The Culex pipiens complex include species such as the common house mosquito Culex pipiens pipiens that is a wide spread species distributed over extensive areas in Europe and Africa and has been introduced in the Americas and in some temperate areas of Asia and Australia [1, 2]. Two forms or biotypes have been described: the molestus and the pipiens forms [1]. Although morphologically indistinguishable as adults, these two forms present genetic, behavioural and physiological differences. The molestus form is stenogamous and autogenous, that is mosquitoes are able to mate in confined environments and lay their eggs in absence of a previous blood meal, respectively. By contrast, mosquitoes of the pipiens form use open environments for mating (eurygamous) and requires a blood meal, as a nutrient source, for oviposition (anautogenous). In North European countries, the two forms use different habitats with Cx. p. pipiens form molestus frequently living in underground environments in areas with human influence, while mosquitoes of the pipiens form are mainly present in aboveground habitats [3, 4]. In countries of the Mediterranean basin, mosquitoes of the two forms are sympatric and, because molestus and pipiens forms are not completely genetically isolated, hybrids are frequent [5–8]. Hybridization between the two forms has been also occasionally reported in North Europe, in Germany [9], Austria [10], the Netherlands [11, 12] and the UK [13]. Hybridization between forms has also been reported under laboratory conditions [14].

The importance of mosquito species for the transmission of vector-borne pathogens is strongly determined by, among other factors, their feeding behaviour [15]. In addition to the differential competence of each mosquito species to develop a particular pathogen, the feeding behaviour of mosquitoes determines the contact rate between infected and susceptible vertebrate hosts. Culex p. pipiens play a key role in the transmission of numerous vector borne pathogens affecting humans, livestock and wildlife including viruses (e.g. West Nile virus, WNV; [16]), protozoa (e.g. avian malaria; [17]) and metazoa (e.g. filarial worms [18]) parasites (reviewed by [2]). Culex p. pipiens females feed mainly on birds (69–97%) [19], but mammals are also an important fraction of their blood meals, compromising over 20% of the blood meals in some populations [19, 20], also see [2, 21, 22]. However, it has been suggested that Cx. p. pipiens form molestus feed mainly on mammals while Cx. p. pipiens form pipiens feed mainly on birds [23], see also [24]. Hybrids show a higher vectorial competence for the transmission of some pathogens than both pipiens and molestus forms (e.g. WNV, [25]), and it has been suggested that due to their intermediate feeding behaviour [23], they may act as bridge vectors for the transmission of pathogens between birds and humans, as in the case of WNV [4].

Avian malaria parasites of the genus Plasmodium are mosquito-borne parasites commonly found infecting birds [26]. Nowadays, approximately 40 different species have been described [26] but, since the seminar paper by Bensch et al. [27], a high genetic diversity of cytochrome b (cyt b) lineages of Plasmodium parasites have been identified. Several mosquito genera transmit avian malaria parasites, including different species of the genus Culex, which may play a central role in the transmission of avian malaria [17]. In particular, Cx. p. pipiens may play a key role in the transmission dynamic of Plasmodium parasites under natural conditions. Plasmodium infected birds are common hosts of this mosquito species, allowing frequent parasite-Cx. p. pipiens encounters [22, 28]. Furthermore, avian Plasmodium parasites, including P. relictum, successfully develops in Cx. p. pipiens as supported by experimental infections of mosquitoes [17, 29–31], and different avian Plasmodium lineages have been molecularly isolated from mosquitoes trapped in the wild [32–37].

The aims of this study were (1) to investigate the spatial distribution of pipiens and molestus forms and their hybrids in relation to the type of habitats (natural, rural and urban areas), (2) to compare the feeding patterns of both forms in these habitats and (3) to determine avian malaria prevalence and diversity in these mosquitoes. Previous evidence suggest that the two Cx. p. pipiens forms and their hybrids use differential habitats [3, 10], thus a higher frequency of the molestus form is expected in urban areas than in natural ones, while the opposite pattern is expected for mosquitoes of the pipiens form. Furthermore, the molestus form may feed mainly on mammals, blood meals from mosquitoes of the pipiens form should be predominantly avian-derived [38] while their hybrids may show an intermediate feeding behaviour [23, 38]. If the feeding pattern differs between mosquito forms, the prevalence of infection by avian malaria parasites will be higher in the pipiens form due to the expected higher frequency of avian-derived blood meals with a decreased prevalence in mosquitoes of the molestus form and hybrids as the proportion of mammal-derived blood meals increased.

