Reproductive physiology of Anopheles gambiae and Anopheles atroparvus

Transcription

1 Vol. 30, no. 1 Journal of Vector Ecology 11 Reproductive physiology of Anopheles gambiae and Anopheles atroparvus Luís Fernandes and Hans Briegel Institute of Zoology, University of Zürich, CH-8057 Zürich, Switzerland Received 2 March 2004; Accepted 15 November 2004 ABSTRACT: When exposed to a human host, Anopheles gambiae started probing 4 h post-eclosion, but 95% successfully blood-fed by h with maximal blood volumes of 5-10 µl per female. When fed sugar, the 95% feeding was not observed until h post-eclosion; sugar meals appeared to interfere with blood meals. Similarly in An. atroparvus, maximum volumes were 10 µl when starved but only 6 µl when fed sugar. This species did not bite before 2 d, and 95% biting was by 4 d. Given single blood meals to water-kept An. gambiae, a threshold body size for oogenesis was detected. With wing lengths below 2.8 mm, eggs never matured, but when sugar-fed, females of all sizes matured eggs including the synthesis of maternal deposits. Although sugar feeding interfered with blood feeding, more lipid was transferred to the yolk. In water-kept An. atroparvus only 5% of the females produced eggs. When sugar-fed for 4 d, all females matured eggs, so in this species sugar feeding appeared to be essential for oogenesis. An. gambiae always took multiple blood meals, tested at any time after the first ones, leading to 120 mature eggs/female. Yolk composition was 3.9 mcal protein and 3.8 mcal lipid/oocyte when kept on water, but 2.8 mcal protein and 4.3 mcal lipid/oocyte with intermittent sugar meals, thus marking a surprising flexibility in synthesis of yolk protein and lipid that strongly depends on additional carbohydrates sources. Only 80% of water-fed An. atroparvus re-fed 2 d after a first blood meal with small females taking three blood meals but they still showed reduced fecundity. Only the large water-fed females matured eggs, with blood volumes higher than 9-12 µl. When fed sugar, the blood meal input was reduced, but oogenesis was possible, whereas water-fed females required three blood meals to reach the caloric level comparable to pre-feeding sugar-fed females. Water-fed An. gambiae could survive on daily blood meals alone, but survival was further extended by intermittent sugar meals. When offered a blood donor daily, there was a behavioral difference. Females maintained alone showed a more or less regular 3 d feeding and oviposition activity, while females kept in groups fed daily followed a daily oviposition pattern, suggesting gonotrophic discordance. Journal of Vector Ecology 30: Keyword Index: An. gambiae, An. atroparvus, ovarian development, multiple blood feeding, survival. INTRODUCTION Many species of the genus Anopheles are notorious as vectors of Plasmodium species that can cause malaria in vertebrates. Anopheles gambiae Giles sensu stricto, the African malaria mosquito, causes an extremely heavy burden for humans on that continent. Yet its reproductive physiology has not been adequately investigated. The same is true for An. atroparvus Van Thiel, formerly a malaria vector of the Northern hemisphere. Therefore, we have studied comparative aspects of their physiology during the reproductive cycle. In most mosquitoes, survival depends on plant sugars, which is the only food source for males. Sugar feeding by female An. gambiae in the field is uncertain. Although earlier reports described them feeding on sugar in the field (Laarman 1968, McCrae et al. 1969), more recent studies point to the opposite (Beier 1996). In the laboratory, (Briegel 1990b) found An. gambiae females to complement their nutritional deficiencies at eclosion by sugar feeding. However, sugar did not increase fecundity in An. gambiae, which obtained the same energetic Università degli Studi di Perugia, Dipartimento di Medicina Sperimentale e Scienze Biochimiche, Sezione Microbiologia, Via del Giochetto, Perugia, Italy. input from more frequent blood feeding (Straif and Beier 1996, Gary and Foster 2001). Feeding exclusively on blood tends to increase the biting frequency, as also observed in An. quadrimaculatus and Ae. aegypti (Foster and Eischen 1987, Canyon et al. 1999). In contrast to Aedes, blood meal utilization for oogenesis in Anopheles is characterized by low efficiency (Briegel 1990b). Anopheles females with their limited midgut volume have evolved the mechanism of blood meal concentration (Briegel and Rezzonico 1985) in combination with the ingestion of multiple blood meals, observed under laboratory conditions in An. albimanus, An. gambiae, and An. stephensi (Briegel and Hörler 1993). This is accompanied by hostseeking behavior during the gonotrophic cycle (Klowden and Briegel 1994). Multiple blood meals have also been reported from the field for An. gambiae s.l. and An. funestus (Gillies 1954, Beier 1996), even though there is an inhibition of hostseeking in large An. gambiae individuals for at least 40 h after blood feeding (Takken et al. 2001). The issue of multiple blood meals was addressed by (Gillies 1954, 1955), whose studies implicate that a first blood meal might promote the ovarian follicles from Christophers stage I to the resting stage II (Detinova 1962) before oogenesis could proceed with a second blood meal. He coined the term pre-gravid state for such field-caught An. gambiae and An.

