Wolbachia density and cytoplasmic incompatibility in Aedes albopictus: concerns with using artificial Wolbachia infection as a vector suppression tool.

The mosquito Aedes albopictusi is a competent vector of harmful human pathogens, including viruses causing dengue and chikungunya. Cytoplasmic incompatibility (CI) induced by endosymbiotic Wolbachia can be used to produce functionally sterile males that can be released in the field as a suppression tool against this mosquito. Because the available sexing methods are not efficient enough to avoid unintentional release of a few transinfected females, we assessed the CI pattern in crosses between wPip Wolbachia-transinfected (ARwP) females and wild-type males of Ae. albopictus in this study. Quantitative polymerase chain reaction was used to monitor the titer of the Wolbachia strains that naturally infect Ae. albopictus, that is, wAlbA and wAlbB, in age-controlled males and females. Data were coupled with incompatibility level detected when the above-mentioned males were crossed with ARwP females. Wolbachia infection titer was also monitored in samples of wild caught males. Incompatibility level was positively correlated only with wAlbA density. Crosses between wild-type males having very low wAlbA density (<0.001 wAlbA/actin copy numbers) and ARwP females were partially fertile (CIcorr = 68.06 ± 6.20). Individuals with low wAlbA titer were frequently found among sampled wild males (30%-50% depending on the site and period). ARwP males can be as considered as a very promising tool for suppressing Ae. albopictus. However, crosses between wild males having low wAlbA density and ARwP females may be partially fertile. In the case of local establishment of the transinfected mosquito line, this occurrence may favor the replacement of the wild-type mosquitoes with the ARwP line, thus reducing the long-term efficacy of incompatible insect technique. Various alternative strategies have been discussed to prevent this risk and to exploit Wolbachia as a tool to control Ae. albopictus.

Introduction

Aedes (Stegomyia) albopictus (Diptera: Culicidae), commonly known as Asian tiger mosquito, is one of the most invasive insect species worldwide [1,2]. It shows high vector competence for several arboviruses, including viruses causing dengue (DENV) and chikungunya (CHIKV) [3]. Although Ae. albopictus is not considered as the primary vector of DENV, it is becoming a major cause of viral outbreaks because of rapid changes in its overall distribution. In the last decade, Ae. albopictus caused new DENV epidemics in different countries such as Hawaii [4], Mauritius [5], and China [6]. With respect to CHIKV, a recent mutation in genes encoding envelope glycoproteins of CHIKV of an African lineage enhanced the adaptability of this virus to Ae. albopictus [7], leading to outbreaks in Indian Ocean islands in 2005–2006 [8] and in temperate regions such as Italy [9] and France [10]. The CHIKV mosquito-human transmission cycle recently has established in the Caribbean [11]. The associated epidemic is currently out of control, with more than 500,000 cases in a few months, and models are being developed to predict the possible spread of the virus in the Americas [12].

In recent years, interest in mosquito control strategies based on the principles of autocidal control, such as release of radio-sterilized males (Sterile Insect Technique, SIT) in field [13], has increased, thus providing new approaches by using innovative biotechnology (such as insect transgenesis and endosymbiont manipulation) [14-17]. The potential to combine these recent approaches with the classical SIT may offer viable solutions to overcome intrinsic limitations of the current conventional vector control strategies that mostly rely on insecticides and community participation. In addition to exerting negative effects on non-target insects and having toxicological impact on humans and environment, insecticides result in the establishment of resistant strains, as shown in some recent reports [18,19]. Therefore, alternative strategies to control mosquitoes are being sought worldwide.

Research on the endosymbiont Wolbachia pipientis (Alphaproteobacteria: Rickettsiales) has increased steadily in the last two decades, driven by the possibility of exploiting its biological properties as tools for insect pest and vector control [16,20]. This maternally inherited bacterium manipulates host reproduction and is carried by approximately 40% arthropod insect species as well as some crustaceans, mites, and filarial nematodes [21]. Cytoplasmic incompatibility (CI) is the most common reproductive phenotype observed in arthropod species infected with specific Wolbachia strains [,22-27]. First described in mosquito Culex pipiens [28-30], CI is a conditional embryonic lethality that occurs when males infected with CI-inducing Wolbachia strains are crossed with uninfected females (unidirectional CI [Uni-CI]) or with females carrying other incompatible Wolbachia strains (bidirectional CI [Bi-CI]).

