasitology, Vetsuisse Faculty, University of Zürich, Zürich, Switzerland. 13Institut Pasteur de Nouvelle-Calédonie, URE Dengue et Arboviroses, Nouméa, New Caledonia. 14Institut Pasteur de Nouvelle-Calédonie, URE Entomologie Médicale, Nouméa, New Caledonia. 15Institut Pasteur du Cambodge, Medical Entomology Unit, Phnom Penh, Cambodia. 16Institut Pasteur, Université Paris Cité, Arboviruses and Insect Vectors, F-75015 Paris, France. 17 VECCOTRA, Rivière Salée, Martinique. 18Laboratory of Entomology, Wageningen University & Research, Wageningen, The Netherlands. 19Unité des Virus Emergents (UVE), Aix Marseille Université, IHU Méditerranée Infection, Marseille, France. e-mail: anna-bella.failloux@pasteur.fr Nature Communications | (2022)13:4490 1 Article https://doi.org/10.1038/s41467-022-32234-y Viral pathogens with high epidemic potential have marked human Results history by massive losses of life and economic declines. Pandemics like Culex spp. mosquitoes were resistant to most ZIKV strains, successfully the Spanish flu1 have left traces and the fear of new viral emergences blocking infection with an overall infection rate (IR) of 0.2% (2 mos- materializes today with the Severe Acute Respiratory Syndrome- quitoes out of 1089 tested) and complete absence of both dis- Coronavirus-2 (SARS-CoV-2) pandemic. Over the past decades, efforts semination efficiency (DE) and transmission efficiency (TE). As such, invested into vaccination programs and development of antiviral they were not included in more complex vector competence analyses. treatments have led to major medical progress resulting in measles Aedes spp. showed different patterns depending on species and ZIKV decline, smallpox eradication, appropriate treatments of Human strain used for infection. For Ae. aegypti and Ae. albopictus, the Immunodeficiency virus and Hepatitis C virus, and improved control experimental design required nested, within-country random effects. of SARS-CoV-2. As major viral pathogens, arthropod-borne viruses In the case of Ae. japonicus, the country and mosquito population (arboviruses) are also on the rise. Arboviruses have developed a effects were fully colinear, therefore analyses only relied on logistic strategy to accomplish their transmission cycle via replication in both regressions. vertebrate and arthropod hosts. Their RNA genome contributes to the generation of large populations of genetically distinct variants per- Differential susceptibility of Aedes spp. mosquitoes to ZIKV mitting adaptations to environmental changes including host switch- Depending on mosquito species and ZIKV strain, we observed differ- ing and transmission2. In the last 30 years, mosquito-borne viruses ent patterns of viral progression within the vector’s organs. In Ae. have dramatically expanded their distribution range leading to aegypti infected with ZIKV strains, there were no notable differences increasingly frequent and large epidemics3. In 2016, Zika was desig- between IR and DE at 21 days post-infection (dpi), with IR = 94.3% (95% nated as a “Public Health Emergency of International Concern” by the CI: [92.1%–96.0%]) and 69.1% [64.0%–73.9%], and with DE = 89.4% World Health Organization, subsequently to the large spread of Zika [86.6%–91.8%] and 64.3% [59.1%–69.3%] for respectively, African and virus (ZIKV) in the Pacific islands and the Americas with unusual noti- American strains. However, there was a large and significant decrease fications of microcephaly in newborns and different neurological dis- for TE = 71.9% [68.0%–75.6%] and 30.0% [25.3%–35.1%] for respectively, orders (e.g. Guillain-Barré syndrome); this happened 70 years after its African and American strains as can be seen in Fig. 1. This is consistent first isolation from a human case in East Africa4. Local cases of Zika with efficient viral progression through the midgut, and partial barriers were detected in 2019 in Southern France additionally corroborating to viral progression before reaching the salivary glands. In contrast, the role of the Asian tiger mosquito (Aedes albopictus) in ZIKV trans- when infected with Asian ZIKV strains, both midguts and salivary mission in addition to chikungunya and dengue viruses5. glands mitigated progression of the virus: IR = 68.9% [64.3%–73.3%], Aedes mosquitoes are important vectors of arboviruses, DE = 56.7% [51.8%–61.5%], and TE = 28.9% [24.7%–33.5%]. acquiring the virus through a blood meal from an infected host. In Ae. albopictus mosquitoes, viral progression was more efficient Subsequently, the virus replicates within the mosquito and is with the African ZIKV strain than with American or Asian ZIKV strains transmitted to a new host. The virus must overcome two anatomical for all three outcomes. For the African ZIKV strain, a regular decrease barriers (midgut and salivary glands) and circumvent major antiviral was observed between IR (73.9%), DE (69.2%) and TE (55.8%), con- pathways before being delivered in saliva6. Geographic populations sistent with the absence of a marked barrier to viral progression of a same mosquito species may not share the same immunological through the midgut and into the salivary