2 12 Journal of Vector Ecology June 2005 funestus. The need for more than one blood meal to complete oogenesis was explained by (Briegel 1990a, b) by the fact that female Anopheles in general have much lower teneral reserves than culicine mosquitoes. The significance of body size in the reproductive physiology of mosquitoes has been recognized and emphasized by Briegel et al. (2001a), Briegel and Timmermann (2001), and Briegel et al. (2002). Body size is determined by environmental conditions during the larval period (Briegel 1990a, b, Lyimo and Koella 1992, Timmermann and Briegel 1993, 1999, Briegel et al. 2001a, Briegel and Timmermann 2001, Briegel et al. 2001b). Small mosquitoes have several disadvantages compared to large ones, including decreased survival, smaller teneral reserves, and reduced fecundities (Takken et al. 1998). A higher probability of survival was related to larger body size in An. arabiensis (Ameneshewa and Service 1996) and An. dirus (Kitthawee et al. 1992), and blood meal size and fecundity also improved with size (Hogg et al. 1996). Ramasamy et al. (2000) found a positive effect of body size on teneral reserves and the number of ovarian follicles in An. tessellatus. Body size also affected infection rates with Plasmodium falciparum (Lyimo and Koella 1992). Lyimo and Takken (1993) found that wild An. gambiae females with wing lengths above 3 mm were able to mature eggs from the first blood meal, while smaller females required two or three blood meals to mature the first batch of eggs. These results were later confirmed under laboratory conditions (Takken et al. 1998). Our study of nutritional status before a blood meal, the mobilization of teneral reserves, and the reserve synthesis from sugar meals may explain the extent of blood consumption and the efficiency of its utilization for either reserve synthesis or oogenesis. The pre-gravid state described by Gillies (1954, 1955) will be related to the body size of a female and its dietary history, both determining the course of oogenesis and survival. The issue of multiple blood feedings is addressed in relation to diverse metabolic parameters and compared between the two vector species with their different physiotypes. MATERIALS AND METHODS Mosquito colonies Our Anopheles (Cellia) gambiae s.s. colony originated from Lagos, Nigeria, and has been routinely colonized for 20 years at 27 ±1 C and 85±5% R.H. under long day conditions (14L:10D), with 40 min dimming periods simulating sunrise and sunset. Routinely, 400 newly-hatched larvae were counted into 16 x 24 x 5 cm pans and fed standardized volumes of pulverized TetraMin daily. In order to obtain small mosquitoes, at least 800 larvae were counted per pan, since intraspecific competition at the larval stage causes a decrease in adult body size (Timmermann and Briegel 1993). Anopheles (Anopheles) atroparvus has been colonized continuously for 9 years in our laboratory at room temperature (22±1 C) under long day conditions (14L:10D). Routinely, 300 newly-hatched larvae were counted into polystyrol pans and fed a different amount of TetraMinBaby daily with the same set of feeding spoons. At least 800 larvae were counted to obtain small adults. For colony maintenance all blood meals were given from restrained guinea pigs. Experimental procedures Mosquitoes used in experiments originated from both rearing conditions to create the full range of body sizes, and females were not mated. Body size of all mosquitoes was determined by wing length measurements from the alula to the tip, excluding the fringes. Measurements were converted to cubic values (WL 3 ), which are more suitable for scaling physiological data than linear dimensions (Schmidt-Nielsen 1984, Briegel 1990a). Teneral females were sampled not longer than 1 h after eclosion and fixed immediately for caloric measurements, or they were kept with access to water or sucrose concentrations ranging from 0.1% to 20% in round cages (11.4 cm diameter x 13.0 cm height) for survival and blood feeding experiments. The different sucrose concentrations were used to determine the ideal concentration for mosquito survival. An. gambiae was always studied at 27 C and An. atroparvus at 22 C, but survival experiments of both species were performed at both temperatures, 27 C and 22 C. The extent of reserve mobilization under nutritive stress was measured in females starved to death with access to water only. These experiments revealed the minimal irreducible amounts (MIA), necessary for survival (Van Händel 1984, Briegel 1990a, b). Females kept with access to 10% sucrose were sampled at regular intervals to estimate lipid and glycogen synthesis. Development of avidity to blood ingestion was studied in females maintained on water or 10% sucrose from eclosion and exposed to a host in 0.2 dm 3 cages at 4 h intervals in An. gambiae and daily intervals in An. atroparvus. At 22 C mosquitoes develop more slowly than at 27 C, so timing of experiments was adapted to the different development times of the two species. An. atroparvus females seldom ingested blood from a human arm, so a guinea pig was offered instead. Either the human arm or a guinea pig were exposed for 15 min, or until all females ceased probing. Shortly afterwards, each female was checked for the presence of blood with transmitted light and kept individually in vials (13 x 100 mm) with access to water. Feces were collected for later determination of hematin. After blood meals were digested (44-50 h for An. gambiae and 80 h for An. atroparvus), the females were dissected for measurements of follicle and yolk lengths and for oocyte counts. Wings were kept for the measurement of body size. Mosquitoes with undigested blood in the midgut were discarded. The development of ovaries before a blood meal was also studied at various times after eclosion in females kept with access to water or 10% sugar. The ovaries were examined at 300x to determine the Christophers stages as described by (Detinova 1962), and the follicle lengths were measured at 100x. Their caloric content of protein, lipid, and glycogen was also measured. To reach measurable amounts, the resting stage ovaries from 30 to 70 An. gambiae females, and 25 An. atroparvus females had to be pooled for each sample, depending on the developmental stages. To study oogenesis, groups of water-fed An. gambiae

