Experimental analysis of the blood-sucking mechanism of female mosquitoes

Only female mosquitoes bite humans to suck blood to obtain protein for their eggs. They have been studied mainly in connection with the transmission of tropical diseases, such as malaria and yellow fever (Dekker et al., 2005; Riehle et al., 2003). To suck blood from a host through a long and slender proboscis, female mosquitoes utilize pumping organs located in their head. The heart of a mosquito controls the circulatory system of the body (Glenn et al., 2010). However, their blood-sucking mechanism has not been studied yet. In addition, the main functions of pumping organs are different from those of the heart that controls the circulatory system of the body (Glenn et al., 2010).

The tiny pump system of female mosquitoes has the ability to suck more than three times the body mass of blood within a short period (Clements, 1992). Rapid intake of a sizeable amount of blood by blood-sucking insects is commonly induced by successive contractions and relaxations of their extrinsic visceral muscles, which function as dilators (Nayar and Sauerman, 1975). Pumping muscles in the head of fluid-feeding insects constitute a single pump (for example in the Reduviidae) or two pumps (e.g. Cicadidae and Culicidae) (Champman, 1998).

Mosquitoes are armed with a two-pump system and their systematic operation allows a large pressure gradient to develop. The two pumps, the cibarial dilator pump (CP) and pharyngeal dilator pump (PP) are located inside their head (Schiemenz, 1957). Blood is extracted from a host and transported to the main body of the insect through a food canal in the proboscis using the pressure gradient generated by the two pumps.

Researchers have attempted to explain the presence of the two pumps, especially in relation to the blood-sucking mechanisms of female mosquitoes. The most persuasive argument to date is that they provide a rapid and highly efficient pumping mechanism that may reduce mortality significantly during the blood-sucking period (Gillett, 1969). Daniel and Kingsolver found that there is an optimal hematocrit at which the rate of protein intake, or the rate of blood cell intake, can be maximized for blood-feeding female mosquitoes (Daniel and Kingsolver, 1983). The intake rate of blood is important because it is directly related to egg production (Kingsolver and Daniel, 1995); however, there have been few investigations on the blood-sucking mechanism and functional features of the two-pump system of the female mosquito. There have been only a few theoretical studies on blood-feeding insects using morphological information and a numerical approach (Reddy and Kesavan, 1988). However, such approaches are limited in determining the in vivo processes (Kesavan and Reddy, 1985).

Recently, the transport of liquid in the food canal was investigated experimentally using a micro-particle image velocimetry (PIV) technique (Lee et al., 2009). Velocity information from the food canal can be used to study indirectly the main features of the blood-sucking phenomena of a female mosquito. In our study, the dynamic behavior of the two pumps in the head of the female mosquito was visualized using a synchrotron X-ray micro-imaging technique to show how the two pumps work during the blood sucking process. It is a powerful tool for observing the internal structure of opaque bio samples (Westneat et al., 2003; Westneat et al., 2008). The volume variation of the two pumping organs was investigated in relation to the streamwise velocity signal of the sucking flow inside the food canal of a mosquito.

Mosquitoes (Aedes togoi Theobald 1907) were reared and maintained in an air-conditioned room at 27°C, 80% relative humidity, with a 16 h:8 h light:dark photoperiod. Larvae were hatched in distilled water and fed a slurry of ground fish food and baker's yeast. After pupation, mosquitoes were transferred to a cage made of a fine-mesh net. Upon emergence, they were fed using a 10% sucrose-soaked cotton rod placed on the bottom of the cage. Mosquitoes more than 4 days post-emergence were selected for the experiments.

The mouthparts (proboscis) of a mosquito consist of six piercing slender stylets that are appressed in a fascicle and this is surrounded by the labium. Among the six piercing stylets, the labrum, which serves as the food canal is the largest and stiffest (Clements, 1992). The labium has a feather-like opaque cuticle that prevents the observation of blood flow through the food canal (Fig. 1A). To visualize blood flow inside using the PIV technique, the opaque cuticle of the proboscis was removed microsurgically with sharp forceps. Through this treatment, the food canal becomes transparent as shown in Fig. 1B.

