Mosquito larvae experience multiple environmental stressors that may modify how subsequent adults interact with pathogens. We evaluated the effect of larval rearing temperature and intraspecific larval competition on adult mosquito immunity and vector competence for Sindbis virus (SINV). Aedes aegypti larvae were reared at two intraspecific densities (150 and 300 larvae) at 20° C and 30° C and the adults were fed artificially on citrated bovine blood containing 105 plaque forming units of SINV. Expression of cecropin, defensin, and transferrin was also evaluated in one- and five-day-old female adults. There was a direct relationship between larval density and SINV infection and dissemination rates at low temperature (20° C) and an inverse relationship between larval density and SINV infection rate at high temperature (30° C). Cecropin was only expressed in five-day-old adults that were raised at high temperature as larvae and was 20-fold over-expressed at low compared to high density treatments. Defensin and transferrin were under-expressed in one-day-old adults and over-expressed in five-day-old adults in all competition-temperature combinations relative to low density treatments at 20° C. These findings suggest that interaction between biotic and abiotic conditions of the larval environment may alter adult mosquito immunity resulting in enhanced vector competence for arboviruses.
The global expansion of mosquito vectors of arthropod-borne viruses (arboviruses) including Aedes aegypti, Aedes albopictus, and Culex pipiens is a serious threat to international health. These species are changing the epidemiology of arboviral diseases in their new geographical regions by introducing novel arboviruses or becoming novel vectors of native or independently introduced arboviruses (Gubler 1997, Juliano and Lounibos 2005). The recent outbreaks of dengue virus in South and Central America (Gubler 2002), Chikungunya virus in Italy and Indian Ocean islands (Ligon 2006, Charrel et al. 2007), and West Nile virus in North America (CDC 2008) are examples of threats posed by invasive mosquito species and associated arboviruses. Currently, vector control is the most effective means of interrupting transmission of arboviruses since there are no effective antiviral therapies and vaccines are only available for a limited number of arboviruses. Unfortunately, the full potential of vector control is yet to be achieved in part due to lack of adequate knowledge of the complex interactions among the pathogens, vectors, hosts, and their environment.
Mosquitoes exhibit a complex life cycle with aquatic immature stages and an adult stage that disperses to an ecologically different terrestrial environment. The disparate ecological niche occupied by immature and adult stages exposes them to entirely different environmental conditions that may have variable effects on adult mosquito fitness and vector competence. A significant amount of literature is available on how environmental conditions experienced by the adults affect their fitness (Braks et al. 2006, Vrzal et al. 2010, Xue et al. 2010) and competence for pathogens (Turell 1989, 1993, Richards et al. 2007). In contrast, studies investigating the impact of larval environment on pathogen transmission are relatively rare but have been increasing in recent years. These studies have consistently shown that the effects of larval environment carry over to the adult stage and influence their fitness and competence for pathogens (Grimstad and Walker 1991, Reiskind and Lounibos 2009, Muturi et al. 2011).
Intraspecific larval competition is one of the numerous environmental stressors that are likely to influence adult mosquito fitness and competence for pathogens. This competition is manifested directly through interference or indirectly through resource exploitation (Blaustein and Chase 2007, Murrell and Juliano 2008). The effects of larval competition operate at the individual level and often impact negatively on individual survival, development, size, longevity, reproduction (Moore and Fisher 1969, Black et al. 1989, Murrell and Juliano 2008, Reiskind and Lounibos 2009, Muturi et al. 2010), and immunity (Suwanchaichinda and Paskewitz 1998). These effects are translated into population level effects by adversely impacting on one or more population processes that reduce the per capita growth rate such as reproduction and survival (Juliano 1998). An increasing number of studies suggest that larval competition can affect how subsequent adults interact with pathogens. Aedes albopictus (Alto et al. 2005) and Aedes aegypti females (Muturi et al. 2011) from high-density treatments were more susceptible to Sindbis virus (SINV) than females from low-density treatments. Similarly, the presence of Ae. albopictus resulted in large Aedes triseriatus females that were more susceptible to LaCrosse virus than females from monospecific cultures (Bevins 2008). In contrast, large Ae. aegypti females from varying conditions of larval density and resources were more susceptible to dengue 2 virus than smaller females (Sumanochitrapon et al. 1998). In nature, however, larval competition operates in the presence of other environmental factors that may suppress, enhance, or even reverse its effects on vector competence. These factors include chemical contaminants (Muturi et al. 2011), predators (Juliano et al. 2009), desiccation (Costanzo et al. 2005), and type and amount of larval resources (Yee et al. 2007, Murrell and Juliano 2008).