3 Vol. 30, no. 1 Journal of Vector Ecology 13 were blood fed on a human arm 24 h after eclosion, or after 5 d when fed on sugar, and their ovaries were dissected in 3-12 h intervals after blood ingestion. Due to their slower development and decreased blood hunger An. atroparvus were blood fed on a guinea pig 8 d after eclosion, which included 5 d of sugar feeding plus 3 d of starvation. Some had their ovarioles teased apart and their follicle and yolk lengths measured, while others had their entire ovaries used for determinations of protein, lipid, and glycogen contents. In addition, females were analyzed for their total caloric contents before and after oviposition to assess the processing of a blood meal for oogenesis, for energetic demands, or for maternal reserve synthesis. The significance of multiple blood meals during one gonotrophic cycle was studied thoroughly in both species of all size classes. An. gambiae females were maintained with access to only water and blood fed 24 h after eclosion. Some females were then offered a second blood meal 24 or 48 h later (48 or 72 h after eclosion). An. atroparvus were also kept with access to water, and offered a guinea pig daily. After digestion was completed, the females and their ovaries were analyzed and their oocytes counted. Survival with daily blood meals from humans was studied in An. gambiae with access to water or 10% sucrose from eclosion in square 8 dm 3 cages. Between 50 and 100 sugarfed males were added daily to the cages containing females (depending on experiment) for the first week to ensure insemination. The cages had an oviposition cup and were screened daily for dead mosquitoes, which were removed and had their wing length measured. The number of mosquitoes taking a blood meal and their gonotrophic status was recorded and ovipositions were counted daily. A group of 50 females eclosed the same day, was kept with 10% sucrose, and used as a control. Analytical procedures For biochemical analyses, mosquitoes were fixed individually in reacting tubes (10 x 75 mm) with 100 µl of ethanol and heated to 90 C for 3 min. The Kjeldahl digestion and subsequent Nesslerization of ammonium sulfate was used to measure the protein content, as described by Briegel (1986). Lipid and carbohydrates were measured in the same specimen, fixed in 13 x 100 mm tubes, following procedures by Van Handel and Day (1988). Lipid contents were measured on the basis of the vanillin reaction and carbohydrate by the anthrone method. All data were converted to calories per insect to allow direct energetic comparisons, based on the proportion of 110 mg lipid and 250 mg carbohydrate or protein for 1 cal (=4.18J); the corresponding conversion factors are cal per mg lipid and cal per mg carbohydrates or protein. To normalize caloric measurements for body size, the calories were divided by WL 3, leading to a size-specific caloric content (SSCC-values) (Timmermann and Briegel 1993). Fecal material of individual females was eluted in 1% lithium carbonate for measurement of hematin (Briegel 1980). The amount of blood ingested was quantified retrospectively from the hematin readings for each female based on the stoichiometric relationship between fecal hematin and hemoglobin ingested. The molar proportions between vertebrate hemoglobin and fecal hematin allowed us to calculate the blood consumption of a female, first in terms of hemoglobin, then in total protein, and finally as µl of host blood ingested (Briegel 1986). This method accurately reflects the amount of protein ingested, but the total volume of blood should be taken as a reference and not as an actual value, because Anopheles mosquitoes concentrate their blood meal during feeding (Briegel and Rezzonico 1985), and fecal hematin reflects the hemoglobin digested by the midgut. Blood expelled during blood feeding (prediuresis) was discarded. The hemoglobin titer of the blood donor was monitored by Drabkin s solution (Briegel et al. 1979). RESULTS Teneral conditions Body size and total protein, lipid, and glycogen contents were measured within 1 h of eclosion for all possible body sizes. In An. gambiae, wing lengths were in the range of mm; females were slightly larger than males. Their caloric values were about two-thirds of protein and close to another third of lipids, while carbohydrates always remained below one-tenth of the total. Plotting the individual caloric values of females against their body size (WL 3 ), the following linear regressions were obtained: protein y=0.037x (N=118, r 2 =0.925, t=37.91, p<0.0001); lipid y=0.030x (N=174, r 2 =0.803, t=26.46, p<0.0001); glycogen y=0.004x (N=154, r 2 =0.501, t=12.34, p<0.0001). When normalized for body size (WL 3 ) and plotted as SSCC-values against body size, equally linear regressions were revealed for all three components: protein y=0.0003x (N=118, r 2 =0.236, t=5.99, p<0.0001); lipid y=0.0008x (N=174, r 2 =0.595, t=15.90, p<0.0001); glycogen y= x (N=154, r 2 =0.067,t=3.31 p<0.01). Protein and glycogen showed isometric relationships with body size, while total lipids revealed a weak positive allometry. In An. atroparvus, body sizes of females were larger than males, within the ranges of mm and mm wing length respectively. The caloric contents of females revealed significant linear regressions when plotted against body size (WL 3 ): protein y=0.024x (N=135, r 2 =0.865, t=29.23, p<0.0001); lipid y=0.024x (N=149, r 2 =0.80, t=24.26, p<0.0001); glycogen y=0.004x (N=149, r 2 =0.572, t=14.02, p<0.0001). Regression formulas for males were somewhat different (data not shown). When the caloric values of females were transformed to SSCC-values, the following regressions were obtained: protein y= x (N=135, r 2 =0.139, t=4.641, p<0.0001); lipid y=0.0002x (N=149, r 2 =0.643, t=16.26, p<0.0001); glycogen y= x (N=149, r 2 =0.201, t=6.08, p<0.0001). There were no major differences between protein and glycogen of small and large females, but lipids were in positive allometry. Survival and reserve mobilization Under starvation conditions with access to only water from eclosion, all An. gambiae died within 4 d and An.

