Rift Valley fever virus and European mosquitoes: vector competence of Culex pipiens and Stegomyia albopicta (= Aedes albopictus)

Rift Valley fever (RVF) is a mosquito‐borne disease caused by the Rift Valley fever virus (RVFV). Rift Valley fever affects a large number of species, including human, and has severe impact on public health and the economy, especially in African countries. The present study examined the vector competence of three different European mosquito species, Culex pipiens (Linnaeus, 1758) form molestus (Diptera: Culicidae), Culex pipiens hybrid form and Stegomyia albopicta (= Aedes albopictus) (Skuse, 1894) (Diptera: Culicidae). Mosquitoes were artificially fed with blood containing RVFV. Infection, disseminated infection and transmission efficiency were evaluated. This is the first study to assess the transmission efficiency of European mosquito species using a virulent RVFV strain. The virus disseminated in Cx. pipiens hybrid form and in S. albopicta. Moreover, infectious viral particles were isolated from saliva of both species, showing their RVFV transmission capacity. The presence of competent Cx. pipiens and S. albopicta in Spain indicates that an autochthonous outbreak of RVF may occur if the virus is introduced. These findings provide information that will help health authorities to set up efficient entomological surveillance and RVFV vector control programmes.


Introduction
Rift Valley fever (RVF) is an arthropod-borne zoonotic disease caused by Rift Valley fever virus (RVFV), an arbovirus of the Phlebovirus genus, belonging to the Bunyaviridae family. Rift Valley fever is a zoonotic disease transmitted by infected mosquitoes to a wide range of hosts, including both domestic (especially sheep and goat) and wild (African buffalo, waterbuck, camel, rat) animals (Olive et al., 2012). Rift Valley fever virus has been isolated in more than 50 mosquito species from seven different genera, although the majority belong to the Culex and Stegomyia (Aedes) genera (Linthicum et al., 2016). The virus was first described in 1931 in the Rift Valley province of Kenya (Daubney et al., 1931). Since then, RVFV has caused significant human and animal outbreaks in several African countries, including South Africa (1950( -1951( ), Egypt (1977( , 1997( , 2003, Mayotte (2007Mayotte ( -2008, Madagascar (2008), Kenya (1997Kenya ( -1998Kenya ( , 2006Kenya ( -2007, Tanzania (2007), Somalia (2007) and Mauritania (2010)(2011)(2013)(2014) (Gerdes, 2004;Nanyingi et al., 2015;Linthicum et al., 2016;Métras et al., 2016). In 2000, RVFV was reported for the first time outside the African continent, in Saudi Arabia and Yemen (Ahmad, 2000). The impact of RVFV on public health and the economy can be very high, as was reported after the epidemic outbreak in Saudi Arabia in 2000. During this outbreak, 883 people were infected, resulting in 124 human deaths, and 40 000 animals died or were aborted (Al-Afaleq & Hussein, 2011). Another important outbreak occurred in Egypt during 1977, during which 200 000 human clinical cases, 600 deaths and economic losses of more than US$115 m were reported (Meegan et al., 1980). The presence of RVFV outside the African continent and especially in countries bordering the Mediterranean Sea, such as Egypt, highlights the possibility that RVF may be introduced into Europe. The risk for the introduction of RVFV into Europe has been reviewed (Chevalier et al., 2010;Rolin et al., 2013;Mansfield et al., 2015). These authors classified the risk for the introduction of RVFV into countries within the European Union as low, mainly because of restrictions imposed by the EU on the import of livestock, differences in climate and seasonal variations in vector and host densities in comparison with those present in Africa. Nonetheless, the illegal importation of infected livestock, especially between Africa and southern Europe and between the Middle East and central Europe, has been indicated as the most likely source of virus introduction into Europe (Chevalier et al., 2010). Climate is a key factor in estimating the risk for RVF outbreaks (Gerdes, 2004 (Garrett-Jones, 1964;Smith et al., 2012). The climatic effects of El Niño are also expected to affect several European countries with rainfall anomalies (USDA, 2015). Furthermore, several unpredictable factors, such as bioterrorism and the intentional introduction of the virus may increase the risk for RVFV introduction, but entomological research and knowledge of vector ecology may improve estimations of risk for RVF (Rolin et al., 2013). Perceptions of the risk for the introduction of RVFV into Europe are further enhanced by evidence from countries in northwest Africa (Mauritania and Senegal) in which outbreaks have been reported (Nanyingi et al., 2015), serological evidence of RVFV antibodies in camels in Morocco (El-Harrak et al., 2011) and the presence of stable competent vector populations in Algeria, Morocco and Tunisia (Moutailler et al., 2008;Amraoui et al., 2012).