glands. In contrast, DE was background, leading to different vector competences7. The out- significantly lower for American (15.6% [11.5–20.4%]) and Asian (11.2% come of infection depends on specific pairings of mosquito and [7.7–15.5%]) ZIKV strains, consistent with a lack of viral progression pathogen genotypes described under genotype-by-genotype (G x through the midgut. G) interactions8. These interactions could structure viral popula- Time since infection showed an increasing, approximately linear tions through adaptations to their local vector populations8. Vector association with DE and TE in Ae. aegypti and Ae. albopictus mosqui- competence data have been used to derive risk maps informed with toes, suggesting an accumulation pattern and an overall absence of distribution of mosquito species, to help focus and strengthen viral clearance within the vector, while IR reached a peak value within 7 vector control measures9. Nevertheless, other factors, non-genetic, dpi and remained at comparable levels until 21 dpi. In contrast, Ae. biotic and abiotic, should be considered in the functioning of the japonicus infected by the African ZIKV strain showed that IR peaked at vectorial system making it even more complex than initially 14 dpi, with decreasing DE and TE values over time (TE7dpi = 15.0% thought10. A global vector competence analysis of potential Zika [7.1–26.6%], TE14dpi = 15.0% [7.1–26.6%] and TE21dpi = 6.7% [18.5–16.2%]). vectors (Aedes aegypti, Ae. albopictus, Aedes japonicus, Culex pipiens pipiens, Culex pipiens molestus, and Culex quinquefasciatus) was Transmission efficiency is higher with the African ZIKV performed to assess the risk of Zika disease outbreaks. Once the Although all species of Aedes studied were competent to transmit vector is well established in a region, associated arboviruses have ZIKV, infected Ae. aegypti species were on average more prone to the potential to emerge through imported human cases. To this successful dissemination than Ae. albopictus and Ae. japonicus aim, we analyzed the vector competence of 50 mosquito popula- (TEaegypti = 34.0% [24.0–45.8%], TEalbopictus = 17.6% [11.4–26.0%], tions from 15 locations and 12 countries (Brazil, Cambodia, TEjaponicus = 7.2% [2.9–16.9%]). This was even more pronounced at 21 Cameroon, Congo, Cuba, France, Gabon, Guadeloupe, Haiti, La dpi when TE was significantly higher for Ae. aegypti than Ae. albo- Reunion, Madeira, The Netherlands, New Caledonia, Spain, and pictus (46.7% [35.3–58.5%] vs. 25.3% [17.3–35.5%], p < 0.0001) and Ae. Switzerland) which were challenged with different ZIKV genotypes japonicus (46.7% [35.3–58.5%] vs. 10.8% [4.4–24.1%], p = 0.0002), the (West Africa, Asia, America) using standardized experimental pro- latter two not reaching statistical significance (p = 0.09). However, tocols (mosquito infections, viral titrations, parameters of vector many combinations of ZIKV strains and mosquito populations were competence). In this work, we elaborated a vector competence not available for Ae. japonicus for which results should be con- data-driven prediction for ZIKV transmission to inform on the cur- sidered less robust. rent risk of Zika transmission at a global scale. We show that Ae. The African ZIKV strain was the most efficient in infecting and aegypti populations tested are highly competent to transmit the disseminating, therefore posing a greater risk in case it was newly different ZIKV genotypes and more specifically, the African geno- introduced in a region: the predicted expected marginal means 21 dpi type. Moreover, other mosquito species such as Ae. albopictus or Ae. were TE = 82.7% [78.5–86.1%] for Ae. aegypti and TE = 52.4% japonicus can play a secondary role in transmission while Culex [45.4–59.3%] for Ae. albopictus (Fig. 2). American and Asian strains had mosquitoes are definitively non-susceptible to ZIKV. comparable TE regardless of time since infection: (1) in Cameroon, Nature Communications | (2022)13:4490 2 Article https://doi.org/10.1038/s41467-022-32234-y Fig. 1 | Infection rate, dissemination and transmission efficiencies of ZIKV observed means of IR, DE and TE. (n = 30 biologically independent mosquitoes for strains (Africa, America, Asia) according to time since infection in three Aedes each combination of species, country, mosquito population, days post-infection mosquito species. Panels show the rates achieved by averaging over all sampled and ZIKV strain were studied, unless stated otherwise. See Supplementary Table 1 mosquitoes, i.e., by pooling together mosquitoes from all countries used in the for complete break-down of sample sizes). Colors correspond to the studied out- study. Error bars are exact 95% binomial confidence intervals centered on the come (IR, DE, TE). Congo, and Gabon, TE was low (~10%), suggesting African mosquitoes locations of strains within a country. For Ae. japonicus, no obvious are mostly susceptible to local ZIKV strains; (2) in Brazil, Haiti, Gua- pattern could be detected with only 5.2% of the observed variance deloupe, Madeira, La Reunion and New Caledonia, TE reached mod- explained by available covariates suggesting that