4 14 Journal of Vector Ecology June 2005 atroparvus died within 8 d. Small females died earlier as previously reported by Takken et al. (1998). When provided with sugar solutions, however, survival was substantially extended. Cohorts of 50 large (WL 3.0 mm) and 50 small An. gambiae females (WL 2.6 mm) with access to 10% sucrose showed maximal survival times of 20 and 15 d (50% survival times 9 and 11 d), respectively. In An. atroparvus, access to sugar solutions extended survival to 33 and 40 d (50% survival times 16 and 21 d) for 50 large and 50 small females, respectively. Survival curves largely overlapped, so body size had no major consequences on survival times when sucrose was available as an energy source. Various sugar concentrations from 0.1% up to 20% in identical cages were tested for female survival of An. gambiae. Sucrose concentrations below 0.5% revealed no differences in survival to water, whereas higher concentrations led to a gradual increase of the 50% survival times. Maximum survival times for An. gambiae were obtained with concentrations of 5-10%; 20% solutions increased mortalities. In An. atroparvus, the effect of different sugar concentrations (0.1%-20%) was similar. Concentrations higher than 0.4% gradually increased the 50% survival times, and maximum survival was also reached with 10% sucrose. The effect of temperature was compared between An. atroparvus and An. gambiae with 10% sucrose (Figure 1). In both species the lower temperature extended their survival. The caloric protein, lipid, and glycogen contents of females that died during these survival experiments were determined. The protein contents of dead females were similar to teneral values; in An. gambiae mean contents were Figure 1. Comparison of female survival between An. gambiae (triangles), and An.atroparvus (squares), and between two temperatures (open and filled), with access to 10% sucrose. (27 C, N=609; 22 C, N=98); An. atroparvus (27 C, N=149; 22 C, N=236). Table 1. Oogenesis in An. atroparvus with access to water or sugar from eclosion, given a blood meal at successive days (M±SE). Time of blood meal (d) Water Sugar N Gravid (%) eggs N Gravid (%) eggs (1) (7) (4) 134± (81) 194± (3) 139± (100) 184± (100) 158± (100) 166± (100) 154± ±0.08 cal protein (N=32, range ), 0.21±0.15 cal lipids (N= 43, range ), and 0.05±0.03 cal glycogen (N=43, range ). In An. atroparvus, the corresponding values were 1.90±0.13 cal protein (N=13, range of ), also similar to their teneral values (t=1.16, p>0.2). The caloric lipid content was 1.19±0.97 (N=56, range cal), and glycogen was 0.29±0.31 (N=56, range cal). Complete starvation, i.e. access to water only, uncovered the minimal irreducible amounts (MIA) required for survival. The differences to the teneral values indicate maximal mobilization of reserves. Depending on body size, in An. gambiae from 5-36% of the teneral protein was catabolized by smallest and largest females, respectively. At the same time 25-80% of the lipid and roughly 60-80% of the glycogen was mobilized. Thus, lipid contributed the largest absolute segment of teneral reserve mobilization, with a 30-fold difference between the extreme sizes; obviously, the extent of mobilization of teneral lipid and protein reserves for survival was strongly affected by body size. The time course of lipid and glycogen mobilization was followed in 4 h intervals, requiring h to reach the MIA levels. This period coincided with the aforementioned survival time. In An. atroparvus large females utilized 5-42% of the teneral protein content, 37-85% of the teneral lipid, and 40-50% of glycogen, always depending on size. Large females had a clear advantage; they could utilize more teneral reserves, particularly in absolute amounts, which explains their longer survival under starvation over An. gambiae. The lipid and glycogen contents of females which died either early or late during the survival experiments often varied considerably. Therefore, death may not always be caused by exhausted reserves because dead insects sometimes contained large amounts of sugar. Reserve synthesis from sugar meals Beginning at eclosion, female An. gambiae with access to 10% sucrose were sampled for the first 2 d at 4 h intervals and from day 3 onwards for another 3 wk at 2 d intervals. They were screened for body size and analyzed for accumulation of lipid and glycogen (Figure 2). Glycogen was steadily synthesized from eclosion for 36 h to its maximal levels of roughly 0.5 cal in large and 0.25 cal in small females,

5 Vol. 30, no. 1 Journal of Vector Ecology 15 Figure 2. Lipid and glycogen content of large female An. gambiae kept for three weeks with 10% sucrose (M±SE, N=12-32 per point). The teneral values corresponding to their mean body size (WL ±3.25, N=465) are indicated by a filled symbol. Figure 3. Lipid and glycogen content of large female An. atroparvus kept for four weeks at 22 C with 10% sucrose (M±SE, N=7-23 per point). Teneral values are indicated by filled symbols for an average size of WL ±8.33 (N=295). representing a 6-7-fold increase over teneral levels for both size classes. During the next 3 wk the glycogen contents gradually declined. Lipogenesis followed a different pattern (Figure 2). In large females it took about 2 d until teneral levels had roughly doubled the lipid contents. Thereafter, this remained at this level for another 2 wk, although with great variability. In small females the lipid content oscillated irregularly around teneral levels during the first day and then slowly increased, but rarely doubled; after 5 d it decreased. The same experiments with An. atroparvus are summarized in Figure 3 for large females. Glycogen synthesis varied between 0.5 and 1.5 cal with an average maximum of 1.2 cal, i.e. a 4-5-fold increase over teneral values. Synthesis of lipids was clearly delayed; it took about 2 wk to reach the maximum levels. By day 3 a substantial lipogenesis had started. Large females attained maxima more than 3-fold the teneral values and in small females the lipid maximum was reached 2 d earlier with a 4-fold increase. Compared to An. gambiae, An. atroparvus is characterized by remarkably high caloric lipid accumulations, even when normalized for body size (data not