In 2009, an incompatible Ae. albopictus strain (ARwP) was generated to support a SIT project against Ae. albopictus in Italy [31]. ARwP was obtained by replacing natural Wolbachia double infection with a single heterologous strain of Wolbachia (wPip) taken from Culex pipiens molestus [32]. The transinfected wPip strain was attributed to the wPip-IV incompatibility group (Mylene Weill, pers. com.) according to a classification by Atyame et al. [33]. Infection parameters (maternal inheritance and fitness costs), male mating competitiveness performances, and CI features (induction of complete and not age-dependent CI when wPip-transinfected males are crossed with wild-type females) were investigated both at laboratory level and on a semi-field scale [34,35]. The results of these studies were consistent with the traits needed to use the ARwP strain as the provider of ready-made sterility-inducer males for incompatible insect technique (IIT), which is an alternative autocidal approach based on the release of biologically incompatible males rather than irradiated males [15,36,37].

However, some major drawbacks have to be overcome, particularly against mosquito vector species, before IIT can be practiced in field. One of the major constraints is the inevitable co-release of females infected by a non-native Wolbachia strain that results in the release of incompatible males. This is because of the absence of an efficient sexing technology for the perfect separation of male and female pupae. In fact, at least 1% female contamination [38] is expected during the release of males. Model simulations based on laboratory experiments clearly predict that in cases where IIT strategy relies on a strong Uni-CI pattern (i.e., when the target population is uninfected), accidental co-release of Wolbachia-infected males and females may lead to an unwanted replacement of uninfected targeted population with the new infected population [39]. This would make IIT progressively ineffective. However, prediction becomes more complex when the target population harbors an incompatible Wolbachia infection type (Bi-CI), (as an example, if using the ARwP strain against the naturally infected Ae. albopictus). In this case, local coexistence of two different Wolbachia infection types may result in an unstable equilibrium that will evolve over time, resulting in the fixation of either one of the two infection types. According to Dobson et al. [39], this outcome would theoretically depend on two main factors: (i) pattern of CI (Uni-CI or Bi-CI) between the two populations and (ii) their competition both at larval and adult stages.

In this study, we provide new evidence on the first factor, i.e., CI pattern between naturally infected Ae. albopictus males (coinfected with wAlbA and wAlbB Wolbachia strains) and wPip-transinfected ARwP females, that we recently proved to be partially bidirectional [34]. Crosses between ARwP females and wild-type males were partially fertile when wild-type males were aged more than two weeks. On one hand, this weakness in the reproductive barrier between wild-type and transinfected mosquitoes may be convenient because it allows us to easily outcross the ARwP line with wild-type populations, thus restoring genetic variability periodically. However, the possible effects of this in field should be carefully considered.

First, we considered fundamental to provide additional data on wAlbA and wAlbB density in wild type Ae. albopictus from two sites for comparison with previous reports [40,41]. Next, we coupled the molecular determination of wAlbA and wAlbB bacterial titers in age-controlled superinfected males with the CI level observed in their crosses with ARwP females.

At last, comparison of laboratory results with the outcomes of Wolbachia density monitoring performed on captured wild-type males has been discussed as a necessary step for developing a bio-ecologically safe, long-term, and area-wide suppression strategy against Ae. albopictus based on the exploitation of Wolbachia-induced CI.

Materials and Methods

Ethics Statement

Research performed on invertebrates such as mosquitoes does not require a specific permit according to the directive 2010/63/EU of the European Parliament and of the Council on the protection of animals used for scientific purposes. Ae. albopictus is not an endangered or protected species, and no specific permissions are needed for collecting its eggs or adults in Italy. Samples were not collected from private or protected areas (see below for locations). Two of the authors (MC and RM) voluntarily used their arms for blood feeding during the experiments. According to the ethics committee of ENEA, this practice is not considered human experimentation.