Mosquitoes belonging to the Culex pipiens complex are known to be efficient vectors of RVFV (Turell et al., 1996) and have been proposed as the principal vectors during the 1977 outbreak in Egypt (Meegan et al., 1980). Mosquito species of the Culex genus (Culex theileri Theobald, Culex perexiguus Theobald and Culex antennatus) have also been considered as potential vectors as a result of their bio-ecology in terms of abundance, biting activity, feeding habits and longevity [European Food Safety Authority (EFSA), 2013]. Despite the scant data on the possible role of Stegomyia albopicta (= Aedes albopictus) as a vector of RVFV, studies on its host-feeding patterns in rural areas (Valerio et al., 2009;Faraji et al., 2014) would suggest that this species could contribute to RVFV transmission.
Other characteristics of Stegomyia mosquitoes are particularly relevant and may suggest an active role of S. albopicta in RVFV transmission: Stegomyia mosquitoes are able to transmit RVFV transovarially (Linthicum et al., 1985); moreover, the eggs of Stegomyia spp. may enter diapause and survive at temperatures between 0 ∘ C and −15 ∘ C (Thomas et al., 2012). For these reasons, S. albopicta may be important not only for its ability to horizontally transmit the virus, but also for its capacity to maintain viable virus during the coldest winter months. Culex pipiens and S. albopicta are expected to be the main vector species because of their massive presence within the countries of the Mediterranean basin.
To date, only one study of vector competence has been performed in European mosquitoes (Moutailler et al., 2008). The authors of this study tested the capacity of two different strains of RVFV (the virulent strain ZH548 and the avirulent strain Clone 13) to produce disseminated infections in several mosquito species from the Camargue region of France (Turell et al., 1996(Turell et al., , 2007(Turell et al., , 2008aMoutailler et al., 2008;Amraoui et al., 2012). Further, nine species have been tested in Canada (Iranpour et al., 2011), four species in Australia (Turell & Kay, 1998) and 21 species in the U.S.A. (Turell et al., , 2008b(Turell et al., , 2010(Turell et al., , 2013a(Turell et al., , 2013b(Turell et al., , 2015. Given the presence of potential vectors and favourable environmental conditions in some areas (Sanchez-Vizcaino et al., 2013), the possibility of an RVF outbreak event in Spain cannot be excluded. In this context, the vector competence (infection, dissemination and transmission) of two different strains of Cx. pipiens and a strain of S. albopicta were investigated using a virulent strain of RVFV. To the present authors' knowledge, this is the first study to test transmission efficiency in European species using a virulent RVFV strain. Two different approaches involving FTA™ cards and capillary techniques were compared to test viral transmission. Furthermore, experimental mosquito infections were achieved using cycling conditions that simulated environmental conditions. This should allow for a more realistic estimation of vector competence in mosquitoes present in Europe.

Mosquito strains
Two different strains of Cx. pipiens were used. These included a Cx. pipiens form molestus from Empuriabrava (2011) and a hybrid between the pipiens form and molestus form, from Gavà (2012). Molecular characterization of the Cx. pipiens forms was performed as previously described (Bahnck & Fonseca, 2006). Empuriabrava is located near the Aiguamolls de l'Empordà, a wetland area in the north of Catalonia. Gavà is a coastal tourist town within the metropolitan area of Barcelona. The strain of S. albopicta was sourced in 2005 from Sant Cugat de Vallés, a town within the metropolitan area of Barcelona and the site of the first finding of an identified Asian tiger mosquito in Spain in 2004 (Aranda et al., 2006). Mosquitoes were reared under a 14 : 10 h (light:dark) photoperiod with two crepuscular cycles of 30 min inserted to simulate dawn and dusk; mean temperature was 26 ∘ C during the day and 22 ∘ C during the night. Relative humidity (RH) was maintained at a constant 80%. These environmental conditions corresponded to the mean temperature and photoperiod at the latitude of the study area (41 ∘ 24 ′ N, 2 ∘ 10 ′ E) during summer (July and August), when the density and activity of mosquitoes are expected to be at their highest.