shown). Development of blood hunger The first successful ingestion of blood was determined by offering a human arm every 4 h to inexperienced females that had permanent access to water or 10% sucrose after eclosion. With An. gambiae, a few females attempted to bite at 4 h after eclosion, but none successfully fed until 8 h posteclosion. For both treatments, the 95% threshold was observed at h, and by 24 h 100% of the water-kept females had fed. In females with previous access to sugar the 95% feeding threshold was delayed to h. In An. atroparvus, females maintained on water did not feed before day 2 and only by day 4 was the 95% feeding threshold reached. When kept on sugar, An. atroparvus females always showed a decreased tendency (below 30% by day 5) to blood-feed. The amount of blood ingested was determined for An. gambiae on the basis of hematin measurements (Figure 4). Females maintained on water gradually increased their blood meals until maximum volumes were reached at h. The females kept with sugar initially behaved similarly, but their blood meals were slightly smaller and at 24 h they showed a decline in blood meal consumption. This is explained by voluminous crops due to previous sugar consumption. Indeed, in such females only half of the abdomen was filled with bright red blood while the other half appeared translucent, characteristic of filled crops. From h onwards the blood consumption increased and became similar in both treatments. The maximal volumes of 5-10 µl per female were possible

6 16 Journal of Vector Ecology June 2005 Table 2. Total caloric increase above teneral values of An. gambiae after processing one blood meal given after 1 d of access to water. Gravid * WL Caloric increase over teneral values Protein (%) Lipid (%) (62) 0.38 (117) (28) 0.45 (122) (32) 0.49 (111) (33) 0.50 (88) Table 3. Total caloric increase above teneral values of An. gambiae after one blood meal given after 5 days of continuous access to 10% sucrose. Gravid * WL Caloric increase over teneral values Protein (%) Lipid (%) (25) 0.57 (699) (27) 0.69 (548) (19) 0.72 (482) (23) 0.86 (413) (25) 0.21 (254) (26) 0.16 (125) (15) 0.29 (191) (0) 0.85 (260) (0) 0.92 (248) (0) 1.26 (283) (1) 1.09 (191) (44) 0.29 (141) (29) 0.16 (50) (5) 0.25 (67) (10) 0.14 (32) (15) 0.30 (53) *Ovarian values are included. Nonoogenic Nonoogenic (1) 0.51 (623) (23) 0.49 (392) (10) 0.49 (326) (34) 0.57 (275) *Ovarian values are included. Table 4. Multiple blood meals of An. gambiae. Increase of blood meal size, percentage of gravid females, and fecundity of females kept with constant access to water. They were fed blood once at 24 h after eclosion, twice at 24 and 48 h, or at 24 and 72 h. Twenty females even took three blood meals at 24 h intervals. The average body size was 2.75 mm wing length (range mm; M±SE, N in parentheses). Time of blood meal (h) Size of blood meal µg hemoglobin µl blood Gravid females (%) Eggs ±178 (35) 2.2± ±18 (14) 24/48 652±184 (118) 4.4± ±24 (36) 24/72* 446±103 (79) 3.0± ±45 (31) 24/48/72 740±143 (20) 4.8± ±21 (9) *By 72 h, females had already digested blood taken at 24 h, so the values measured at 72 h were reduced. Therefore, total blood consumption was estimated by adding the values measured at 24 h and at 72 h, i.e. 772 µg or 5.2 µl.

7 Vol. 30, no. 1 Journal of Vector Ecology 17 Figure 4. Development of blood consumption in An. gambiae kept with water (open circles) or sugar (filled squares) after eclosion. Amount of blood ingested based on fecal hematin determinations. The hemoglobin values were converted to approximate volumes of blood ingested and concentrated during feeding (Mean±SE, N=13-55 per data point). because of the concentration mechanism during feeding (Briegel and Rezzonico 1985). In An. atroparvus, the maximal blood uptake was about 10 µl, reached within 4 d when kept on water. Providing females first with sugar had an equally negative effect on blood consumption; fewer females ingested blood and they always took smaller meals of roughly 6 µl. This reduced blood feeding persisted until day 15 but could be improved when sugar sources were removed before. Up to 80% of the females kept with sugar for 5 d, and then starved for 3 d took blood, although the blood volume was not increased (data not shown). Ovarian development prior to a blood meal Christophers stages and follicle lengths in ovaries of An. gambiae were compared between water-fed and sugar-fed females. In An. gambiae, Christophers stage I corresponds to 40 µm of follicle lengths, stage I-II to µm, and stage II to µm. There was no significant difference between Table 5. Multiple blood-feeding in An. atroparvus kept with water. Percentage of refeeding females, the amount of blood ingested (µl), the percentage of oogenic females, and fecundity (M±SE, N in parentheses). Time of blood meal (d) Tested Blood-fed* (%) Blood consumption (µl)** Gravid (%) Eggs ±13 (91) 12±4 28 (56) 179± ±13 (57) 15±4 37 (74) 183± ±12 (32) 13±5 13 (45) 163±56 *Percentage based on the number of females that had ingested a previous blood meal (=100%). **These values correspond to mean blood volumes of 12-15µl with a range of 3-26µl per female. water-fed and sugar-fed females, as both had reached stage II within 2 d after eclosion. With sugar however, follicles were slightly larger. The maximum follicle length of 60 µm, attained 36 h after eclosion, corresponded to the stable, pre-gravid state described by Gillies (1954), or the resting stage, or the last stage of the previtellogenic phase, according to Clements (1992). In both water-fed and sugar-fed An. atroparvus, the resting stage was reached only by day 5 with mean follicle lengths of 90 µm. In Christophers stage I the follicles measured µm, in stage I-II µm, and in stage II µm. Large and small females of both species were able to reach stage II when given sugar, with no major differences in oocyte length or caloric content. When starved, small females were not able to reach stage II because they died before 36 h in An. gambiae or 5 d in An. atroparvus. In starved females of An. gambiae, the ovarian caloric content at 12 h was 8.0±3.9 mcal/female protein (N=3), 1.4±0.2 mcal lipid (N=4), and 0.4±0.2 mcal glycogen (N=5). While lipid and glycogen levels remained