Temperature of the larval environment can also alter the outcome of intraspecific competition on vector fitness and competence for pathogens through effects on larval growth and development, larval and adult survival, adult size, blood feeding behavior, fecundity (Bayoh and Lindsay 2004, Vinogradova and Karpova 2006), pathogen replication (Reeves et al. 1994), and extrinsic incubation period (Kay et al. 1989, Kay and Jennings 2002). Within the range at which mosquito survival is possible, higher temperature leads to shorter development times and small adults (Tun-Lin et al. 2000, Bayoh and Lindsay 2004) that are more (Muturi and Alto 2011) or less competent for arboviruses (Kay et al. 1989, Kay and Jennings 2002, Westbrook et al. 2010). In cases where high temperature and intraspecific competition are associated with enhanced vector competence, exposure of mosquito larvae to both stressors may intensify susceptibility to infection. In contrast, if temperature is inversely associated with vector competence, it may suppress the impact of intraspecific competition on vector competence. Enhanced infection, dissemination, and transmission of arboviruses is often attributed to the breakdown of “midgut and salivary gland infection and escape barriers” (Grimstad and Walker 1991, Hardy et al. 1983), but the underlying molecular mechanisms are poorly understood.
In our previous study, we found that Ae. aegypti adults from high larval rearing temperatures were more susceptible to Sindbis virus (SINV, Togaviridae: Alphavirus) than adults from low larval rearing temperatures (Muturi and Alto 2011). Using this model system, we tested the prediction that temperature alters the effects of intraspecific larval competition on Ae. aegypti susceptibility to SINV. We also attempted to elucidate the relationship among enhanced vector competence and baseline expression levels of three immune-responsive genes; cecropin, defensin, and transferrin. These genes are induced by various types of microorganisms and are therefore ideal targets for quantifying the baseline immune status (Hultmark 1993). Defensins and cecropins are two of the three families of antimicrobial peptides (AMPs) identified in mosquitoes (Waterhouse et al. 2007). Defensins are active against gram positive bacteria but are also induced by other types of pathogens including viruses, fungi, and parasites (Lowenberger et al. 1995, Cheng et al. 2001, Magalhaes et al. 2008). Cecropins have lytic activity against both gram positive and gram negative bacteria (Lowenberger et al. 1999), while transferrin is an iron-binding protein involved in iron transport (Nichol et al. 2002) and has antibiotic activity in insects (Yoshiga et al. 1997).
MATERIALS AND METHODS
Life history and vector competence studies
The experiments were conducted using F10 progeny of Ae. aegypti from field collections in Florida. Approximately 12-hour-old 1st instar larvae were reared in 1.6 liters of oak infusion at two intraspecific larval densities (150 and 300 larvae) incubated at 20° C or 30° C. The oak infusion was prepared by fermenting 500 g of senescent leaves of live oak (Quercus virginiana) in 160 liters of tap water for two weeks. Each treatment was replicated four times and supplemented with 0.2 g of yeast: albumin during experimental setup and 0.1 g of the same food eight days later. The containers were monitored daily and pupae were transferred into cotton-sealed plastic vials with de-ionized water until adult emergence. The total number of male and female mosquitoes emerging each day was recorded and both sexes were housed together in paperboard cages at 25° C and provided continuous access to 10% sucrose.
Five to ten-day-old females were allowed to blood-feed artificially on SINV-infected bovine blood (MRE16 strain from Sindbis Health District, Nile Delta, Egypt) as previously described (Muturi and Alto 2011). In brief, adults were sugar-starved for 48 h prior to providing them a 30 min access to citrated bovine blood containing 105 plaque forming units/ml of SINV via the Hemotek membrane feeding system (Lancashire, UK). Blood-fed females were maintained at 25° C in paperboard cages for a 14-day virus incubation period and provided continuous access to 10% sucrose and an oviposition cup lined with a moist paper towel. Unfed females were provided a second and third opportunity to blood feed in a 20-h interval after which all unfed females were killed and their wings measured as an indicator of size. After a 14-day extrinsic incubation period, the females were killed and their wings measured as an indicator of size while the bodies and legs were preserved separately at –80° C and later used to determine SINV infection and dissemination rates, respectively, by real time RT-PCR. The RNA extraction and RT-PCR protocol including primers and probe information has been described in detail elsewhere (Muturi and Alto 2011). Infection rate was the percent of mosquitoes with SINV-infected bodies out of the total fed. Dissemination rate expressed as the percent of mosquitoes with SINV-infected legs out of the total fed was used to estimate the vector competence of the population.
