Background

Species within the genus Anopheles (Diptera: Culicidae) play a major role in the transmission of malaria in Africa, in particular mosquitoes from the An. gambiae complex and the An. funestus group[1, 2]. The An. gambiae complex contains excellent and efficient vectors of malaria (An. gambiae s.s. and An. arabiensis), as well as minor vectors (An. merus) and non-vectors (An. quadriannulatus species A and B)[1]. The An. funestus group contains an important vector of malaria, An. funestus s.s.[2].

Although chemical insecticides have been used successfully in integrated vector control programs[3], many malaria vector control programs are hampered by the development of resistance of the vectors to chemical insecticides[4–6]. In addition to development of resistance, concerns about the non-target effects and persistence of the chemical insecticides have prompted the development of environmentally friendly control agents and control programs[7].

Bacillus thuringiensis subsp. israelensis (Bti) is a Gram-positive, aerobic, spore-forming, bacterium that produces crystalline inclusions that contain crystal (Cry) or cytolytic (Cyt) proteins that are highly toxic to mosquito larvae[8, 9]. Although there is low risk of resistance being developed to Bti[10], there are several disadvantages to using Bti as a control agent[11, 12]. These include its low persistence in the field due to inactivation by UV, ingestion of Bti by other aquatic organisms, and the settling of Bti from the mosquito larval feeding zone[11–13]. One strategy to overcome some of the disadvantages of Bti is to clone the cry genes of Bti into aquatic microorganisms that: (1) are not toxic to other organisms, (2) inhabit and persist in the larval feeding zone, (3) are used by mosquito larvae as a food source, (4) express Cry proteins at levels that are mosquito larvicidal, and (5) have cell walls that reduce inactivation of the Cry proteins by UV[13–15].

Xiaoqiang et al.[15] inserted the Bti cry4Aa, cry11Aa, and p20 genes under the control of two tandem promoters (cyanobacterial constitutive promoter, P psbA , and Escherichia coli T7 early promoter, P A1 ) into a filamentous nitrogen-fixing cyanobacterium, Anabaena sp. strain PCC 7120 (PCC 7120). The Bti genes are integrated into the chromosome of PCC 7120, resulting in a stable recombinant strain[16]. Laboratory bioassays have shown that the resultant recombinant strain, PCC 7120#11, is a very effective larvicidal agent against Aedes aegypti[13, 15, 16].

To our knowledge, no studies have examined the larvicidal activity of PCC 7120#11 against several African malaria vectors. The aim of this study was, thus, to evaluate the larvicidal activity of PCC 7120#11 against five African Anopheles species in order to determine if PCC 7120#11 may have potential as a malaria vector control agent.

Methods

Larvicidal activity of PCC 7120#11 was determined by laboratory bioassays against four species in the An. gambiae complex and one species in the An. funestus group. The An. gambiae complex species used in the study were (origin, colony name, and colonisation date provided): An. gambiae s.s. (Ibadan, Nigeria; NAG; 2001), An. arabiensis (Kanyemba, Zimbabwe; KGB; 1975), An. merus (KwaZulu-Natal, South Africa; MAF; 1988), and An. quadriannulatus species A (Sangwe, Zimbabwe; SANGWE; 1998). The species from the An. funestus group that was used in the study was An. funestus s.s. (Maputo, Mozambique; FUMOZ; 2000).

The anopheline mosquito species were obtained from colonies maintained at the National Institute of Communicable Diseases (Johannesburg, South Africa). Since the activity of PCC 7120#11 against A. aegypti had been previously evaluated[15], we included it as a control in the bioassays. The A. aegypti larvae were obtained from the South African Bureau of Standards (Pretoria, South Africa).

PCC 7120 and PCC 7120#11, were cultured in BG-11 medium[15], at 30°C under continuous illumination (2000 lux) with constant agitation[13, 16]. The PCC 7120 and PCC 7120#11 cells were harvested by differential centrifugation and the cell concentration was determined by haemocytometer counts. Two millilitres of the appropriate dilution (covering an in-cup concentration range of 1.00 × 104 to 3.20 × 107 cells/ml) of either PCC 7120 or PCC 7120#11 was added to 130 ml plastic cups that contained 98 ml sterile distilled water and 20 third-instar mosquito larvae. In the case of An. merus, a sterile 5 M saline solution was used instead of sterile distilled water.

Larvicidal activity was determined 24 hours post-inoculation, with larvae presumed dead if they did not move when prodded. An untreated control (sterile distilled water) was included in the bioassays. If mortality in the controls (PCC 7120 and untreated) exceeded 5%, the test was discarded and repeated. Each bioassay was repeated in triplicate on different days. The lethal concentration (LC50 and LC90) for each species was determined by probit analysis[17]. For each species, probit analysis was based on the mortality data obtained from five PCC 7120#11 concentrations.