fairly stable between 12 h and 36 h, protein levels fluctuated. In sugar-kept females the ovarian contents at 108 h after eclosion was 10.0±2.5 mcal/ female protein (N=3), 2.6±0.3 mcal lipid (N=3), and 1.1±0.1 mcal glycogen (N=3). These values were maintained for 4-5 d at the teneral level. Ovarian lipid always remained below 3 mcal/female and glycogen below 1 mcal/female. In An. atroparvus, the teneral ovarian content was 10.3±0.9 mcal/female protein (N=3), 2.4±0.3 mcal lipid (N=4), and 0.5±0.1 mcal glycogen (N=4). When kept with water, the ovarian protein doubled within 1 d and remained constant until death occurred 3 d later. With access to sugar for 2 d however, the ovarian contents were 34.6±1.7 mcal/ female for protein (N=2), 3.1±1.4 mcal for lipid (N=3), and 0.7±0.2 mcal for glycogen (N=3). The ovarian protein tripled within 2 d, whereupon it remained stable, and ovarian lipid doubled. The ovarian lipid never exceeded 6 mcal, and glycogen always remained within 2 mcal. Ovarian protein appeared to be more flexible in both species. Oogenesis with single blood meals Water or sugar-fed An. gambiae received single blood meals at 24 h after eclosion. Their feces were collected during the period of blood digestion and oogenesis for determining their blood meal sizes, and the females were either dissected for egg counts or fixed for quantification of total protein, lipid, and glycogen. The amount of blood ingested was clearly sizedependent (Figure 5). All non-oogenic females were significantly smaller than oogenic ones (t=17.6, p<0.0001) and had ingested significantly smaller blood meals (t=22.2, p<0.0001). Figure 6 shows the gravidity data for each body size and a threshold body size for oogenesis. None of the water-fed females equal to or smaller than 2.7 mm became gravid, and with body sizes of 2.8 mm only 11% were able to mature eggs, while larger-sized females were increasingly successful in developing eggs. Among sugar-fed females, however, most entered and completed oogenesis, although with a size-dependent frequency. No size threshold was encountered. Among 295 females of An. atroparvus kept on water,

8 18 Journal of Vector Ecology June 2005 Table 6. Overview of female and ovarian caloric contents in both Anopheles species given only one or two blood meals. An. atroparvus without sugar rarely started oogenesis. The values of non-oogenic females are not included, but were similar in protein and glycogen to teneral females, while lipid was increased over one-third (mean cal/female). Protein Lipid Glycogen Σ An. gambiae Teneral d 1* Non-oogenic Before oviposition After oviposition Ovaries d 1/2* Before oviposition After oviposition Ovaries An. atroparvus Teneral d 3** Before oviposition After oviposition Ovaries d 3/5* Before oviposition Ovaries *No sugar **Sugar available Table 7. Overview of ovarian and oocyte development in An. gambiae and An. atroparvus. Follicle length of ovaries at resting stage II, together with oocyte caloric content after one blood meal, when kept with water or sugar previously and after multiple blood meals. An. gambiae An. atroparvus Resting stage ovaries* Follicle length (µm) (mcal/female) Protein** Lipid** Single blood meal Eggs/female Water-kept Protein (mcal/oocyte) Lipid Total Sugar-kept (mcal/oocyte) Protein Lipid Total Multiple blood meals Eggs/female Water-kept Protein (mcal/oocyte) Lipid Total *Oocyte numbers not determined. **Maximum ovarian values of sugar-kept females.

9 Vol. 30, no. 1 Journal of Vector Ecology 19 Figure 5. Blood meal sizes of non-oogenic (A) and oogenic (B) An. gambiae fed single blood meals 24 h after eclosion with acess to water. Regression formula for hemoglobin on body size are: A y=31.9x (N=241, r 2 =0.654, t=21.25, p<0.0001); B y=19.6x (N=216, r 2 =0.133, t=5.73, p<0.0001). only 7 entered oogenesis after one blood meal with up to 150 eggs (Table 1). Females kept on water died by day 5-6. With access to sucrose the percentage of oogenic females rapidly increased between day 2 and 3, and by day 4 all females became gravid, maturing eggs. The caloric contents of gravid An. gambiae after single blood meals again correlated significantly with body size. The following regressions were obtained for previously starved females: protein y=0.037x (N=42, r 2 =0.422, t=5.41, p<0.0001); lipid y=0.046x (N=50, r 2 =0.215, t=3.62, p<0.001); glycogen y=0.008x (N=50, r 2 =0.284, t=4.36, p<0.0001). The regressions for protein and lipids had slopes similar to teneral conditions but were increased by roughly 25% for protein and 33% for lipids. Blood meals that did not allow oogenesis led to the formation of maternal deposits that approximated half the values of yolk. This is demonstrated in Table 2 for each size class of gravid and nonoogenic An. gambiae that had access to water before. In gravid females, lipid had by far the highest increase with 0.5 cal; with protein it was cal per female and glycogen increased only by cal. Non-oogenic females had smaller values and smaller increases. These data express the efficiency of blood meal utilization for the synthesis of yolk in oogenic females and of reserves in non-oogenic females. Gravid females had utilized 5-11% of the blood meal as protein, 10-11% as lipid, while glycogen remained below 2%. Non-oogenic females, however, utilized a total of 9-24% of the caloric input in an inverse relation with body size: protein accounted for 1-8%, lipid 4-15%, and glycogen 2% or less. Together, these results suggest that nonoogenic An. gambiae were able to use similar proportions of the blood meal as did oogenic females. Almost as much of a first and single blood meal was converted to the synthesis of maternal reserves as was transferred by oogenic females to yolk. Similar data are compiled in Table 3 for An. gambiae that had access to sugar before the blood meals. In gravid females, the regressions of caloric content on body size were the following: protein y=0.028x (N=55, r 2 =0.701, t=11.15, p<0.0001); lipid y=0.072x (N=48, r 2 =0.826, t=14.79, p<0.0001); glycogen y=0.013x (N=48, r 2 =0.588, t=8.10, p<0.0001). In gravid females, the relative gain in lipid was 2-7-fold of teneral values, but only up to 27% in protein, both in an inverse relationship to body size.