Mosquito lines and rearing

This study included three Ae. albopictus populations: a wPip-transinfected population (ARwP) [32] and two naturally superinfected populations (SCRE and SANG) obtained from eggs collected in North and Central Italy, respectively. The first collection site is an urban area in Crevalcore in Bologna province (CRE: 44°43′12.10″N, 11° 8′54.94″E) while the second collection site is in Anguillara Sabazia, a suburban area 25 km north of Rome (ANG: 42°5′29.32″N, 12°16′17.57″E). The three mosquito populations were reared as described below.

Larvae were brought to adulthood in 1.5-l larval trays, at a density of 5 larvae/ml. Larval food was provided as described in Sinkins et al. [42]. Adult mosquitoes were kept in 40 × 40 × 40-cm cages placed in a climatic chamber (T = 27 ± 2C°, RH = 70 ± 10%, L:D = 14:10 hours) and were fed with sucrose solution (10%) soaked on cotton.

Age- and sex-specific Wolbachia density in SCRE and SANG mosquito lines

Adult males and females from the two natural Wolbachia-infected lines were isolated at the emergence and were pooled to be aged in the following six age groups (1–3, 4–6, 7–9, 10–15, 16–20, and >20 days). Males and females from each mosquito line and age group were analyzed by quantitative polymerase chain reaction (qPCR; see below) to evaluate the variation in the mean density of each Wolbachia strain with aging.

Crosses for CI strength assessment

Naturally infected males were crossed with 1-week-old ARwP females to determine eventual correlations between wAlbA and wAlbB densities and induced CI level. This study was planned regardless of the geographical origin; for this reason, only SCRE males were used.

In all, 50 virgin Ae. albopictus males belonging to 3 different age groups (3 ± 1, 11 ± 1, and 19 ± 1 days) were singly placed in 20 × 10 × 5-cm mating-oviposition cages with a single 1-week old ARwP virgin female (ARwP ♀ × SCRE ♂).

Because nuclear genes may be involved in generating incompatibility between populations [43], the ARwP strain was outcrossed for 5 generations with Wolbachia-cured SCRE males before setting up the CI experiments.

After copulation, the males from incompatible crosses were stored in ethanol and frozen for subsequent qPCR to determine the titer of Wolbachia strains of naturally infected Ae. albopictus. Once mated, females were fed with blood on human arms and isolated. Eggs laid from each single female were collected on oviposition devices made of wet strips of crepe paper and were stored in an incubator (temperature, 27°C; RH, 90%) for 5 days. The percentage of hatched eggs was used to compare CI levels with those obtained using fertile control crosses. Females whose eggs did not hatch were dissected to determine whether their spermathecae were filled or not with spermatozoa and in the latter case were excluded from the analysis.

Concurrently, ARwP compatible crosses (ARwP ♀ × ARwP ♂) were set up using 1-week-old virgin females and males belonging to the three age groups mentioned above. Ten crosses were performed for males belonging to each age group to determine the background embryonic mortality unrelated to CI and to compute CIcorr index (see below). In addition, 20 1-week-old virgin females were mated with males aged 3 ± 1 days and were then fed with blood. After egg laying, qPCR analysis was performed to quantify wPip Wolbachia titer and to ascertain whether wPip density and fertility in ARwP females were correlated in the above laboratory conditions.

Wolbachia density in wild caught males

Wolbachia titer of randomly captured wild males was surveyed at collection sites of the two studied populations (Crevalcore and Anguillara Sabazia in September 2013 and July 2014, respectively). Samples were collected using manual aspirators and by catching males flying around human operators. These Wolbachia density data were used to evaluate the risk of CI failure in crosses between wild-type males and ARwP females that were accidentally released or locally established after hypothetical IIT-based suppression programs.

Wolbachia genotyping and qPCR

DNA purification and qPCR

Total DNA was extracted from the whole body of a single Ae. albopictus mosquito by using ZR Tissue & Insect DNA Kit MicroPrep (Zymo) according to manufacturer's instructions. Strain-specific primers were used to amplify wsp. The wAlbA-wsp, wAlbB-wsp and wPip-wsp loci were amplified using previously described oligonucleotide primers pairs 328F/QArev2, 183F/QBrev2, wPF/wPR respectively, [34,40,44] to obtained 200-, 112- and 271-bp fragments, respectively.