mosquitoes were less erate values ~20%. In Cambodia however, Asian and American ZIKV susceptible to tested ZIKV strains. strains had much higher TE values, denoting mosquito populations highly susceptible to ZIKV. Discussion Our results confirmed that human-adapted Aedes mosquitoes prepare Breaking down variable importance in ZIKV susceptibility the field for new arbovirus emergences; in addition to other factors, We evaluated the extent to which each covariate of interest con- some combinations of vector-virus are more prone to cause outbreaks, tributed to explanatory value of fitted models. For both Ae. aegypti and but globalization will smooth out differences leading most emergences Ae. albopictus, TE showed low variability among mosquito populations to turn into pandemics owing to the large territories covered by vectors. originating from the same country (Fig. 3). Another striking result from During the 2015–2016 Latin American and the Caribbean epi- Fig. 3 confirmed the higher susceptibility to African ZIKV strains demic, the role of Culex as ZIKV vector has been the subject of major among all tested mosquitoes, both within and across countries. This scientific debates which, if verified, would have revised the modus suggests that within a given country, mosquito populations were operandi of vector control11. As ZIKV has also been detected in field- equally able to disseminate ZIKV regardless of other extrinsic factors. collected Culex mosquitoes12,13, their susceptibility to ZIKV was then We found that 51.7% of variance in Ae. aegypti TE could be explained by tested; from 36 studies on vector competence of Culex mosquitoes for a statistical model accounting for mosquito strain, ZIKV strain, and dpi ZIKV, seven studies showed that ZIKV was present in the salivary (Table 1). The relative importance of these factors was assessed by glands of Cx. restuans, Cx. quinquefasciatus, Cx. tarsalis, and Cx. Cor- estimating the proportion of variance explained solely by these fac- onator. Asian and American genotypes of ZIKV may occasionally infect tors. For Ae. aegypti TE, ZIKV strain was the most important factor Culex mosquitoes but not African strains14. From our dataset15–18, we explaining 33.5% of variance, followed by dpi (30.1%). Of the factors showed that ZIKV can accumulate in Cx. pipiens saliva only after forced considered, mosquito strain had the least impact on TE, with only 6.6% infection by intrathoracic inoculation, short-circuiting the midgut of variance explained by the country of a sampled mosquito strain. An barrier17; after oral infection, the virus was unlikely to disseminate in additional 2.6% of variance was explainable when we considered the mosquito ruling out the role of Culex mosquitoes in ZIKV Nature Communications | (2022)13:4490 3 Article https://doi.org/10.1038/s41467-022-32234-y Fig. 2 | Model-predicted transmission efficiency of Aedes mosquitoes according TE. (n = 30 biologically independent mosquitoes for each combination of to time since infection, split by country of mosquito sampling. TE predicted for species, country, mosquito population, days post-infection and ZIKV strain (A) Ae. aegypti and (B) Ae. albopictus. Countries are ordered to reflect geo- were studied, unless stated otherwise. See Supplementary Table 1 for complete graphic proximity. Error bars show asymptotic 95% confidence interval from break-down of sample sizes). Colors correspond to continent-aggregated the mixed regression models centered on the average predictions of IR, DE and ZIKV strains. Fig. 3 | Transmission efficiencies of ZIKV strains at the regional level for Ae. clustering that reflected their phylogeny. Vertical bars help identify different aegypti and Ae. albopictus mosquitoes sampled at one or more locations in countries, shown on the x-axis with individual colors. Empty cells represent absence every studied country, at 21 dpi. ZIKV strains were pooled in a geographical of data. Color gradients correspond to TE values (ranging 0–100%). transmission. Viral replication was not detected in Culex which was outside Africa21. The “domestication” of this species (named Ae. aegypti unable to trigger the RNAi pathway as no viral replication signatures aegypti) allowed its world-wide dispersion coinciding with the first such as 21 nt viral siRNAs were detected19. Even though vectors of the twentieth century pandemics caused by arboviruses, dengue22, and flavivirus West-Nile virus, a phylogenetically ZIKV-related arbovirus, then chikungunya23 and Zika24. In addition to increased vector-host mosquitoes of the Cx. pipiens complex are not competent for ZIKV contacts, Ae. aegypti aegypti has developed an enhanced permissive- transmission and are very unlikely to play a role of vector during a Zika ness to some arboviruses7. At the mosquito level, successful trans- outbreak. Thus, to control Zika, targeted interventions should focus mission implies that the virus passes through two anatomical barriers, largely on Aedes mosquitoes11,20. midgut and salivary glands25. Ae. aegypti mosquitoes we tested were Aedes aegypti is an African mosquito (named Ae. aegypti formosus) more prone to impose