Mosquitoes were captured using BG-sentinel traps baited with BG-lure and dry ice as a source of CO₂. Overall, from April to December, 5–6 trapping sessions were conducted at each site. In each trapping session, three traps were operated for 24 h at 45 different sampling sites in Cadiz, Huelva and Seville provinces (15 per each province) once every 45 days (Fig. 1). The sampling sites were grouped in triplets including one natural habitat (conserved landscapes with low density of human and livestock), one rural habitat (farms with livestock) and one urban habitat (urbanized densely populated areas), see [39] for further details of sampling protocol.

Adult mosquitoes were preserved in dry ice and stored frozen until morphological identification. Mosquitoes were sorted on a Petri plate on a chill table under a stereomicroscope and morphologically identified to species level following [40, 41]. Culex p. pipiens females with a recent blood meal in the abdomen were selected for this study. The abdomen of engorged mosquitoes was separated from the head-thorax using sterile tips. Genomic DNA of each the abdomen and the head-thorax of each mosquito was extracted using the DNA Kit Maxwell^® 16LEV kit [42].

The DNA extracted from the mosquito abdomen was employed to identify the vertebrate origin of the blood meals using a nested PCR following Alcaide et al. [43]. Briefly, this method amplifies a 758-base pairs fragment (excluding primers) of the mitochondrial cytochrome oxidase 1 (COI) gene (barcoding region) of vertebrate species. Amplicons were sequenced in the Macrogen sequencing service (Macrogen Inc., The Netherlands). Vertebrate hosts of mosquitoes were identified by comparisons of the sequences obtained from mosquitoes with those available in public databases (GenBank DNA sequence database, National Center for Biotechnology Information Blast, and/or the Barcode of Life Data Systems).

The DNA extracted from the head-thorax of each engorged mosquito was used to identify the Cx. p. pipiens form and the presence and identity of avian malaria parasite lineages. The mosquito forms were identified following Bahnck and Fonseca [44]. This method is based in the amplification of the 5′ flacking region of CQ11 microsatellite and has been routinely used to identify the Cx. p. pipiens forms in Old-world countries [5, 7, 8], including Spain [45]. The presence and identity of Plasmodium and Haemoproteus parasites in the head-thorax of mosquitoes was determined following Helgreen et al. [46]. Parasite lineages were determined by BLAST comparison of the sequences with those deposited in Genbank and the morphospecies was determined based on the information available in MalAvi [47].

Statistical analyses were conducted using the GLIMMIX procedure in SAS 9.2 with a binomial distribution using a Laplace approximation. Firstly, generalized linear mixed models were used to identify potential differences in the occurrence of each form (molestus or pipiens) and their hybrids between habitats. The triplet (see sampling sites above) was included as a random factor in order to control for geographical pseudo-replication of the samples. Secondly, differences in the presence of avian-derived blood meals between mosquito forms were tested using the presence/absence of avian derived blood meals as the dependent variable, mosquito forms and habitat category as fixed factors and triplet as a random factor. Finally, differences in the prevalence of blood parasites between mosquito forms and habitats were tested by including parasite infection status as the dependent variable, mosquito forms and habitat category as fixed factors and triplet as a random factor.

A higher percentage of the pipiens form was found in natural than in urban areas, while the percentage of mosquitoes of the molestus form and hybrids did not differ between the three habitat categories considered in this study. Traditionally, anautogenous mosquitoes, a characteristic typically assigned to the form pipiens, are more commonly found in aboveground and rural habitats, while the autogenous form, corresponding to Cx. p. pipiens form molestus, is usually found in underground urban environments [1, 4, 48, 49]. However, reports of the differential distribution of genetically characterized mosquito forms in areas with different degrees of anthropization are scarce, especially in Southern Europe. In Portugal, the proportion of mosquitoes of the pipiens form was higher in farms located in peri-urban areas without dwellings in the vicinity than indoors in residential areas, supporting a negative association between the frequency of this form and the degree of urbanization [38]. In addition, although the frequency of molestus forms did not significantly differ between urban and peri-urban habitats, the proportion of molestus mosquitoes tended to be higher in indoors traps in areas densely populated by humans rather than in small villages, suggesting the higher presence of this form in urbanized environments [38]. Our results support those from Osorio et al. [38], in spite of the differences between studies regarding the sampling locations (indoors/outdoors).