Actin gene of Ae. albopictus was used as a nuclear reference and was amplified using primer pair actAlbqPCRsense (CCCACACAGTCCCCATCTAC) and actAlbqPCRantisense (CGAGTAGCCACGTTCAGTCA) to obtain a 119-bp amplification product.

Amplification reactions were performed using 20 μl of FluoCycle II SYBR Master Mix (Euroclone). Total DNA (2 μl) from each mosquito was used as a template for PCR and each reaction was performed in triplicate. PCR was performed in ABI Prism 7100 (Applied Biosystems) thermal cycler using the following amplification program: initial activation at 95°C for 5 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Presence of specific amplification products was verified using dissociation curves.

Construction of plasmids for obtaining qPCR standard curves

Specific DNA sequences encoding wAlbA-wsp, wAlbB-wsp, wPip-wsp and actin (quantitative reference) for qPCR amplification were cloned from the total DNA extracts. DNA fragments of wAlbA-wsp (382 bp) and wAlbB-wsp (501 bp) were amplified from the total DNA extracted from field-caught Ae. albopictus by using primer pairs 328F/691R and 183F/691R, respectively [44]. DNA extracted from Cx. pipiens was used as a template to amplify a wPip-wsp gene fragment (404 bp) with primers 183F and wPR. All the amplicons were cloned in pCR 2.1 vector plasmid (TA Cloning Kit, Invitrogen).

The amplified sequences were assembled to obtain the plasmids pBS-A-B-act (containing wAlbA-wsp, wAlbB-wsp and actin gene fragments) and pBS-Pip-act (containing wPip-wsp and actin gene fragments) by using the following procedure. Actin gene fragment was transferred from pCR 2.1 into BamHI-NotI sites of pBluescript II SK (+) vector to produce pBS-act plasmid. Next, wAlbB-wsp and wPip-wsp fragments were cloned from pCR 2.1 into NotI-SacI sites of the pBS-act plasmid to produce pBS-B-act and pBS-Pip-act plasmids. Finally, wAlbA-wsp fragment was cloned from pCR 2.1 into KpnI-XhoI sites of the pBS-B-act plasmid to produce pBS-A-B-act plasmid. All the obtained constructs were sequenced to assess the correct assembly and absence of unwanted sequence variations.

Statistical analysis and CI computation

Direct correlation between Wolbachia density (wAlbA and wAlbB) and mosquito age was investigated in both the sexes by using Spearman correlation test.

To test the correlation between CI level and wAlbA and wAlbB bacterial loads in males, a graph was plotted for Wolbachia density in each male used in each single-pair crossing experiment against observed CI expression.

Values of Wolbachia (wAlbA and wAlbB) density in each male used in incompatible crosses were grouped in Wolbachia density classes. For each cluster of Wolbachia density, we associated mean CI values found in crosses involving the corresponding males.

If the data set met the assumptions of normality (Shapiro–Wilk test, P > 0.05), one-way analysis of variance (ANOVA) was performed, followed by Bonferroni–Dunn multiple comparisons test. If the data set did not meet the assumptions of normality, non-parametric Kruskal–Wallis test was performed to determine whether this clustering highlighted significant differences between Wolbachia density in males and mean CI expressed in crosses. Dunn's test was used for pairwise comparison of mean wAlbA and wAlbB densities in males and CI values.

CI expression was calculated using egg mortality observed in each single-pair incompatible cross (ARwP ♀ × SCRE ♂) and was compared with the mean hatching rate (mean ± SEM) observed in single-pair compatible crosses (e.g., ARwP ♀ × ARwP ♂) by using CIcorr index [45]. This index does not overestimate the CI level caused by other mortality factors such as mean embryonic mortality observed in compatible crosses and male age effects that generally decrease fertility. The formula used for calculating CIcorr index is as follows: where UnE is the proportion of eggs that did not hatch in crosses between different infection types and UnE is the proportion of undeveloped embryos in compatible crosses. The value attributed to this last parameter was derived from ARwP compatible crosses performed using three classes of age-controlled males as described above.

