Effects of mosquito species, food type and quantity
The rearing methods and experimental material that were analyzed in the observations reported here have been published previously . Briefly, 80 L1 An. gambiae ‘G3’ or Ae. aegypti ‘New Orleans’ strain larvae were reared in 150 mm Petri dishes containing approximately 100 ml of standard rearing water consisting of 0.3 g of pond salts (API, McLean, VA USA) per liter of type II purified water. After pupation, the containers that were to be reused were scrubbed with a wet sponge, rinsed in type II water and thoroughly air-dried. Alternatively, new containers were used. Four diets were used; two of these were custom formulations of a diet specifically designed for mosquitoes . This diet consists of a 2:2:1 ratio (by weight) of bovine liver powder, tuna meal and Vanderzant vitamin mix. One formulation was prepared at CDC in Atlanta, GA using ‘Now’ brand liver powder (Bloomingdale, IL USA), tuna meal (AA Baits, Rock Ferry, Birkenhead, UK) and Vanderzant vitamin mix (Bio-Serv, Flemington, NJ, USA). The other two diets were commercially available fish foods; TetraMin Plus Flakes (Tetra GmbH, Melle, Germany) and Doctors Foster and Smith Koi Staple diet (Rhinelander, WI USA). Koi pellets were ground in a Miracle Model MR–300 Electric Grain and Flour Mill (Danbury, CT USA) followed by sieving and saving the particles that passed a 600 standard sieve. The TetraMin was ground in a Black and Decker ‘SmartGrind’ coffee grinder (Beachwood, OH USA) to a consistency that passed through a 600 sieve. These diet types will be identified as CDC, Frontier, TetraMin, and Koi respectively. Diet was fed at two concentrations, 1.6 or 3.2% on alternate days. We also fed one diet, Koi, as a pellet, the form in which it is supplied, or as a powder. Due to the experimental design, the equivalent of 2.7% of pellet was provided in two pellets rather than 1.6% of the slurry. For details, consult the previously mentioned paper . There were three dishes for each species, food and concentration.
Estimating Biofilm Abundance
A quantitative estimate of the amount of biofilm was made in situ by staining with crystal violet (Product no. C0775–25G, Millipore Sigma, Billerica, MA USA), essentially by the method of O’Toole . A 0.1% solution of crystal violet was prepared and 10mL were added to each Petri dish. Dishes were stained with gentle agitation for 15 min and then rinsed 3 times with type II water and air-dried overnight. The bound crystal violet was eluted by adding 20mL of 30% acetic acid and incubated at room temperature for 15 min. The OD590, which is the maximum absorbance of crystal violet, was then measured using a general-purpose UV/VIS spectrophotometer (720 series, Beckman Coulter, Atlanta, GA USA) in 1 cm light path polystyrene cuvettes. We confirmed that maximal absorbance was 590nm in the 30% acetic acid solution rather than 550 nm as described by O’Toole. Background absorbance was measured on eluant from new dishes and subtracted from the absorbance values of all samples. The spectrophotometer was zeroed against a blank containing 30% acetic acid after approximately every 5 measurements to be certain there was no drift in observations.
To explore whether the form in which the food was provided might affect the amount of biofilm that formed, we established an orthogonal set of six dishes for each species, half of which were provided with two Koi pellets and half were provided a 1.6% Koi slurry. Dishes were treated with crystal violet as mentioned above, and the OD590 was determined. As a side note, we observed that Aedes aegypti are active feeders that disintegrated the pellets rapidly whereas the pellets remained largely intact for up to 2 days when placed in dishes containing An. gambiae which we thought might affect the amount of biofilm. It should be kept in mind that the weight of Koi food when fed as a pellet was greater than when fed as a slurry.
The effect of overnight bleaching on biofilm abundance was also determined. Thirty, previously-used dishes (without knowledge of the specific use) were allocated into two equal groups of 15; one group was treated overnight in a solution of 0.1% sodium hypochlorite prepared from an 8.5% sodium hypochlorite solution of concentrated Chlorox (Oakland, CA USA) bleach diluted in type II water. This formulation of bleach contained no fragrances or sodium hydroxide which are found in some household bleach solutions. After bleach treatment, the dishes were thoroughly rinsed three times in type II water and fully air-dried. The OD590 was determined as above.
Statistical analyses were performed using R version 3.5.1 “Feather Spray” . As OD590, the measure of biofilm abundance, was a continuous variable, an ANOVA was used. The OD590 data were strongly skewed and had a single outlier, these data were Ln-transformed for tractability in the analyses. The transformed values of the response variable (optical density) led to substantially improved model fit in all cases. As model fit was good, in no case was there evidence of particular influence by the outlier apparent in the un-transformed data.
Full models were fit with all main effects and interactions and the importance of terms was assessed by stepwise simplification. As the data are low replication, p<0.01 was used as a threshold for assessing the significance of interactions to avoid over-interpretation; p<0.05 was used for main effects. The occurrence (presence/absence) of total larval mortality which indicated microbial blooms as a function of whether the dish was previously used or new was analyzed by Fisher’s Exact Probability test. The effect of bleaching on the residual biofilm abundance was estimated by t-test.