The mosquito strains were tested for the presence of viruses from the genera Flavivirus (family Flaviviridae), Alphavirus (family Togaviridae) and Phlebovirus (family Bunyaviridae) by reverse transcription nested polymerase chain reaction (RT-nPCR) (Sánchez-Seco et al., 2001, 2003 to confirm the absence of other viral infections. The strains were also tested for the presence of Wolbachia spp. by PCR analysis of a fragment of wsp gene as previously described (Braig et al., 1998). All strains were positive for Wolbachia spp. but negative for Flavivirus, Alphavirus and Phlebovirus (data not shown).

Design of vector competence assays
Infection rate (IR), disseminated infection rate (DIR) and transmission efficiency (TE) were evaluated. The IR is defined as the proportion of mosquitoes in which the body (abdomen, thorax and head) is infected among all tested mosquitoes. In these mosquitoes, the virus was able to overcome the midgut infection barrier (MIB). The DIR is defined as the proportion of mosquitoes in which the legs and wings are infected among all mosquitoes in which the body is infected. In these mosquitoes, the virus was able to overcome midgut infection and escape barriers. The TE is defined as the proportion of mosquitoes with infectious saliva among the total number of mosquitoes tested (Jupille et al., 2016). In these mosquitoes, the virus was able to overcome the salivary glands infection and escape barriers.
Three assays were performed. In the first assay, the two forms of Cx. pipiens were tested using two different viral doses: 5.7 log 10 TCID 50 /mL and 7.0 log 10 TCID 50 /mL. This first trial was designed to elucidate the IR and DIR of the two forms of Cx. pipiens, mimicking low and medium-high viraemia. The presence of viral RNA in saliva was evaluated using FTA™ Cards (GE Healthcare, Little Chalfont, U.K.). In the second and third assays, two different approaches were used to test saliva samples, using, respectively, FTA™ Cards and a capillary for the direct extraction of saliva from the mosquito. In the second assay, the Cx. pipiens hybrid strain was tested using a viral dose of 7.5 log 10 TCID 50 /mL. In the third assay, the S. albopicta strain was tested using a viral dose of 6.2 log 10 TCID 50 /mL.
In all assays, female mosquitoes aged 7-9 days that had never blood fed were used. Mosquitoes were housed in 0.5-L plastic cages with mesh screening and fed on a 10% sucrose solution ad libitum. The sucrose solution was removed 30 h before the mosquitoes were given infectious bloodmeals. The mosquitoes were fed using a Hemotek feeding system (Discovery Workshop, Accrington, U.K.) at 38 ± 0.5 ∘ C with a specific pathogen-free chicken skin as a membrane (Valo BioMedia GmbH, Osterholz-Scharmbeck, Germany). The mosquitoes were fed with heparinized bovine blood (Universitat Autònoma de Barcelona, Barcelona, Spain) doped with RVFV (maximum virus:blood ratio: 1 : 3) and adenosine 5 ′ -triphosphate (ATP) (5 × 10 −3 M) (Sigma-Aldrich Corp., St Louis, MO, U.S.A.). After feeding on the infected blood, the mosquitoes were anaesthetized with carbon dioxide (CO 2 ) and fully engorged females were selected. The infected blood was titrated in Vero cells; some specimens from each group were killed and analysed to provide inoculum control. The rest of the mosquitoes were individually transferred to cardboard cages (Watkins & Doncaster, Leominster, U.K.) sealed with mesh screening on top and stored inside a climatic cabinet under the environmental conditions described above. Sucrose solution was administered on soaked cotton pledgets placed on the mesh screen and changed every day. All assays were performed in Biosafety Level 3 facilities at the Centre de Recerca en Sanitat Animal (CReSA).

Sample collection
In all assays, FTA™ Cards were used to take saliva samples at different time-points, including at 14 days post-infection (d.p.i.) in the first assay, and at 5 d.p.i. and 14 d.p.i. in the second and third experiments, respectively. FTA™ Cards were soaked in Manuka honey (Manuka Health New Zealand, Te Awamutu, New Zealand) and a blue alimentary colorant. The FTA™ Cards were left for 24 h on the top of the mesh screen of all individual cardboard cages to allow the mosquitoes to feed from the cards. Subsequently, the FTA™ Cards were collected, re-suspended in 0.3 mL of phosphate-buffered saline (PBS) and stored at −80 ∘ C until tested.