Data were checked for normality and homogeneity of variance and analyzed using the SAS statistical package (SAS Institute 2002). Two-way Analysis of Variance was used to test for the effect of temperature and larval density on survivorship to adulthood (females + males). Where significant differences were detected, the means were separated by Tukey-Kramer test. Separate MANOVAs were used to test for treatment effects on female life history traits (development time and size) and infection parameters (SINV infection and dissemination rates). For each significant treatment effect, standardized canonical coefficients were used to determine the relative contribution and relationship among life history traits and between infection parameters (Scheiner 2001).
Gene expression studies
One- and five-day-old adult females derived from identical experiments as described above were used to determine the impact of larval rearing temperature and intraspecific larval density on expression of antimicrobial peptides (defensin and cecropin A) and transferrin, an iron-binding protein that also serves as an antimicrobial agent in insects. The five-day-old females were maintained on 10% sucrose. RNA from these females was extracted using a Qiamp® virus Biorobot® 9604 kit according to the manufacturer's protocol (Qiagen). RNAs were quantified using nanodrop readings and 500 ng total RNA were treated with DNase 1 (Biolabs) in 20 μl aliquots using the manufacturer's protocol. qRT-PCR was conducted in 7300 real time PCR system (Applied Biosystems) in 20 μl reaction volumes containing 10 μl 2× one-step RT-PCR master mix, 0.5 μl 40× multiscribe and RNase inhibitor mix, 0.5 μl of each 10 μM forward and reverse primer stock, 1 μl 20× SYBR Green dye, 5.5 μl double distilled water, and 2 μl template RNA. RT-PCR thermocycling conditions were 50° C for 60 min, 95° C for 10 min followed by 40 cycles of 95° C for 15 s, 55° C for 30 s, and 72° C for 30 s. The primers used were as follows; defensin; forward 5-GCCACCTGTGAT-CTGCTGAGCGGA-3, reverse 5-GGAGTTGCAGTAGCCTCCCCGAT-3, cecropin A; forward 5-ATTTCTCCTGATCGCCG-TGGCTG-3, reverse 5-GAGCCTTCTCGGCGGCATTGAA-3, and transferrin; forward 5-CGGCCAGCTGGAGGATAACATTG-3, reverse 5-CGCTTTGAA-CTCGCCGAACATCTC-3. Relative gene expression was performed according to 2–DDCT method (Livak and Schmittgen 2001) and normalized using ribosomal protein L8 (AERPL8). For all treatments, four biological samples were analyzed in three replicates. Because mosquitoes perform better at low rather than high larval densities (Reiskind and Lounibos 2009) and at low rather than high temperatures (Bayoh and Lindsay 2004), we chose to use a low-density treatment at low temperature (150, 20° C) as our control. This was achieved for defensin and transferrin, but since cecropin was not expressed at low temperature, we used low-density treatments at high temperature (150, 30° C) as the control for this gene. Results are expressed as mean expression ratios (± SE) between “treatments” and “controls” with genes showing at least a two-fold over- or under-expression considered to be differentially expressed.
Life history and vector competence studies
Survival was significantly higher at low rather than high larval density (F = 12.47, df = 1, 12, P= 0.004; Table 1) but was not influenced by temperature or temperature by density interactions. There was significant temperature by density interactions on female development time to adulthood and size with the former contributing the larger portion of observed variation (Table 2). The standardized canonical coefficients were positive, indicating that development time and size responded similarly to treatment combinations. High temperature resulted in smaller mosquitoes but negative density-dependent effects on female sizes were only observed at low temperature (Table 1). Low larval density and higher temperature resulted in shorter development time to adulthood (Table 1).
Table 1. Least square means (± standard error) for effect of temperature and intraspecific larval competition on Aedes aegypti survival to adulthood (male and females combined), female development time and female wing lengths. n values are presented in brackets. Means followed by different lower case letters show significant differences for pair-wise comparisons.