During the synchrotron X-ray imaging experiments, living mosquitoes were immobilized by sticking their legs with instant glue to a polyimide film, which transmits over 99.9% of X-ray energy. Mosquitoes were handled very carefully in order to minimize the stress caused by sample preparation and fixation on the sample holder.

To determine the temporal variation of the streamwise velocity component of the blood flow in the food canal, we employed a micro-PIV velocity field measurement technique (Adrian, 1991; Meinhart et al., 1999). The micro-PIV system consists of a vertical microscope (Eclipse 80i, Nikon, Nishiohi, Tokyo, Japan), a continuous Nd:YAG laser (λ=532 nm, SLOC, Shanghai, China) as a light source, a high-speed CMOS camera (Photoron ultima APX, Fujimi, Tokyo, Japan) and a personal computer for data processing and control (Lee et al., 2009) (supplementary material Fig. S1). In the preparation of human blood samples, we followed the Declaration of Helsinki and the volunteer gave his consent prior to participation. A whole blood sample from a healthy male volunteer (age 28) was taken and stored in anti-coagulated Vacutainer® tubes (BD, Franklin Lakes, NJ, USA). The blood sample was mixed with fluorescent tracer particles (1.0×10⁹ ml^-1) at room temperature (24°C) and then placed near the tip of a mosquito's proboscis. The fluorescent particles were made of polystyrene (Molecular Probes, Eugene, OR, USA) and were ∼1.0 μm in diameter. Such particles absorb green light at a peak wavelength of λ=534 nm and emit at a peak wavelength of λ=554 nm. Fluorescence images of tracer particles were captured through a high-pass filter (λ=550 nm) attached to the microscope.

Synchrotron X-ray microscopy was used to observe the internal structure of the insects. The experiments were carried out at the third generation synchrotron radiation source of the Pohang Accelerator Laboratory (PLS; Pohang, Republic of Korea). The X-ray beam at the 1B2 beamline was monochromatized using a W/B4C double-multilayer monochromator with peak energy at 10 keV. A scintillator composed of a CdWO₄ crystal was placed 70 mm behind the test sample to transform the X-ray images into visible light. The X-ray images were recorded using a CCD camera (MegaPlusII ES2001, Redlake, Tallahassee, FL, USA) at a frame rate of 30 frames s^-1 (Fig. 2). The field of view (FOV) for the X-ray imaging experiments with a ×4 objective lens was 2219×2219 μm. It took ∼30 s from the setting of a mosquito on the sample stage to the start of X-ray imaging. A starved mosquito usually starts to suck as soon as the food reaches the proboscis. In order to record the blood-sucking process from the start, a diluted iodine solution was injected from outside the hutch using a silicon tube connected to a syringe pump. The sample holder was placed on a motorized stage to be able to remotely control the position of the test sample. To minimize the exposure time of the insect to the X-ray beam in order to maximize its survival time, a mechanical shutter was used so that the test mosquito was only exposed to the X-rays when recordings were being made. Mosquitoes remained alive for up to 5 min during the consecutive X-ray imaging experiments. Mosquitoes usually finished the blood-sucking process within 30 s to 1 min.

The boundaries of the two-pump system could not always be detected clearly because of the complicated anatomy of the pumping organs that are covered with dilator muscles. In order to visualize the fluid flow inside the pumping organs, a diluted iopamidol solution (600 mg ml^-1) was used as a contrast medium. The iopamidol solution has a similar viscosity range to that of the test blood sample, indicating similar shear thinning effects of a non-Newtonian fluid (supplementary material Fig. S2). Therefore, the iopamidol solution represents, reasonably well, the biophysical characteristics of blood under the experimental condition of the present study.