a strong bottleneck to the African ZIKV in the living primarily in rainforests colonizing natural breeding sites which in salivary glands leading to decrease ZIKV transmission. Nevertheless, all a subsequent step, has become adapted to living in the human envir- 11 Ae. aegypti populations transmitted at least 2–3 times more the onment (developing a human-biting behavior and colonizing human- African ZIKV than the American and Asian ZIKV strains underlining the made containers) before undergoing different waves of dissemination high potential of African ZIKV to trigger an outbreak outside and inside Nature Communications | (2022)13:4490 4 Article https://doi.org/10.1038/s41467-022-32234-y Table 1 | Contribution of random effects and fixed effects in species-wise regression models Total variance explained Variance uniquely explained Mosquito species Within-country (mosquito) Between-country (mosquito) ZIKV strain Days post-infection Ae. aegypti 51.7% 2.6% 6.6% 33.5% 30.1% Ae. japonicus 5.2% 1.0% 0% 1.0% Ae. albopictus 41.5% 1.7% 0% 29.0% 21.2% The contribution of random effects was measured as the adjusted intra-class correlation coefficient (as defined in Nakagawa et al.52), while the contribution of fixed effects was assessed by the semi- partial R2 measure (from Stoffel et al.51), i.e. the share of variance uniquely explained by a given covariate. The total variance explained is the R2 measure of the saturated model. Africa. It is now admitted that African ZIKV strains were more patho- predominate over Ae. aegypti in habitats where both species are genic than the Asian strains26,27. African Ae. aegypti (Cameroon, Congo, sympatric42 and will certainly promote arbovirus spillovers at the Gabon) were more resistant to American and Asian ZIKV than to edge of their natural sylvatic cycle owing to its generalist behavior African ZIKV suggesting a local adaptation between a ZIKV strain and a (biting both humans and animals, and living in rural areas)41. Ae. mosquito population. There is evidence today that Ae. aegypti for- albopictus from Central Africa are likely originated from South mosus is found both in urban and forest habitats in sub-Saharan America which themselves were introduced from North America in Africa28 and a mixture of Ae. aegypti aegypti and Ae. aegypti formosus is 198643. European Ae. albopictus are presumably a genetic mixture likely present in cities we sampled in Cameroon, Congo and Gabon29. of populations from United States and La Reunion43, sharing As Ae. aegypti aegypti is more susceptible to ZIKV than Ae. aegypti similar profiles of ZIKV transmission with mosquitoes from La formosus, an intermediate level of susceptibility can be expected7. Ae. Reunion16,44,45. American ZIKV predominant during the 2015–2016 aegypti formosus, which is predominant in Benoué, was previously epidemic, was efficiently transmitted by Ae. albopictus, in Europe5 found less susceptible to ZIKV compared to the other populations and Brazil46. In the absence of Ae. aegypti, Ae. albopictus is more from Cameroon30. Nevertheless, mosquitoes were only identified likely to trigger ZIKV outbreaks when ZIKV comes from Africa. according to morphological criteria31 and not based on their geno- Another invasive species, Ae. japonicus also experimentally types. As such, the observed local reduction in vector competence transmits the African ZIKV47. Native to East Asia, Ae. japonicus arrived in may also result from other inherited factors not detected through Europe in 200041 and is established in 12 European countries up to morphological examinations. Our experiments showed a decreased now48. Its role in pathogen transmission is still unclear even if arboviral susceptibility within the country, with Ae. aegypti from Benoué having ZIKV RNAs have been detected in field-collected mosquitoes49. Its high lower TE (34.4% [20.2–52.1%]) than those sampled elsewhere in ability to become infected, and to a lower extent, to ensure viral dis- Cameroon (ranging from 62.5 to 83.3%, Supplementary Table 2) semination and transmission, concomitantly to its human-biting except in Maroua, located close to Benoué. Intriguingly, human epi- behavior gives Ae. japonicus the status of being a competent vector demics of ZIKV associated with Ae. aegypti have been observed in a of ZIKV47,50. These findings are however subject to caution, as the limited number of countries in Asia or Africa following the 2015–2016 experimental design may not allow for generalization of our observa- Latin American and the Caribbean epidemic24. The Asian Ae. aegypti tions: Ae. japonicus could only be sampled and bred in European population from Cambodia we tested, transmitted the most efficiently countries (France (n = 90), Switzerland (n = 90), the Netherlands all three ZIKV strains, African, American and Asian, corroborating its (n = 62)). To add to this limitation, only two ZIKV strains were tested good vector competence to ZIKV. Recent systematic screenings of against these mosquitoes and fully collinear with the country of mos- patients revealed that ZIKV was circulating in Singapore, Thailand, quito sampling: French- and Swiss-sampled