All statistical analyses were performed using GraphPad Prism Software 6.0.1 version (GraphPad Software Inc., San Diego, CA, USA).

Results

Age- and sex-specific wAlbA and wAlbB densities

Ae. albopictus males showed remarkable individual variability in wAlbA density (Fig. 1A). In fact, wAlbA titers ranged from 0.0001–0.165 per actin copy number (wAlbA/actin). Complete loss of this Wolbachia strain was rarely observed in males (only 1 case out of 174); however, most males infected at low wAlbA densities (<0.001 wAlbA/actin) yielded negative results for standard PCR (S1 Fig.). Despite this high variability, we observed an evident reduction in the variation range with an increase in age. In fact, in young males (age, 1–3 days), wAlbA/actin ratios were widely different, ranging from <0.001 to >0.06 (mean ± SEM, 0.024 ± 0.005; N = 60). However, very low wAlbA levels (<0.001 wAlbA/actin) were detected more frequently in older males (age, >16 days), with a narrow range of variability (0.005 ± 0.002 wAlbA/actin; N = 35).

Fig 1

Age dependent variation of wAlbA density in Ae. albopictus males (a) and females (b) belonging to two Italian populations (SANG from Anguillara Sabazia, Rome: empty squares; SCRE from Crevalcore, Bologna: full circles).

Data trend is displayed via a polynomial trend-line together with the associated function (dashed line for SANG, solid line for SCRE).

Despite the high individual variability, a significant negative correlation was observed between age of males and wAlbA density (Spearman r = −0.35; P ≤ 0.0001).

In general, SCRE males were infected with lower mean wAlbA titer (0.010 ± 0.004 wAlbA/actin) than SANG males (0.020 ± 0.005 wAlbA/actin) irrespective of their age. However, this difference was not statically significant (ANOVA, F(1,94) = 1.28; P > 0.05).

In contrast to the trend observed in males, wAlbA density was positively correlated with age in females (Spearman r = 0.92, P ≤ 0.0001; Fig. 1B). The highest average wAlbA titer was observed in SANG females (2.420 ± 1.005 wAlbA/actin) and not in SCRE females (1.800 ± 0.680 wAlbA/actin); however, data were not sufficiently robust to highlight a statistically significant difference with respect to the geographical origin of the females (ANOVA, F(1,45) = 0.15; P > 0.05).

We confirmed that wAlbB was more abundant than wAlbA (t-test, P < 0.0001; Fig. 2). On an average, wAlbB was 40–50- and 7–9-fold more concentrated than wAlbA in males and females, respectively. In Ae. albopictus males (Fig. 2A), mean wAlbB titer increased with age (Spearman r = 0.38, P = 0.002). On the other hand, wAlbB density increased in females aged 10–15 days (Spearman r = 0.69; P ≤ 0.0001) and declined in older females (age, 19–21 days; Fig. 2B). Like wAlbA, wAlbB was more abundant on an average in both the sexes of the SANG population (0.80 ± 0.10 wAlbB/actin in males; 19.28 ± 3.76 wAlbB/actin in females) than in those of the SCRE population (0.52 ± 0.16 wAlbB/actin in males; 14.00 ± 2.81 wAlbB/actin in females). However, the differences were not statistically significant (ANOVA, F(1,24) = 1.09; P > 0.05).

Fig 2

Age dependent variation of wAlbB density in Ae. albopictus males (a) and females (b) belonging to two Italian populations (SANG from Anguillara Sabazia, Rome: empty squares; SCRE from Crevalcore, Bologna: full circles).

Data trend is displayed via a polynomial trend-line together with the associated function (dashed line for SANG, solid line for SCRE).

Effect of male wAlbA and wAlbB densities on CI expression

A general trend of significant decrease in the percentage of egg hatching related to male aging was observed in compatible ARwP ♀ × ARwP ♂ crosses by using 1-week-old females. In fact, mean percentages of egg hatching were 74.51 ± 4.89, 65.62 ± 3.22, and 49.10 ± 5.88 when males were 3 ± 1, 11 ± 1, and 19 ± 1 days old, respectively. Particularly, the oldest males were significantly less fertile than younger males (ANOVA, F(3,36) = 9,419; P < 0.0001).