At 14 d.p.i., each mosquito was anaesthetized with CO 2 and dissected. The legs and wings were detached from the body and both parts were separately homogenized in 0.5 mL of Dulbecco's modified Eagle's medium (DMEM) (Lonza Group AG, Basel, Switzerland). The samples were homogenized at 30 Hz for 1 min using TissueLyser II (Qiagen GmbH, Hilden, Germany) and stored at −80 ∘ C until tested for RVFV.
In the second and third trials, saliva was extracted from each mosquito at 14 d.p.i. using a capillary technique, as previously described (Dubrulle et al., 2009). Briefly, after the dissection of the legs and wings, the proboscis was inserted into a P20 pipette tip filled with 7 μL of a 1 : 1 solution of fetal bovine serum (FBS) and 50% sucrose solution. To stimulate salivation, 1 μL of 1% pilocarpine (Sigma-Aldrich Corp.), prepared in PBS at 0.1% Tween 80, was applied to the thorax of each mosquito. After 60 min, the solution containing the saliva was expelled into 1.5-mL tubes containing 193 μL of DMEM; 150 μL were used for viral RNA extraction and the remaining samples were used for RVFV isolation in a monolayer of Vero cells. Cells were incubated for 7 days (37 ∘ C, 5% CO 2 ) and cytophatic effects were evaluated.

Virus detection
Viral RNA was extracted from samples (bodies, legs and wings, FTA™ Cards and saliva) using NucleoSpin ® RNA Virus (Macherey-Nagel GmbH & Co. KG, Düren, Germany) according to the manufacturer's recommendations. The RT-PCR was performed as previously described with minor modifications (Drosten et al., 2002). Reverse transcription quantitative PCR (RT-qPCR) was carried out using AgPath-ID™ One-Step RT-PCR Reagents (Applied Biosystems, Inc., Foster City, CA, U.S.A.) without adding supplementary MgSO 4 . The samples were amplified using a 7500 Fast Real-Time PCR System (Applied Biosystems, Inc.) programmed as follows: 48 ∘ C for 10 min; 95 ∘ C for 10 min, and 45 cycles at 95 ∘ C for 15 s and at 57 ∘ C for 35 s. The limit of detection was estimated at 0.09 TCID 50 per reaction.

Mosquito feeding and mortality
The two forms of Cx. pipiens behaved differently during artificial feeding; data are shown in Table 1. The mean ± standard deviation (SD) feeding rate was higher in Cx. pipiens hybrids (25.6 ± 6.62%) than in the Cx. pipiens form molestus strain (3.8 ± 1.96%). Mean ± SD mortality rates at 14 d.p.i. were 3.3 ± 1.10% and 6.2 ± 8.83% in Cx. pipiens hybrids and Cx. pipiens form molestus, respectively. In S. albopicta, the feeding rate was 24.5% and the mortality rate at 14 d.p.i. was 7.3%.

Mosquito infection and dissemination
Infected mosquito bodies were detected for both Cx. pipiens forms tested using the lowest viral dose (5.7 log 10 TCID 50 /mL). However, disseminated infection was not detected in any mosquito. By contrast, a viral dose of 7.0 log 10 TCID 50 /mL was able to induce both infection and disseminated infection in the Cx. pipiens hybrid form. At the same viral dose, Cx. pipiens form molestus presented infection, but not dissemination. Both infection and disseminated infection were detected in S. albopicta. The IRs and DIRs are summarized in Table 2. The virus was able to cross the midgut barriers in mosquitoes of the Cx. pipiens hybrid form and S. albopicta showing disseminated infection (positive legs and wings).

Transmission of RVFV
Rift Valley fever virus was detected in saliva of the Cx. pipiens hybrid form and S. albopicta using both FTA™ Cards and the capillary technique. A positive saliva sample indicates that the virus was able to cross salivary gland barriers. All positive saliva samples (FTA™ Cards and collected saliva) are reported in Table 3 and are related to the corresponding samples of legs and wings and to the results of isolation in Vero cells (presence/absence of cytophatic effects).
In the first assay, no FTA™ Cards tested positive, regardless of the titre used or the mosquito strain tested.