Ae. japonicus were infected Cambodia, Myanmar and Vietnam suggesting that ZIKV has been with an African ZIKV strain, whereas mosquitoes from the Netherlands present in Asia at a low but sustained level for years or even were infected with an American ZIKV strain (see data structure in decades32–35. Moreover, human population immunity could also mod- Supplementary Fig. 1). Therefore, we cannot distinguish between ulate the risk of outbreaks36; a large population exposed to dengue mosquito susceptibility and ZIKV strain potency. viruses may develop neutralizing antibodies against dengue viruses We conclude that from our comprehensive assessment of vector which cross-react with ZIKV adjusting the outcome of ZIKV infection36. competence, invasive Aedes mosquitoes set the bed for pandemic Thereby, Asian populations can be partially protected against ZIKV37. arboviruses; in the absence of the historical vector of human arbo- At an intra-country level, mosquito populations (in our study, e.g. five viruses, Ae. aegypti, other species such as Ae. albopictus or Ae. japo- in Brazil18) share the same profiles of transmission with a highest nicus can take over as ZIKV vectors and transmit even more efficiently transmission level with African ZIKV. Once introduced, the African when infected with the African ZIKV. ZIKV has the potential to cause an outbreak whatever the origin of Ae. aegypti populations. Our assessments are based on vector competence Methods which is only one aspect of vectorial capacity; it results from complex Experimental procedures interactions of multiple factors influencing pathogen transmission by a All partners of the ZIKAlliance project (https://zikalliance.tghn.org/) vector, including internal factors such as genetic, evolution, immunity, performed their own mosquito experimental infections using a or interactions with other microorganisms, modulated by external common standardized protocol (Supplementary Fig. 2). Some factors such as abiotic (like climate or topography) and biotic (like deviations from this protocol could not be avoided (e.g. availability nutrition or hosts)38. of blood source, feeding via blood droplets instead of via mem- In the tropical belt, Ae. aegypti and Ae. albopictus share the brane). Fifty populations (Fig. 4 and Supplementary Table 1) were same ecological niche39. Ae. albopictus mosquitoes we tested were collected as immature stages and reared in insectaries under con- more susceptible to African ZIKV suggesting that the species can trolled conditions. Mosquito females were exposed to an infectious contribute to an epidemic like it happened in Gabon in 200740. A blood meal containing one third of viral suspension and two third of lower transmission of American and Asian ZIKV was presumably washed rabbit erythrocytes (Supplementary Fig. 3) at a final titer of related to a strong mosquito midgut barrier30. Since its arrival in ~107 TCID50/ml. After 1 h, engorged females were isolated in boxes Central Africa in 2000, Ae. albopictus plays a major role in the (28°±1 °C and 80 ± 10% humidity) and fed with 10% sucrose until transmission of different arboviruses41. Ae. albopictus tends to analysis. Nature Communications | (2022)13:4490 5 Article https://doi.org/10.1038/s41467-022-32234-y Fig. 4 | Distribution of mosquito species experimentally infected with Zika countries sampled. Colors correspond to mosquito species. The map was built viruses. The color code used represents the different mosquito species and using the open source map site https://cmap.comersis.com/cartes-Monde- numbers refer to the number of mosquito populations tested. In green, are the WORLD.html. Different geographic ZIKV strains were used (Supplementary combinations of mosquito populations and ZIKV strains, according Fig. 4): (1) Africa (Dakar; KU955592; isolated in 1984 from Aedes taylori to time since infection. Time since infection was considered a in Senegal and passaged four times on BHK21 cells), (2) Asia (Cambo- categorical covariate. Prior to any modeling approach, all out- dia; KU955593; isolated from a human case in 2010 and passaged three comes were visualized with heat maps similar to those presented in times on Vero cells and Malaysia; KX694533; isolated in 1966 from Supplementary Figs. 6, 8 and 9, for all combinations of ZIKV strain Aedes aegypti in Malaysia and passaged four times on Vero cells), and and mosquito population and at every value of time since infection. (3) America (Guadeloupe; LR792671.1; isolated from a human case in The resulting experimental design, consisting of combinations of Guadeloupe in 2016 and passaged two times on Vero cells, Martinique; country of study, mosquito species and populations as well as time KU647676; isolated from a human case in 2015 and passaged three since infection, was used to guide the development of generalized times on Vero cells, and Suriname; KU937936; isolated from a human linear (mixed) models. serum in Suriname in 2016 and passaged 5–6 times on Vero cells). ZIKV Owing to complex hierarchical experimental design, each mos- Dakar, Malaysia, and Martinique provided by EVAg (https://www. quito species was modeled