10 20 Journal of Vector Ecology June 2005 Figure 6. Oogenesis in An. gambiae after a single blood meal depending on body size. Females had either access to water for 1 d (A), or to sugar for 5 d (B) before the blood meal. The percentage of gravid females is given for each size class. With water 457 females were tested and 216 became gravid (47%), with sugar this was 150 females (81%) out of 185 tested. However, in absolute terms, large females accumulated considerably more calories than smaller ones. Non-oogenic females also gained up to 34% protein and 2-6-fold lipid (Table 3). Although sugar-feeding had a deleterious effect on blood meal volume as shown before (Figure 4), it had a positive effect on egg production (Table 3). Sugar-kept females had much lower protein contents that were apparently compensated by increased lipid contents derived from previous sugar meals. In An. atroparvus, the total protein and lipid contents after digestion of a single blood meal were also in significantly linear regression with body size: protein y=0.029x (N=70, r 2 =0.717, t=13.12, p<0.001); lipid y=0.031x (N=74, r 2 =0.344, t=6.14, p<0.001) for sugar-fed oogenic females. For water-fed non-oogenic females regressions were: protein y=0.018x (N=89, r 2 =0.603, t=11.49, p<0.001); lipid y=0.011x (N=87, r 2 =0.147, t=3.82, p<0.001). Oogenic females had roughly doubled their teneral protein and doubled or tripled their teneral lipid, depending on body size. In non-oogenic females the teneral protein and lipids remained at similar levels, thus no maternal deposits were discerned. Multiple blood meals, fecundity, and reserve accumulation Sugar meals reduced the tendency to feed on blood, so only water-fed females were tested. An. gambiae showed a continued avidity for blood; after a first blood meal all females readily took a second meal at any time tested. Up to three blood meals were ingested on consecutive days after eclosion (Table 4). With two blood meals in 24-h intervals almost all females developed eggs, compared to only 40% with one blood meal. With a second blood meal at 48 h the mean number of mature eggs increased from 56 to 93 per female and when fed at 24 and 72 h, mean fecundity rose to 120 eggs (Table 4). Because digestion of a single blood meal took 40 to 44 h, the females with the second meal at 72 h had largely digested their first meal, but oviposition was denied. In females fed three-times, fecundity was somewhat lower, because of a volumetric interference between the second and third blood meals that limited total ingestion. We analyzed the caloric content of mature oocytes in females taking multiple blood meals. Among females that had no sugar before, the caloric content per mature oocyte was fairly constant: 3.9±0.4 mcal of protein, 3.8±0.4 mcal of lipid, and 0.4±0.2 mcal of glycogen (N=17-19). Protein and lipid as the main constituents of yolk occurred in roughly equicaloric amounts, while glycogen always remained below 10%. Females that blood fed once with previous access to sugar ingested smaller blood meals and matured oocytes with lipid as a dominant component: 2.8±0.7 mcal protein, 4.3±0.6 mcal lipid, and 0.5±0.3 mcal glycogen (N=22-23). The total was 7.6 mcal, significantly less than in water-kept females with 8.1 mcal per oocyte (t=8.11, p<0.005). In a similar approach, An. atroparvus with access to water were tested for multiple feedings. This species was not as prone to take multiple blood meals as the previous one. The results are presented in Figure 7 in relation to body size. Females were offered their first blood meal on day 3 after eclosion, a second blood meal on day 4 or 5, or three blood meals at days 3, 4, and 5. Even though females were starved, there were always some that did not bite. The large females ingested sufficient amounts of blood to mature eggs (Figure 7), but there appeared to be a critical blood meal size between cal protein (9-12 µl blood), above which oogenesis was possible. The highest average refeeding of 80% was observed with an interval of 2 d between two blood meals (Table 5). Correspondingly, blood meal size and gravidity were maximal. Fewer females took blood on three consecutive days, and the ones that did were almost all of a smaller size. Smaller females ingested relatively smaller volumes, which in turn affected oogenesis. Therefore, even though females took three blood meals, fecundity appeared reduced (Table 5). Sugar feeding was advantageous to An. atroparvus females because it allowed oogenesis with a single blood meal, while in water-kept females three blood meals were required to attain similar caloric amounts to females that had access to sugar before the first blood meal. However, females with access to sugar took considerably smaller blood meals. The caloric content per mature oocyte also differed between both treatments. Sugar-maintained An. atroparvus had oocytes that contained 5.8±1.0 mcal of protein, 6.8±1.0 mcal of lipid, and 0.3±0.2 mcal of glycogen (N = 39), i.e. a total of 12.9 mcal, significantly higher than water-kept females with 5.1±0.4 mcal of protein, 4.8±0.5 mcal of lipid, and 0.3±0.1 mcal of glycogen (N = 11-14), totaling 10.2 mcal (t=5.84, p<0.005). An overview of female and ovarian caloric contents is given for both species in Table 6. Non-oogenic An. gambiae increased their lipid deposits by an average of 0.2 cal,

11 Vol. 30, no. 1 Journal of Vector Ecology 21 Figure 7. Blood consumption of An. atroparvus kept with water from eclosion, and blood-fed twice, i.e. day 3 and 4, or day 3 and 5, or three times, i.e. at day 3, 4 and 5 after eclosion (N=129). The grey area illustrates the blood meal size assumed to be critical for oogenesis: Oogenic females (filled circles; N=78), were clearly above this amount and non-oogenic ones (open squares; N=51) below. The broken line indicates the regression (y = 3.821x ; N=183, r 2 =0.096, t=4.38, p<0.001) for blood consumption with single blood meals of females fed sugar before and all were oogenic. Figure 8. Survival of An. gambiae with daily blood meals. Either they had permanent access to water (filled circles; N=47) or sucrose (filled triangles; N=36). For comparison, survival of females with access to sucrose only (open squares; N=50).