Results of qPCR showed that fertility in compatible crosses was not correlated to wPip density in ARwP females (Spearman r = 0.012; P = 0.958; Fig. 3). Findings allowed us to exclude wPip titer as a factor determining egg mortality under the tested experimental conditions.

Fig 3

Correlation between percentage of egg hatching in ARwP × ARwP compatible crosses and wPip Wolbachia density in females.

Data trend is displayed via a polynomial trend-line together with the associated function.

Results of qPCR analysis relative to wAlbA and wAlbB densities in males from incompatible crosses are reported in Fig. 4 along with the observed CI levels (CIcorr index). Our results showed that induction of strong CI (CIcorr index ≅ 100) corresponded to males with wAlbA titers ranging from 0.01 to 0.11 wAlbA/actin (Fig. 4A). It is worth highlighting that almost all crosses showing incomplete CI (<80%) involved older males and some young males apparently harboring low wAlbA density (<0.001 wAlbA/actin) since emergence. Furthermore, old males with relatively high wAlbA titers (>0.01 wAlbA/actin) induced complete or approximately 100% CI.

Fig 4

Correlation between wAlbA (a) or wAlbB (b) Wolbachia density and expressed CI levels (CIcorr index) in Ae. albopictus males when crossed with ARwP females at three different ages (3 ± 1; 11 ± 1; 19 ± 1).

Data trend is displayed via a polynomial trend-line together with the associated function.

Male wAlbA densities were clustered into three classes <0.001, 0.001−0.010, and >0.010 wAlbA/actin, and their related mean CI levels were calculated (Table 1). The average CI level (CIcorr = 68.06 ± 6.20) induced by males with lowest wAlbA density was significantly lower than that in other classes (Kruskal–Wallis and Dunn multicomparison test P < 0.05).

Table 1

Density of wAlbA Wolbachia and associated CI level expressed by Ae. albopictus males when crossed with ARwP females.

wAlbA density classes(wAlbA/actin)	N	CI corr(mean ± SEM)
<0.001	17	68.06 ± 6.20*
0.001−0.010	17	95.57 ± 2.68
>0.010	16	98.08 ± 1.43

Density values have been clustered in three density classes.

* − statistically significant difference, by Bonferroni’s multiple comparison test with α = 0.05.

In contrast, male wAlbB titer showed no apparent correlation with CI penetrance in crosses involving ARwP females (Fig. 4B). To evaluate the specific role of wAlbB as a CI inducer, we restricted the correlation analysis to cases in which wAlbA density was low (<0.001 wAlbA/actin). By following this method, no significant correlation was observed between wAlbB concentration and CI expression (Fig. 5).

Fig 5

Correlation between wAlbB density and expressed CI level (CIcorr index) in Ae. albopictus males characterized by very low wAlbA titer when crossed with ARwP females.

Data trend is displayed via a polynomial trend-line together with the associated function.

Results of qPCR analysis of wild collected Ae. albopictus are reported in Fig. 6 based on the same wAlbA density classes defined in the previous experiment. In September 2013, approximately 50% males collected from Crevalcore were infected with very low wAlbA titers (0.0001−0.001 wAlbA/actin). The males belonging to the above density class decreased to 46.15% in samples collected in July 2014 in favor of males infected with higher wAlbA titers (>0.01 wAlbA/actin). Among Ae. albopictus males collected from Anguillara Sabazia, proportion of males with infection titers below the lowest Wolbachia density class (0.0001−0.001 wAlbA/actin) ranged between 36.36% in September 2013 and 30.77% in July 2014.

Fig 6

Density of wAlbA Wolbachia in wild caught Ae. albopictus males from two Italian sites (Crevalcore, CRE, Bologna; Anguillara Sabazia, ANG, Rome) and two periods (September 2013; July 2014).

Density values have been clustered in three density classes. The individuals belonging to each density class are reported as percentages of the whole amount. CI level is expected to decrease to about 68% in crosses between ARwP females and males with wAlbA density values <0.001 wAlbA/act.