separately. When several mosquito popu- european-virus-archive.com/) were used in most infections. ZIKV lations were sampled at different locations within a given country, the Cambodia was used for the mosquito populations Haiti, Funchal, variations resulting from the population effect was modeled by a Mont, Corse, ZIKV Suriname for Best, Amsterdam and Lelystad, and random effect nested within the country, in a mixed regression fra- ZIKV Guadeloupe for HAV-PT, HAV-PRG. mework. The contribution of each fixed-effect term was assessed by At 7, 14, and 21 days post-infection (dpi), a batch of mosquitoes refitting univariable models and measuring the semi-partial R-squared was processed to estimate if viral particles were present in the body value (as defined by Stoffel et al.51). Likewise, the contribution of ran- (thorax and abdomen) as the marker for infection, head for dis- dom effects was measured as the adjusted intra-class correlation semination and saliva for transmission; 0 refers to no viral particles coefficient (as defined by Nakagawa et al.52). Analyses were conducted and 1 to at least one viral particle detected. Presence of viral particles using the R software (v. 4.2.0)53 with the following packages: lme4 (v. was asserted by cytopathic effects observed on a monolayer of Vero 1.1-27.1)54, emmeans (v. 1.7.2), performance (v.0.8.0)55 and partR2 CCL-81 cells in 96-well plates45. Three parameters were measured: (1) IR (v. 0.9.1). corresponding to the proportion of mosquitoes with infected body The robustness of results was assessed using sensitivity analyses, (thorax and abdomen) among examined mosquitoes, (2) DE referring where the same regression models were applied with data aggregated to the proportion of mosquitoes with virus detected in the head at larger geographical scales for covariates of interest. Three variations among examined mosquitoes, and (3) TE representing the proportion of the same analyses were conducted: (1) by aggregating country of of mosquitoes with virus detected in saliva among examined mos- study by continent (therefore assuming random variations across quitoes. TE was used as the primary endpoint to evaluate vector countries) and preserving individual ZIKV strains, (2) by aggregating competence with respect to ZIKV strain. It is important to note that, by ZIKV strains by continent (corresponding to relatedness between definition, TE ≤ DE ≤ IR, the denominator of each being the total strains in the phylogenetic tree, see Supplementary Fig. 4) and pre- number of mosquitoes dissected. serving individual countries of study, and (3) aggregating both country and virus strain by continent. The results from aggregation detailed in Statistical analysis point (2) are presented in the main text as they provide the most Success for transmission, dissemination and infection were inves- comprehensive set of results. The results corresponding to aggrega- tigated as binomial outcomes to assess how their population-level tion levels from points (1), (3) and (4) are presented in Supplementary counterpart (TE, DE, IR) were associated with available Figs. 5 through 10. Nature Communications | (2022)13:4490 6 Article https://doi.org/10.1038/s41467-022-32234-y Reporting summary 18. Fernandes, R. S. et al. Vector competence of Aedes aegypti, Aedes Further information on research design is available in the Nature albopictus and Culex quinquefasciatus from Brazil and New Cale- Research Reporting Summary linked to this article. donia for three Zika virus lineages. Pathogens 9, 575 (2020). 19. Lourenço-de-Oliveira, R. et al. Culex quinquefasciatus mosquitoes Data availability do not support replication of Zika virus. J. Gen. Virol. 99, The data that support the findings of this study are fully available in an 258–264 (2018). electronic Supplementary table (Microsoft Excel XLSX format) at 20. Lourenço-de-Oliveira, R. & Failloux, A.-B. Lessons learned on Zika https://zenodo.org/record/6788898#.Ysgu0oTP2Uk as well as in the virus vectors. PLoS Negl. Trop. Dis. 11, e0005511 (2017). GitLab repository found in the Code availability section. 21. Gloria-Soria, A. et al. Global genetic diversity of Aedes aegypti. Mol. Ecol. 25, 5377–5395 (2016). Code availability 22. Guzman, M. G. & Harris, E. Dengue. Lancet 385, 453–465 (2015). All analyses conducted in this manuscript and associated code have 23. Weaver, S. C. & Lecuit, M. Chikungunya virus and the global spread been uploaded to a Git repository at https://gitlab.pasteur.fr/tobadia/ of a mosquito-borne disease. N. Engl. J. Med. 372, 17289-zikalliance and are readily available. The results of the present 1231–1239 (2015). manuscript are based on commit 1477a8be1554c91d2a5ea03896440 24. Weaver, S. C., Charlier, C., Vasilakis, N. & Lecuit, M. Zika, Chi- 7e88499ddf9. kungunya, and other emerging vector-borne viral diseases. Annu. Rev. Med 69, 395–408 (2018). References 25. Kramer, L. D. & Ciota, A. T. Dissecting vectorial capacity for 1. Patterson, K. D. & Pyle, G. F. The geography and mortality of the 1918 mosquito-borne viruses. Curr. Opin. Virol. 15, 112–118 (2015). influenza pandemic. Bull. Hist. Med. 65, 4–21 (1991). 