12 22 Journal of Vector Ecology June 2005 Figure 9. Frequencies and pattern of blood feeding and oviposition in 16 water-kept, inseminated An. gambiae kept individually in 8 dm 3 cages after blood meals. The oviposition followed a more or less regular 3-d pattern after the peaks of blood meals. Figure 10. Feeding and oviposition frequency of An. gambiae kept in groups of 18 in 8 dm 3 cages. A: blood meals were taken almost daily. B: The oviposition pattern was continuous. No clear gonothropic concordance was recognized.

13 Vol. 30, no. 1 Journal of Vector Ecology 23 demonstrating once more that, for them, the blood meal is an important energy source in addition to providing the raw material for oogenesis. In Table 7 an outline of ovarian and oocyte development is presented for both species. The composition of yolk was flexible and varied with diet, and providing sugar before a blood meal caused a proportional increase in lipid content. While An. gambiae females revealed the highest total contents per oocyte when given blood alone, in An. atroparvus highest values were observed when given access to sugar before the blood meals. Survival with only access to blood We tested the survival of An. gambiae with blood alone as the dietary substrate. Between the blood meals, females had access to water. The number of females taking blood, their gonotrophic status, and the number of eggs oviposited were recorded daily; females could not be screened for insemination. Maximal survival time of blood-fed females was 54 d when they had access to sugar and 42 d with water. The corresponding 50% survival times were 27 d and 16 d respectively, clearly extended when compared to sugar alone (20 d, Figure 8). The survival curves for large and small females largely overlapped. Sugar-kept females with daily blood meals matured and oviposited an average of 606 eggs per female, more than double the average number of 229 eggs of blood-fed females maintained on water in-between. Evidently, blood meals as the only food source allow a considerably extended survival of An. gambiae, and an appreciable fecundity. However, additional sugar sources further improved total fecundity. In these experiments, the intestinal and gonotrophic status were also evaluated. Females feeding on the host, while dark blood still was present in their midguts, were judged to refeed. Females containing yellowish yolk material that were re-feeding were considered to be gonoactive, while a third group of females approaching the host appeared empty. It was impossible to determine whether they were at the end or at the beginning of another gonotrophic cycle, because some might have had ingested and digested a small blood meal without producing eggs. Because this evaluation was not error-free as a result of small blood meals that might have escaped visual detection, an attempt was made to follow the oviposition pattern. After mating, females were either kept individually or in groups in large cubic cages (8 dm 3 ) with continuous access to water, and offered blood every day for at least 10 min or until all had stopped probing. When kept singly, the majority of the blood meals were taken in a 3-day pattern, and oviposition followed 3 d later in almost the same pattern (Figure 9). Thus, these females showed regular gonotrophic cycles of 3 d, indicating their concordance. Some females apparently required two blood meals to complete their gonotrophic cycles. Females with empty spermathecae were discarded. However, when kept and blood fed in groups, females behaved differently. At 7 out of 12 d they regularly took blood meals, and their oviposition pattern was followed (Figure 10). Apparently females were capable of taking blood meals almost daily and oviposited continuously, thus lacking gonotrophic concordance. Probably, females kept in groups might experience disturbances in their feeding activity, apparently disrupting their gonotrophic concordance. Furthermore, oviposition activity of An. gambiae was affected by cage size: the smaller the container, the more reluctantly they oviposited. In small cups (27 cm 3 ), mature oocytes were retained for over 5 d, despite the presence of sperm in the spermathecae. DISCUSSION Body size is well-established as a crucial parameter of fecundity and fitness in many mosquitoes and particularly in several Anopheles species (Briegel 1990b, Kitthawee et al. 1992, Lyimo and Takken 1993, Ameneshewa and Service 1996, Takken et al. 1998, Charlwood et al. 2003). However, there exist more complex physiological interactions and consequences than just linear relationships between size, ingested blood volume, and fecundity, as was generally noted for Aedes aegypti (Briegel 1990a, Clements 1992). Some farreaching consequences of variable body sizes relate to the vectorial capacity and to the concept of the gonotrophic cycles in the Anophelini. In a series of classical field studies, Gillies (Gillies 1953, 1954, 1955, 1961, Gillies and Wilkes 1965) reached several important conclusions with respect to An. gambiae. He observed and introduced the pre-gravid status of females approaching hosts in the field, referring to newly emerged unmated females, which led him to suggest that two blood meals were required to mature the first batch of eggs (Gillies 1954). He further maintained that their first blood meal could be taken independent of mating, because reaching the pre-gravid state was considered more important. An. gambiae females were recently observed taking blood meals before mating in the field (Charlwood et al. 2003). Laarman (1968), however, found sugar feeding to be part of the normal feeding behavior of An. gambiae and An. funestus, and more recently, Gary and Foster (2001) investigated the effects of sugar feeding in An. gambiae on its vectorial capacity, as did Okech et al. (2003), who promoted the crucial significance of sugar feeding for survival and thus vectorial capacity of An. gambiae. Recently, Foster and Takken (2004) found that newly emerged An. gambiae prefer sugar over blood. The females used in our experiments were able to mature eggs without being inseminated, even though in other species mating influences egg development (Marchi et al. 1978, Klowden and Chambers 1991, Lounibos 1994). Many of these findings may be explained by our physiological and quantitative studies on body size and reserve status of teneral females. We have found a critical body size (wing length of 2.8 mm) below which the females teneral reserves are insufficient to allow oogenesis with a single blood meal, unless these females had obtained sugar before. Takken et al. (1998) recorded no oogenesis in small females after 4 d of sugar-feeding. Only after 5 d of exclusive sugar-feeding, some of our smallest females were able to mature eggs with a single blood meal. Larger females however, had sufficient lipid and protein reserves available from the larval growth period, so that ingested blood permitted oogenesis with a