Development time (days)
Wing length (mm)
87.76 ± 1.60a (600)
15.35 ± 0.19a (216)
2.97 ± 0.02a (211)
77.08 ± 6.12b (1200)
17.38 ± 0.16b (355)
2.71 ± 0.02b (340)
86.00 ± 2.48a (600)
8.58 ± 0.09c (192)
2.48 ± 0.01c (192)
70.75 ± 2.71b (1200)
11.49 ± 0.14d (345)
2.46 ± 0.02c (337)
Table 2. MANOVA results for the effect of temperature and intraspecific larval competition on Aedes aegypti life history traits (size and development time) and vector competence for SINV (infection and dissemination rates).
Standardized canonical coefficients (SCC)
Life history traits
T × D
T × D
Significant temperature and density interactions were observed for SINV infection and dissemination rates with the former contributing to most of the variation (Table 2). There was a positive relationship between larval density and SINV infection and dissemination rates at low temperature and an inverse relationship between larval density and SINV infection rates at high temperature (Figure 1). Larval density had no significant effect on SINV dissemination rate at high temperature (Figure 1).
Gene expression studies
Cecropin was only expressed in five-day-old adults derived from immatures reared at high temperature and was 20-fold over-expressed at low rather than high-density treatments. Relative to low-density treatments at 20° C, defensin and transferrin were under-expressed in one-day-old adults and over-expressed in five-day-old adults in all competition-temperature combinations (Table 3).
Table 3. Relative expression of defensin and transferrin in Aedes aegypti adults in response to temperature and intraspecific larval competition. Expression levels were normalized with housekeeping gene AERPL8 and are shown as expression ratios relative to the controls (low density treatments (150) at 20° C). Genes that were two-fold over-expressed or under-expressed relative to the controls were considered to be differentially expressed.
0.14 ± 0.05
0.01 ± 0.00
83.12 ± 14.21
68.47 ± 11.02
62.15 ± 9.31
478.05 ± 27.88
29.13 ± 6.35
30.85 ± 7.81
Intraspecific larval competition impacts negatively on individual and population performance (Reiskind et al. 2004, Reiskind and Lounibos 2009) and may also enhance susceptibility of subsequent adults to arboviruses (Alto et al. 2005, Bevins 2008). However, the direction and/or the magnitude of these effects may be altered by other environmental factors. The current study evaluated whether larval environmental temperature can modify the impact of intraspecific larval competition on adult mosquito life history traits, immunity, and vector competence for SINV. The Aedes aegypti-Sindbis virus model system is known to be suitable for studying how mosquitoes interact with other Alphaviruses such as Chikungunya and Eastern Equine Encephalitis (Myles et al. 2004).
Our findings suggest that temperature can modify the outcome of intraspecific competition. As previously observed, survival to adulthood decreased with increasing larval density (Ho et al. 1989, Juliano 1998) but was unaffected by larval-rearing temperature (Lounibos et al. 2002, Muturi and Alto 2011). In contrast, temperature altered the effect of intraspecific larval competition on development time and size of resulting adults. Relative to low-temperature treatments, high temperature resulted in a 44% and 34% reduction in development time to adulthood at low and high larval density, respectively. In addition, negative density-dependent effects on female wing lengths were detected at low but not at high temperatures. Accelerated development of mosquito larvae at high temperatures may interfere with nutrient intake and accumulation nullifying any nutritional advantage that low-density treatments may have over the high-density treatments (Korochkina et al. 1997). This may partly account for the lack of density-dependent effects on adult size at this high temperature.
The effect of intraspecific larval competition on SINV infection and dissemination rates was contingent upon the larval rearing temperature. At low temperature, SINV infection and dissemination rates were consistent with previous findings that larval competition enhances vector competence for arboviruses (Alto et al. 2005, Alto et al. 2008, Bevins 2008). However, at higher temperatures, high larval density resulted in lower SINV infection rates relative to low-density treatments despite both groups having similar-sized adults. These findings suggest that at high temperature, similar-sized adults in the low- and high-density treatments were produced by different mechanisms that influenced their vector competence differently, confirming previous observations that mosquito size is not necessarily a good indicator of vector competence (Muturi et al. 2011). In addition, our findings are inconsistent with the previous argument that larger mosquitoes from low-temperature treatments are likely to imbibe larger volumes of blood and are thus more likely to be infected compared to small mosquitoes from high-temperature treatments (Westbrook et al. 2010).