26. Aubry, F. et al. Recent African strains of Zika virus display higher 2. Poirier, E. Z. & Vignuzzi, M. Virus population dynamics during transmissibility and fetal pathogenicity than Asian strains. Nat. infection. Curr. Opin. Virol. 23, 82–87 (2017). Commun. 12, 916 (2021). 3. Huang, Y.-J. S., Higgs, S. & Vanlandingham, D. L. Emergence and re- 27. Smith, D. R. et al. African and Asian Zika virus isolates display emergence of mosquito-borne arboviruses. Curr. Opin. Virol. 34, phenotypic differences both in vitro and in vivo. Am. J. Trop. Med. 104–109 (2019). Hyg. 98, 432–444 (2017). 4. Dick, G. W. A., Kitchen, S. F. & Haddow, A. J. Zika Virus (I). Isolations 28. Powell, J. R. Mosquitoes on the move. Science 354, 971–972 (2016). and serological specificity. Trans. R. Soc. Tropical Med. Hyg. 46, 29. Weetman, D. et al. Aedes mosquitoes and Aedes-borne arbo- 509–520 (1952). viruses in Africa: current and future threats. Int. J. Environ. Res. 5. Giron, S. et al. Vector-borne transmission of Zika virus in Europe, Public Health 15, 220 (2018). southern France, August 2019. Eurosurveillance 24, 30. Kamgang, B. et al. Different populations of Aedes aegypti and 1900655 (2019). Aedes albopictus (Diptera: Culicidae) from Central Africa are sus- 6. Sim, S., Jupatanakul, N. & Dimopoulos, G. Mosquito immunity ceptible to Zika virus infection. PLoS Negl. Trop. Dis. 14, against arboviruses. Viruses 6, 4479–4504 (2014). e0008163 (2020). 7. Aubry, F. et al. Enhanced Zika virus susceptibility of globally inva- 31. McClelland, G. A. H. A worldwide survey of variation in scale pat- sive Aedes aegypti populations. Science 370, 991–996 (2020). tern of the abdominal tergum of Aedes aegypti (L.) (Diptera: Culi- 8. Lambrechts, L. et al. Genetic specificity and potential for local cidae). Trans. R. Entomol. Soc. Lond. 126, 239–259 (1974). adaptation between dengue viruses and mosquito vectors. BMC 32. Ho, Z. J. M. et al. Outbreak of Zika virus infection in Singapore: an Evolut. Biol. 9, 160 (2009). epidemiological, entomological, virological, and clinical analysis. 9. Mariconti, M. et al. Estimating the risk of arbovirus transmission in Lancet Infect. Dis. 17, 813–821 (2017). Southern Europe using vector competence data. Sci. Rep. 9, 33. Ruchusatsawat, K. et al. Long-term circulation of Zika virus in 17852 (2019). Thailand: an observational study. Lancet Infect. Dis. 19, 10. Viglietta, M., Bellone, R., Blisnick, A. A. & Failloux, A.-B. Vector 439–446 (2019). specificity of arbovirus transmission. Front. Microbiol. 12, 34. Tun, M. M. N. et al. Detection of Zika virus infection in Myanmar. 773211 (2021). Am. J. Trop. Med. Hyg. 98, 868–871 (2018). 11. Roundy, C. M. et al. Lack of evidence for Zika virus transmission by 35. Quyen, N. T. H. et al. Chikungunya and Zika virus cases detected Culex mosquitoes. Emerg. Microbes Infect. 6, 1–2 (2017). against a backdrop of endemic Dengue transmission in Vietnam. 12. Guo, X. et al. Culex pipiens quinquefasciatus: a potential vector to Am. J. Trop. Med. Hyg. 97, 146–150 (2017). transmit Zika virus. Emerg. Microbes Infect. 5, 1–5 (2016). 36. Wen, J. & Shresta, S. Antigenic cross-reactivity between Zika and 13. Guedes, D. R. et al. Zika virus replication in the mosquito Culex dengue viruses: is it time to develop a universal vaccine? Curr. quinquefasciatus in Brazil. Emerg. Microbes Infect. 6, 1–11 (2017). Opin. Immunol. 59, 1–8 (2019). 14. Viveiros-Rosa, S. G., Regis, E. G. & Santos, W. C. Vector compe- 37. Luo, X. S., Imai, N. & Dorigatti, I. Quantifying the risk of Zika virus tence of Culex mosquitoes (Diptera: Culicidae) in Zika virus trans- spread in Asia during the 2015-16 epidemic in Latin America and mission: an integrative review. Rev. Panam. Salud Publica 44, the Caribbean: a modeling study. Travel Med. Infect. Dis. 33, e7 (2020). 101562 (2020). 15. Hery, L., Boullis, A., Delannay, C. & Vega-Rúa, A. Transmission 38. Lefèvre, T., Vantaux, A., Dabiré, K. R., Mouline, K. & Cohuet, A. Non- potential of African, Asian and American Zika virus strains by Aedes genetic determinants of mosquito competence for malaria para- aegypti and Culex quinquefasciatus from Guadeloupe (French sites. PLoS Pathog. 9, e1003365 (2013). West Indies). Emerg. Microbes Infect. 8, 699–706 (2019). 39. Lwande, O. W. et al. Globe-Trotting Aedes aegypti and Aedes 16. Gomard, Y., Lebon, C., Mavingui, P. & Atyame, C. M. Contrasted albopictus: risk factors for arbovirus pandemics. Vector-Borne transmission efficiency of Zika virus strains by mosquito species Zoonotic Dis. 20, 71–81 (2020). Aedes aegypti, Aedes albopictus and Culex quinquefasciatus from 40. Grard, G. et al. Zika Virus in Gabon (Central Africa)—2007: a new Reunion Island. Parasites Vectors 13, 398 (2020). threat from Aedes albopictus? PLoS Negl. Trop. Dis. 8, 17. Abbo, S. R. et al. Forced Zika virus infection of culex pipiens leads to e2681 (2014). limited virus accumulation in mosquito saliva. Viruses 12, 41. Ngoagouni, C., Kamgang, B., Nakouné, E., Paupy, C. & Kazanji, M. 659 (2020). Invasion of Aedes albopictus (Diptera: Culicidae) into central Nature Communications | (2022)13:4490 7 Article https://doi.org/10.1038/s41467-022-32234-y Africa: what consequences for emerging diseases? Parasites Vec- The funders had no role in study design, data collection and inter- tors 8, 191 (2015). pretation, or decision to submit the work for publication. 