Overall, males infected with wAlbA titers compatible with the expected complete or almost complete CI (>0.001 wAlbA/actin) ranged from 50.00% to 69.23% depending on the site and period.

Discussion

Various strategies exploiting Wolbachia infection are currently being investigated for vector control [16]. Of these, Wolbachia-induced CI has been proposed as a method to produce functionally sterile males that can be released in field for vector suppression, in accordance with the principles of SIT [13,14]. Wolbachia transinfection techniques have been used to establish new laboratory lines of important vector species to obtain biological traits suitable for this purpose [46]. Specifically, CI relationships with wild-type populations, male competitiveness in comparison with wild-type males, and fitness parameters contributing to efficient mass rearing should be studied carefully.

Nearly five years after its generation, the wPip-transfected line of Ae. albopictus (ARwP) is presently being evaluated for developing the most appropriate CI-based suppression strategy against this mosquito species. In previous studies, we have ascertained that ARwP line shows some desirable traits of an ideal IIT effector against Ae. albopictus (full CI, male mating competitiveness and female fitness not significantly different from wild-type Ae. albopictus) [32,34,35]. However, we also found that although CI between ARwP males and wild-type females was always complete, naturally infected males were not equally strong CI inducers towards ARwP females [34]. Moreover, although wild-type males were coinfected with wAlbA and wAlbB Wolbachia strains, only wAlbA strain determined a pattern of complete bidirectional incompatibility with wPip-infected females. In addition, male aging seemed to be the factor responsible for the reduction in CI level [34]. In this study, we analyzed wAlbA and wAlbB titers in age-controlled males collected from two sites and the associated CI levels after crossing these males with ARwP females. In fact, establishment of a correlation between Wolbachia density and CI would help in evaluating the risk of bidirectional CI failure by sampling wild-type males before hypothetical field releases.

Results of qPCR supported the results of previous studies [40,41], confirming that wAlbA is constantly maintained at a lower density than wAlbB both in males and females. In addition, density patterns of both the Wolbachia strains were strongly dependent upon sex, with females showing the highest densities and males exhibiting a dramatic decrease in wAlbA titers with age. Although we clearly confirmed that adult age was a fundamental factor influencing the density of both the bacterial strains, we realized that in general, the density of Wolbachia strains in naturally infected Ae. albopictus was an unpredictable individual feature that was partially related to the geographical origin of the population [40] and to environmental conditions (temperature and food availability) in which the larvae developed [42].

Differences observed at the population level were not significant in our study and did not involve the general trend of infection dynamics but confirmed that bacterial load may vary locally and depending on the period of the year. In fact, larger differences between populations have been observed in previous studies investigating wAlbA and wAlbB densities in Ae. albopictus populations from Reunion island, Greece, and Corsica [40].

Correlation studies between CI and Wolbachia density have shown that incompatibility level is positively related to male wAlbA density. However, this correlation is not linear. In fact, only males with very low wAlbA titers (<0.0010 wAlbA/actin) induced significantly lower levels of CI while those with higher wAlbA density did not show CI weakening. These findings relative to wAlbA partially agree with the model proposed by Breeuwer and Werren [47], which states that decrease in CI penetrance in old males is directly proportional to Wolbachia density in the testes or sperm cysts in general [45,48-56]. Notably, a marked decrease in incompatibility level was previously observed when 10-day-old wAlbA mono-infected males were crossed with Wolbachia-free females [41].

In contrast, wAlbB, the most abundant Wolbachia strain in superinfected males, did not show any correlation with the observed CI in crosses with ARwP females. In fact, the observed decrease in the CI level countered the initial increase in wAlbB density in aging males, which was possibly mediated by a concurrent decrease in wAlbA density. Moreover, in males with very low wAlbA titers, CI expression did not change in response to changes in wAlbB density. These results agree with those of a previous study, which showed that crosses between ARwP females and wAlbB or wAlbA mono-infected males were approximately 20% fertile and completely unfertile, respectively [34].