In some mosquito-arbovirus systems such as Culex tarsalis and Western equine encephalitis virus (Hardy et al. 1990), Culex annulirostris and Murray Valley encephalitis virus (Kay et al. 1989), Aedes vigilax and Ross River virus (Kay and Jennings 2002), and Aedes albopictus and Chikungunya virus (Westbrook et al. 2010), vector competence was significantly lower in females derived from higher rather than from lower larval-rearing temperatures. In contrast, Ae. aegypti females from higher larval-rearing temperatures were more competent for SINV than females from lower larval-rearing temperatures (Muturi and Alto 2011). These findings, along with our current findings, suggest that there may be other temperature- and density-dependent physiological and/or morphological alterations that act on mosquitoes to enhance their vector competence for arboviruses. Apparently, our findings suggest that conditions of the larval environment may partly explain the large spatial and temporal variations in vector competence and arbovirus transmission observed within small geographical regions (Kilpatrick et al. 2010).
Little is known about the mechanisms by which larval environmental factors enhance vector competence for arboviruses. In adults from resource-deprived larval environments, this has been attributed to reduced thickness of the basement membrane of the midgut and other organs (Grimstad and Walker 1991). We attempted to elucidate how the larval-rearing temperature and intraspecific larval competition may alter adult mosquito immunity to pathogens by quantifying the baseline levels of three immune-related genes (defensin, cecropin A, and transferrin) in one- and five-day-old females. For all temperature and density combinations, defensin and transferrin were under-expressed in one-day-old females and over-expressed in five-day-old females relative to low-density treatments at low temperatures. Conversely, cecropin was expressed only in high-temperature treatments and in much higher levels at low rather than high larval density. These findings suggest that conditions of the larval environment can alter adult mosquito immunity with detrimental effects on vector competence. In a recent study, a gene encoding a putative anti-bacteria cecropin-like peptide was shown to possess anti-dengue and anti-Chikungunya viral activity (Luplertlop et al. 2011). Because we measured the baseline immune status of mosquitoes prior to arbovirus exposure and not an immune response to the arbovirus infection per se, we cannot directly link the observed up regulation of immune genes to the vector competence results. Future studies assessing tissue-specific expression of immune genes in mosquitoes following acquisition of an infectious blood meal will unravel the relationship between larval stress, expression of immune peptides, and vector competence. Gene silencing by RNA interference could also unravel the role of specific immune genes in defense against arboviruses (Xi et al. 2008).
Because the three immune peptides are responsive to microbial exposure (Lowenberger et al. 1995, Yoshiga et al. 1997, Lowenberger et al. 1999, Waterhouse et al. 2007, Magalhaes et al. 2008), our results suggest that stress due to temperature and larval competition may predispose the mosquitoes to opportunistic bacterial and fungal pathogens with detrimental effects on vector competence. Opportunistic bacteria and fungi may enhance vector competence in two major ways. First, mosquito investment in immune responses against bacterial and fungal infections is costly (Fellous and Lazzaro 2010) and may limit the amount of energy available for defense against subsequent pathogens. This cost was clearly illustrated using Ae. aegypti cell lines where prior immune challenges with Escherichia coli resulted in higher titers of dengue 2 virus (Sim and Dimopoulos 2010). Alternatively, elevated production of immune peptides in response to opportunistic microorganisms under stressful conditions may eliminate the normal midgut microbial flora that serves as a barrier to infection, resulting in enhanced vector competence. Previous studies have shown that mosquito microbiota induces a basal level immune activity that enhances resistance to pathogens (Beier et al. 1994, Hoffmann et al. 1999, Mourya et al. 2002). For instance, Anopheles mosquitoes became more competent for the malaria parasite Plasmodium falciparum when endosymbiotic bacteria were disrupted with antibiotics (Hoffmann et al. 1999). Similarly, the titers of dengue virus were two-fold higher in Ae. aegypti mosquitoes that were reared aseptically (in the absence of their endogenous bacterial flora) compared to the wild type mosquitoes (Xi et al. 2008). Transferrin, however, has dual a role as an antimicrobial agent as well as an iron-binding protein (Yoshiga et al. 1997, Nichol et al. 2002). It is therefore unclear from our results whether its expression pattern was induced by microorganisms or by oxidative stress. Further, we only assessed gene transcription and thus we cannot ascertain whether or not protein translation occurred. Future studies should address these research gaps.
We thank M. Dmitrieva, N. Krasavin, H. Kyrias, and J. Ricci for assistance with daily maintenance of the experiment and R. Lampman and P. Otienoburu for their helpful comments. We thank L.P. Lounibos for providing Ae. aegypti eggs used to start laboratory colonies for this research. This study was supported by the Illinois Waste Tire and Emergency Public Health Funds.