42. Kamgang, B. et al. Temporal patterns of abundance of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) and mitochon- Author contributions drial DNA analysis of Ae. albopictus in the Central African Republic. T.O. performed the writing and editing, data analysis and interpretation. PLoS Negl. Trop. Dis. 7, e2590 (2013). G.G.B., A.I.N., R.S.F., B.K., L.H., Y.G., S.R.A., D.J., U.G., M.V., A.Y., and 43. Vega-Rúa, A. et al. Vector competence of Aedes albopictus C.J.M.K. did the investigation and data acquisition. V.D. did the investi- populations for chikungunya virus is shaped by their demographic gation, methodology, and data acquisition. M.D.R., C.M.A., N.P., S.B., history. Commun. Biol. 3, 1–13 (2020). P.M., A.V.R., E.V., G.P.P., C.P., N.B., and R.L.O. contributed to supervision, 44. Nuñez, A. I. et al. Evidence of Zika virus horizontal and vertical data analysis and interpretation. C.D. and M.T.W. did the data analysis transmission in Aedes albopictus from Spain but not infectious and interpretation. X.D.L. participated in funding acquisition and project virus in saliva of the progeny. Emerg. Microbes Infect. 9, administration. A.B.F. intervened in conceptualization, project adminis- 2236–2244 (2020). tration, supervision, and writing original draft. All authors have read and 45. Vazeille, M. et al. Zika virus threshold determines transmission by agreed with the published version of the manuscript. European Aedes albopictus mosquitoes. Emerg. Microbes Infect. 8, 1668–1678 (2019). Competing interests 46. Parra, M. C. P. et al. Detection of Zika RNA virus in Aedes aegypti The authors declare no competing interests. and Aedes albopictus mosquitoes, São Paulo, Brazil. Infect., Genet. Evolution 98, 105226 (2022). Additional information 47. Glavinic, U. et al. Assessing the role of two populations of Aedes Supplementary information The online version contains supplementary japonicus japonicus for Zika virus transmission under a constant material available at and a fluctuating temperature regime. Parasites Vectors 13, https://doi.org/10.1038/s41467-022-32234-y. 479 (2020). 48. Koban, M. B. et al. The Asian bush mosquito Aedes japonicus Correspondence and requests for materials should be addressed to japonicus (Diptera: Culicidae) in Europe, 17 years after its first Anna-Bella Failloux. detection, with a focus on monitoring methods. Parasites Vectors 12, 109 (2019). Peer review information Nature Communications thanks Sally Paulson 49. Yang, F. et al. Cache valley virus in Aedes japonicus japonicus and the other, anonymous, reviewer(s) for their contribution to the peer Mosquitoes, Appalachian Region, United States. Emerg. Infect. Dis. review of this work. 24, https://doi.org/10.3201/eid2403.161275 (2018). 50. Abbo, S. R. et al. The invasive Asian bush mosquito Aedes japoni- Reprints and permission information is available at cus found in the Netherlands can experimentally transmit Zika http://www.nature.com/reprints virus and Usutu virus. PLoS Negl. Trop. Dis. 14, e0008217 (2020). Publisher’s note Springer Nature remains neutral with regard to jur- 51. Stoffel, M. A., Nakagawa, S. & Schielzeth, H. partR2: partitioning R2 isdictional claims in published maps and institutional affiliations. in generalized linear mixed models. PeerJ 9, e11414 (2021). 52. Nakagawa, S., Johnson, P. C. D. & Schielzeth, H. The coefficient of Open Access This article is licensed under a Creative Commons determination R2 and intra-class correlation coefficient from gen- Attribution 4.0 International License, which permits use, sharing, eralized linear mixed-effects models revisited and expanded. J. R. adaptation, distribution and reproduction in any medium or format, as Soc. Interface 14, 20170213 (2017). long as you give appropriate credit to the original author(s) and the 53. R Core Team. R: A Language and Environment for Statistical Com- source, provide a link to the Creative Commons license, and indicate if puting (R Foundation for Statistical Computing, 2022). changes were made. The images or other third party material in this 54. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed- article are included in the article’s Creative Commons license, unless effects models using lme4. J. Stat. Softw. 67, 1–48 (2015). indicated otherwise in a credit line to the material. If material is not 55. Lüdecke, D. et al. see: An R Package for Visualizing Statistical included in the article’s Creative Commons license and your intended Models. https://doi.org/10.31234/osf.io/m4uax (2021). use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright Acknowledgements holder. To view a copy of this license, visit http://creativecommons.org/ We thank Richard Paul (Institut Pasteur) for correcting English. This study licenses/by/4.0/. was supported by the European Union’s Horizon 2020 research and innovation program under ZIKAlliance grant agreement no. 734548. © The Author(s) 2022 Nature Communications | (2022)13:4490 8