Previous studies have shown that Wolbachia infected various host tissues but was mostly (at least 20 folds more) concentrated in Ae. albopictus gonads [57]. Particularly, in Koh Samui and Mauritius strains, mono-infected with wAlbA, Wolbachia cannot be detected in non-reproductive tissues [58]. In addition, wAlbA and wAlbB strains showed a highly similar tissue tropism [57,58]. Thus, although our data concerned total body molecular quantification, we can confirm that wAlbA plays the main role in determining the complete bidirectionality in CI pattern occurring between native and ARwP Ae. albopictus. Under these premises, ascertaining wAlbA titer in wild collected males is crucial because changing environmental conditions may cause Wolbachia density to differ significantly. In fact, analysis of males from Crevalcore and Anguillara sites showed very low wAlbA titer in almost half of all the males (on average 45% falling in <0.0001 wAlbA/actin density group), which was in accordance with previous field data [40]. This means that ARwP females could be fertile not only with ARwP males but also (at least partially) with some wild-type males, presumably gaining a reproductive advantage compared with wild-type females.

Therefore, the unintentional but repeated release of ARwP females during IIT should be approached with caution for a series of reasons.

First, the lack of a sexing system that guarantees the total absence of females in the released population increases the vulnerability of any autocidal application against vector mosquitoes because released females can blood feed and transmit diseases [13,15]. This vulnerability further increases when we apply the IIT strategy because accidentally released females can mate with the released males and reproduce. In the specific case of the ARwP strain, the weakness of the bi-CI, that we have highlighted in this work, must be considered as further factor affecting the IIT failsafe.

On the other hand, based on the modeling approach [39,59], presence of two reciprocally completely incompatible mosquito populations at one site is desirable because the competition decreases the number of biting females expected in the long-term for a single population. In addition, a recent paper reported about the stable coexistence in a site of two molecular Wolbachia strains of Culex pipiens, characterized, like ARwP and wild Ae. albopictus, by partially bidirectional CI relationships [60] However, the outcome of a partial failure in the bidirectionality of the CI pattern between ARwP and wild-type Ae. albopictus has not yet been properly investigated by experimental validation of specific theoretical models.

For the reasons above stated, awaiting for the advent of a very efficient sexing system, other strategies could be developed to exploit the favorable properties of the ARwP strain.

An advancement of the IIT-based suppression strategy could be the concurrent or sequential release of males belonging to two reciprocally incompatible lines. This approach could drastically reduce the possibility of establishing a transinfected line because of incompatible crosses that would occur in most cases [15]; however, this should be experimentally validated. Suitable Ae. albopictus lines are already available [61,62]; nevertheless, other lines can be specifically established for this purpose.

Furthermore, an alternative strategy using a combination of IIT and SIT involving irradiation of incompatible males at radiation doses just high enough to induce sterility in any females that are not removed but not sufficient to affect male fitness should be considered [16,63,64]. This approach is expected to be promising if, life most insect species, Ae. albopictus females are more radiosensitive than males [65]. An important advantage of this modified IIT strategy over the traditional SIT strategy would be the absence of any residual male fertility, necessarily tolerated by the latter strategy to save male mating competitiveness [66,67].

As long as a method to release large amounts of only males is available and/or specific population dynamic models are experimentally validated, the latter strategy seems to be the easiest to test and propose. Overall, the goal of field application of the ARwP line to control Ae. albopictus based on IIT/SIT approaches seems achievable. However, as shown in this study, a series of necessary steps need to be considered to ensure that the application strategy is effective and safe in the long-term.

Evaluation of PCR sensitivity in relation to increasing wAlbA titers, calculated as wAlbA/actin copy numbers by qPCR on independent mosquito extracts: (1) 0.0001; (2) 0.0010; (3) 0.0039; (4) 0.0108; (5) 0.0582; (6) 0.1671; (7) 2.2938; (8) no template PCR control.

For wspA gene amplification 328F and 691R oligonucleotides were used, according to the following PCR program: 95°C for 3 min; then 35 cycles at 95°C for 30 s, 55°C for 30 s and 72°C for 35 s; finally an elongation step at 72°C for 7 min. The expected wspA amplicon was of 382 bp.

(TIF)

Click here for additional data file.