In the second assay, two FTA™ Cards tested positive. These FTA™ Cards referred to two different specimens of Cx. pipiens hybrid form (M-46 and M-49) and were sampled at different time-points (5 and 14 d.p.i.). Neither mosquito presented disseminated infection. Five saliva samples obtained using the capillary technique in Cx. pipiens hybrids tested positive by RT-qPCR. Three of these produced cytophatic effects when inoculated in Vero cells. Two specimens (M-14 and M-59) with infectious viral particles in saliva presented disseminated infection. Conversely, specimen M-5 did not present disseminated   infection. The FTA™ Cards for these three specimens were negative.
In summary, two patterns of transmission were evidenced in Cx. pipiens, in, respectively, females that were able to transmit the virus without presenting disseminated infection, and females that transmitted the virus and presented disseminated infection.
In the third assay, one FTA™ Card referred to an S. albopicta specimen with disseminated infection (V-3) tested positive at 5 d.p.i., although the FTA™ Card tested negative at 14 d.p.i. However, the saliva sample extracted at 14 d.p.i. from the same mosquito tested positive by RT-PCR and was able to induce a cytophatic effect when inoculated in Vero cells.
Transmission efficiency was estimated at 4.8% (three mosquitoes with infectious saliva out of 62 mosquitoes fed) in Cx. pipiens hybrids and 2.6% (one of 38 mosquitoes) in S. albopicta.

Discussion
Mosquitoes belonging to Culex and Stegomyia (= Aedes) spp. are considered to be the main vectors for RVFV. In the present study, for the first time, two strains of different Cx. pipiens forms and one strain of S. albopicta collected in Spain were demonstrated to be susceptible to RVFV infection. Moreover, Cx. pipiens hybrids and S. albopicta were able to transmit RVFV.
Culex pipiens form molestus exhibited a lower propensity to feed from the artificial feeding system used, and was not found to be a useful laboratory species for vector competence studies.
In the Cx. pipiens hybrid form, rates of infection and dissemination tended to increase proportionally to the viral dose used during blood feeding, as previously observed in several species (Turell et al., 2008b(Turell et al., , 2013b. This finding probably reflects the presence of dose-dependent midgut barriers which the virus must overcome in order to successfully infect and disseminate to the whole mosquito body (Franz et al., 2015). A previous study, performed with the RVFV strain ZH501, suggested the presence of a midgut escape barrier (MEB) in Cx. pipiens form molestus (Turell et al., 2014). This may explain the absence of disseminated infection in this species in the present study, but the low number of Cx. pipiens form molestus that fed successfully did not provide sufficient data to strongly support this hypothesis and thus further studies are required to clarify this point.
Rates of infection and dissemination in Spanish Cx. pipiens were lower than those in Cx. pipiens tested in France (Moutailler et al., 2008) and Canada (Iranpour et al., 2011). A comparison of the experimental procedures used in the earlier studies with those used in the present study shows some significant differences: (a) the titres used for challenges were higher [10 7.9 -10 9.4 plaque-forming units (PFU)/mL]; (b) the source of feeding was a live infected hamster, and (c) the experimental procedures were conducted under a constant temperature of 28 ∘ C in the French study and 25 ∘ C in the Canadian study. As mentioned before, viral dose directly influences both infection and dissemination rates. The use of a living host for feeding improves the competence of the mosquito specimens tested (Turell, 1988;Lord et al., 2006). In the present study, the viral doses used corresponded to the viral loads detected in blood from European lambs experimentally infected with the same virulent RVFV strain (Busquets et al., 2010). It is known that a higher and constant extrinsic incubation temperature (EIT) corresponds to high rates of infection, dissemination and transmission, as has been experimentally demonstrated for different arboviruses (Richards et al., 2007;Kilpatrick et al., 2008;Lambrechts et al., 2011). A previous study on the effect of EIT on the vector competence of Cx. quinquefasciatus for West Nile virus (WNV) suggested that EIT can influence both the MIB and MEB (Anderson et al., 2010). A more recent study found that the cycling of environmental conditions can also affect vector competence for WNV in Cx. pipiens and S. albopicta (Brustolin et al., 2016). Cycling of environmental conditions directly affects vector competence in the strains tested and hence was applied in the present study to mimic environmental conditions in the field in order to better estimate vector competence. With regard to the influence of the viral load used, the findings of the present study can be compared with those of a previous study in which two forms of Cx. pipiens from the U.S.A. were assayed using a similar viral load (10 7.5 PFU/mL) (Turell et al., 2014). However, the authors of the U.S. study used an infected hamster as a blood source and specimens were maintained at a constant EIT of 26 ∘ C (Turell et al., 2014). As result, the infection and dissemination rates obtained were higher than those in the present study.