Results and discussion

The concentration-mortality data for the mosquito species are summarised in Table 1. The heterogeneity factors for the different mosquito species evaluated were all less than one, indicating a good fit of the concentration-mortality data to the probit model[18].

Table 1 Probit analysis of concentration-mortality data for Anabaena PCC 7120#11 against third instar mosquito larvae
Full size table

The concentration-mortality regression slopes indicate the variability in response to a toxin within the vector population being examined[19]. In this study, An. arabiensis had a significantly steeper slope than the other species evaluated, suggesting that An. arabiensis had lower response variability or reduced heterogeneity in its population compared to the other mosquito species examined. The shallower slopes of the concentration-mortality regression lines obtained for An. gambiae and An. quadriannulatus mean that there are larger differences between the LC50 and LC90 values for these species than for the other anopheline species evaluated.

Although slight variation in LC50s between studies may be expected due to differences in experimental conditions such as rearing conditions of the larvae and natural variations in the larval populations[20], the susceptibility of A. aegypti larvae to PCC 7120#11 in this study was comparable to that (LC50 of 0.9 × 105 cells/ml) previously reported by Xiaoqiang et al.[15]. However, the LC50 value of PCC 7120#11 against A. aegypti larvae was significantly lower than those of the anopheline species evaluated (Table 1). The decreased susceptibility of anopheline larvae compared to A. aegypti larvae may be due to differences in feeding behaviour[21]; Anopheles larvae (surface feeders) may ingest fewer PCC 7120#11 cells than A. aegypti larvae that feed at all levels in the water column.

The LC50 values differed significantly between the anopheline species examined, with PCC 7120#11 being most effective against An. merus larvae (Table 1). On the basis of LC50 ratios, the larvae of An. quadriannulatus were 4-fold less susceptible to PCC 7120#11 than the larvae of An. merus. PCC 7120#11 had relatively weak larvicidal activity against An. funestus, even at concentrations as high 3.2 × 107 cells/ml. The LC90 values for An. merus and A. aegypti were significantly lower than the LC90 values for An. arabiensis, An. quadriannulatus, or An. gambiae (Table 1). Furthermore, the LC90 for An. arabiensis was significantly lower than that of An. quadriannulatus or An. gambiae (Table 1).

The significant differences in susceptibility to PCC 7120#11 between the anopheline species may reflect species-specific differences in one or more of the steps in the ingestion-to-toxicity process. For example, differences in the ingestion rate or efficiency of digestion of PCC 7120#11 cell walls by the different anopheline species may significantly affect the concentration of Cry proteins in the larval midgut. However, preliminary evaluations (unpublished data) did not show marked differences in the ingestion or digestion rates between An. funestus and An. arabiensis.

A factor to consider in interpreting the significant differences in susceptibility between An. funestus and species of the An. gambiae complex is the combination of the toxins present in PCC 7120#11. Bti, which naturally produces crystals that contain the toxins Cry4Aa, Cry4Ba, Cry10Aa, Cry11Aa, Cyt1Aa and Cyt2Ba[9, 22], has higher larvicidal activity against mosquitoes than recombinant clones containing a subset of toxins[9, 23, 24]. In a previous study[25], Bti strain HD522 was evaluated against the same anopheline species, with An. funestus larvae having an LC50 that was statistically similar to the LC50 values obtained for An. arabiensis, An. gambiae, and An. merus larvae. The comparatively low larvicidal activity of PCC 7120#11 against An. funestus could be due to the absence of the other Bti Cry and Cyt proteins in PCC 7120#11, which produces only the Cry4Aa and Cry11Aa proteins. Since the larvicidal activity of Cry proteins is often correlated with the high affinity binding of the proteins to specific membrane-bound receptors in the larval midgut[26, 27], the susceptibility differences could be due to inherent differences between the species in the structure or density of midgut receptors for Cry4Aa or Cry11Aa. In this context, it is noteworthy that Cyt1A is known to synergize Cry11A toxicity by functioning as a membrane-bound receptor[28]. Further research is required to determine which combinations of the Cry and Cyt proteins would result in the highest larvicidal activity against An. funestus larvae.

PCC 7120#11 displayed good larvicidal activity against key malaria vectors, including An. gambiae and An. arabiensis. Although PCC 7120#11 displayed comparatively weak larvicidal activity against An. funestus at economically practical concentrations, PCC 7120#11 may have potential as a larvicidal agent in geographic regions where An. funestus is not the predominant vector. Alternatively, PCC 7120#11 could be applied as a larvicidal agent as part of an integrated vector control program that targets An. funestus adults and the larvae of other malaria vectors. However, before PCC 7120#11 can used in vector control programs, its effects on non-target organisms and its persistence in aquatic environments would have to be comprehensively evaluated.