It is generally assumed that pipiens and molestus forms have a bird and mammal blood feeding preference respectively, while hybrids show an intermediate behaviour [23]. This fact suggest that hybrids may play a central role in the transmission of pathogens such as WNV and USUTU virus (USUV), which circulate between birds but occasionally infect humans. In North America, mosquitoes with a higher ancestry from the molestus form feed more frequently on mammals, including humans [24, 50]. Contrary to this pattern, we did not find any significant association between mosquito forms and their feeding sources. The two forms and their hybrids feed mainly on birds, with a relative high percentage (33%) of blood meals derived from mammals, including humans. A similar feeding pattern of Cx. p. pipiens was reported in Portugal, where mosquitoes of the two forms feed mainly on birds with not significant differences between forms [51], but see [38]. One potential explanation is the different approaches used to identify the mosquito forms between studies, while studies from North America genotyped mosquitoes using seven [50] or ten [24] microsatellite markers, here the mosquito forms were assessed with the CQ11 microsatellite. However, Gomes et al. [51] used six microsatellites and obtained very similar results to our study. Alternatively, mosquito feeding pattern is geographically-dependent [1, 21] and varies seasonally [50, 52]. This variability between areas and seasons may partially explain discrepancies in the feeding source of mosquito forms between studies. The similar feeding behaviour of the two mosquito forms and their hybrids found here supports that all of them could be involved in the transmission of pathogens from birds to humans, as WNV or USUV.

Overall, 6.7% of mosquitoes harboured blood parasites in this study. Glaizot et al. [33] found a similar prevalence (6.6%) in Culex pipiens, although slightly higher values were reported in other previous studies [36, 37]. Differences between studies could be due to the temporal variation in the prevalence of Plasmodium usually found in mosquitoes [35, 36], but also due to methodological differences between studies including the protocol used for parasite detection (e.g. using engorged mosquitoes or mosquitoes with a completely digested blood meal) or the method used for mosquito sampling [53]. Furthermore, the number of mosquitoes with a mammal and avian derived blood meal may differ between trapping methods [54] potentially biasing estimates of parasite prevalence. This fact should be take into account in studies on the epidemiology of vector-borne parasites transmitted by mosquitoes with an opportunistic feeding behaviour. Nevertheless, in this study, the prevalence of avian malaria parasites did not differ between the two forms and their hybrids, which may be due at least in part, to the similar blood feeding behaviour reported between them. However, the absence of significant differences between forms could also be due to the low statistical power owing to the number of mosquitoes analysed and the low prevalence of blood parasites. Avian Plasmodium is a mosquito-borne parasite while the related Haemoproteus is mainly transmitted by Culicoides and louse flies [26]. As occurred here, previous studies have reported the presence of Haemoproteus in mosquitoes [22, 34, 35, 37, 55], however experimental evidence available exclude mosquitoes as vectors of this parasite genus [31, 56]. The Haemoproteus passeris lineage Padom05 was isolated from a mosquito with a recent blood meal on a house sparrow, which is a common avian host of this parasite lineage [31, 57]. This, together with the fact that the mosquito had a recent blood meal from this bird species, suggests that the Haemoproteus DNA derived from rests of undigested blood present in the head-thorax of the mosquito. Alternatively, it is possible that a non-infective form of the parasite present in the mosquito head-thorax was amplified [56], but see [31]. Although molecular isolation does not demonstrate vector competence, the Plasmodium parasites isolated from mosquitoes suggest that the two mosquito forms and their hybrids may frequently interact with birds infected by avian malaria parasites. Moreover, the Plasmodium lineages found in this study have been previously isolated from Cx. pipiens mosquitoes in Europe including Switzerland [33, 36], Portugal [58], Spain [35], France [37] and Italy [22]. Avian Plasmodium develops in laboratory reared Cx. p. pipiens form molestus [29, 30] and these parasites have been molecularly identified in field-collected mosquitoes of this form [32]. Complete development of avian Plasmodium has been reported in mosquitoes of the pipiens form [17]. However, to our knowledge, this is the first study testing the potential differences of the two Cx. p. pipiens forms and their hybrids in the prevalence of avian Plasmodium under natural conditions. Results from this study suggest that, given the similar feeding behaviour of mosquito forms in the study area, the two mosquito forms and their hybrids are similarly involved in the transmission of avian Plasmodium in Southern Spain.

This study supports a differential frequency of mosquitoes of the pipiens form between urban and natural areas. While the hypothesis that both forms and their hybrids differ in their feeding sources was not supported by results from this study, with birds as the most common blood source found in all the cases. Consequently, Cx. p. pipiens molestus and pipiens mosquitoes and their hybrids may frequently interact with WNV, USUV and/or avian malaria parasites playing a role in the transmission of these pathogens.