With regard to RVFV transmission, the present results show that the Spanish S. albopicta and the Cx. pipiens hybrid form strains could possibly sustain the RVFV transmission cycle in nature. One positive FTA™ Card at 5 d.p.i. provided evidence of early transmission capacity in the Cx. pipiens hybrid form, as previously observed in Cx. pipiens from the Maghreb region (Amraoui et al., 2012). In the Maghreb populations, the presence of infectious viral particles was observed from 3 d.p.i The rates of RVFV infection, dissemination and transmission observed in the Spanish S. albopicta strain are comparable with those obtained in a previous vector competence study conducted in S. albopicta mosquitoes from Texas  fed with an infectious bloodmeal at a final titre of 10 4.7 PFU. The finding of a positive FTA™ Card at 5 d.p.i. also showed early transmission capacity, which contrasts with that described previously for the Texas specimens, which were able to transmit RVFV only at 14 d.p.i The FTA™ Card was originally designed as a surveillance tool for arbovirus detection in field studies and was intended to avoid the analysis of trapped vectors ( Van den Hurk et al., 2012). The exposure period for FTA™ Cards was 7 days in field studies. It is probable that a shorter period of exposure will limit the possibility that mosquitoes will feed on the card, which will result in a lower number of positive FTA™ Cards compared with the number of positive saliva samples obtained by capillary extraction. The presence of a blue-coloured belly indicated that the specimen had fed from an FTA™ Card soaked in honey. Negative FTA™ Cards from mosquitoes with positive saliva mainly corresponded to specimens without a blue belly, although the blue belly was not always evident to the naked eye. However, some authors have suggested that forcing salivation in a capillary for 30-45 min may produce an inaccurate overestimation of viral transmission (Smith et al., 2006). The differences in the results obtained by FTA™ Cards and those obtained in saliva directly extracted with the capillary technique are likely to reflect several factors: (a) lower sensitivity of the FTA™ Card technique; (b) an insufficient period of exposure of the FTA™ Card, and (c) an overestimation of viral shed in the capillary.
Two Cx. pipiens hybrid mosquitoes (M-5 and M-49) with positive saliva samples, but without dissemination infection, were observed. Previous studies have described the possibility that RVFV might disseminate from the midgut via the trachea (Romoser et al., 2005;Kading et al., 2014). This would provide a direct pathway to the salivary gland without the need for dissemination in haemocoel and other secondary target organs. Therefore, two patterns of RVFV transmission in females are reported: transmission in females with disseminated infection, and transmission in females without disseminated infection. Further experiments with higher numbers of RVFV-infected mosquitoes are required to strengthen this model.
The three strains of mosquito used in the present study were all naturally infected by Wolbachia spp. This may have influenced the vector competence of infected mosquitoes, as has been shown in previous studies (Moreira et al., 2009;Walker et al., 2011). However, further studies regarding this issue are required to elucidate its possible role in arbovirus-vector interactions.
The risk for the introduction of RVFV into regions of Spain in which livestock densities are high and environmental conditions are favourable has been analysed in a previous study (Sanchez-Vizcaino et al., 2013). Several Spanish regions, including Catalonia, were found to be suitable for an RVF outbreak. The findings related to RVFV vector competence presented in the current work would support this possibility as both Spanish S. albopicta and Cx. pipiens hybrid strains appear to be able to sustain the cycle of RVFV transmission.
In conclusion, the data presented in this work provide information that will help in the establishing of effective vector control programmes and surveillance plans to prevent and control possible RVF outbreaks. Additional studies are required to evaluate the vector competence of other European autochthonous vectors and their possible roles during an RVF outbreak.
at the animal facilities of the Centre de Recerca en Sanitat Animal (CReSA) for their assistance in Biosafety Level 3 facilities. This work was financially supported by the Spanish Government (grant no. MINECO AGL2013-47257-P) and by an EU grant (FP7-613996 VMERGE) and is catalogued by the VMERGE Steering Committee as VMERGE017 (http://www.vmerge.eu).