During the blood-sucking process, the deformation index and light intensity in X-ray images of the two pumping organs were carefully examined. The deformation index is defined as the ratio of the major axis to the minor axis of each pumping chamber. Although the temporal variation of the deformation index sufficiently matches the variation of the light intensity, the latter was found to provide clearer signal variation (supplementary material Fig. S3). The light intensity averaged in a square center section (16×16 μm²) of each pump was used to indicate the rhythmical volume variation of the two pumps (supplementary material Fig. S4). To analyze in detail the functional characteristics of the two pumps, the intensity values in the center section for every cycle of each test sample were conditional-averaged by assigning the temporal location of the maximum intensity value in each period as the reference time (t=0). The intensity values used for analyzing the systaltic motion of the pumps are the reciprocal of light intensity, so the increase in light intensity values corresponds to the expansion of the pump volume. As the volume of each pump increases, the X-ray image of the sucked iodine solution in the center region becomes darker, i.e. the light intensity decreases linearly. Time series data were also obtained from each sample to determine the pulsation characteristics such as pulsatile frequency and pumping duration (supplementary material Movie 1). The intensity values of CP and PP were normalized by their maximum values (I₀). The intensity variation of both pumps was recorded for at least 20 cycles.

Fig. 3 shows the internal structure of the food canal and head of a female mosquito sucking iodine solution. At the beginning of the food-sucking cycle, the CP and PP are empty (Fig. 3A). As the CP begins to expand, liquid is sucked into the CP chamber through the food canal, while the PP remains relaxed (Fig. 3B). As a result, the light intensity in the center region of the CP increases gradually as the iodine solution is accumulated. Subsequently, it rapidly decreases as the solution is drained towards the PP through the canal connecting the two pumps (Fig. 3C). Finally, following fast contraction and expansion, the PP completes one feeding cycle. This systaltic muscular motion of the two pumping organs creates a strong hydrostatic negative pressure (Borrell, 2006), sufficient to suck liquid-phase food from the outside into the digestive organ through a long food canal (length ∼2000 μm, diameter ∼30 μm). The pumping process of mosquitoes can be divided into three phases. To analyze these phases in detail, the streamwise velocity signal of the flow in the food canal was synchronized with the corresponding variations in the volume of the two pumping organs in X-ray images (Fig. 4).

Phase 1 corresponds to the period during which the CP takes in fluid from the outside through the food canal (Fig. 4). In this pumping phase, flow velocity in the food canal exhibits a temporary stoppage from the initial low-level velocity. This is attributed to the shrinkage of the end of the elastic food tube that occurs because of the strong negative pressure induced by expansion of the CP. As the shrunken passage near the CP is opened, the flow velocity is increased dramatically to 1.6 cm s^-1. The opening of the shrunken elastic tube by dilator muscles was observed in a previous anatomical study (Schiemenz, 1957). During phase 1, the expansion period of the CP is considerably longer than that of the PP. The expansion time of the CP of female mosquitoes is 2.6 times longer than that of the complementary PP. The CP/PP expansion ratio for male mosquitoes is ∼4.2. It is also worth noting that the contraction times of both pumps are almost the same. The CP/PP contraction ratio is 1.0 and 1.1 for male and female mosquitoes, respectively.

In phase 2, as the CP reaches a crucial threshold volume, the PP start to draw liquid food from the CP when the cibarial-pharyngeal valve (C-P valve), located between the two pump organs, is opened (Schiemenz, 1957; Waldbauer, 1962). This valve is actively controlled by dilator muscles, depending on the operational condition of the two pump organs. It facilitates the flow between the two pumps. In this pumping phase, the maximum expansion of the two pump organs has a definite phase shift (α). This implies that the temporal volume variations of the two pump organs are rhythmically coordinated and synchronized in each feeding cycle. The duration of the CP contraction is almost the same as that of PP expansion. However, the PP starts to expand in advance of the CP, with a time-shift (β; Table 1). When the C-P valve is opened after the CP starts to contract, the volume of liquid sucked up by the PP is somewhat smaller than the volume of the CP. If the time shift is zero (β=0), CP contraction and PP expansion occur simultaneously. In addition, the reverse flow discharged from the CP chamber toward the food canal causes negative inertia against the main feeding flow. In this case, the PP expends more energy than in the earlier expansion of the PP, with a pertinent time shift of β. The earlier expansion of the PP with β saves energy by adding inertial force to the main flow. It also accelerates the expansion and contraction process of the CP, increasing the feeding frequency. Consequently, mosquitoes can suck large amounts of liquid efficiently in a short period.

A slight reverse flow (-0.15 cm s^-1) was observed in the food canal during phase 2. However, the amount of the reverse flow was very small, ∼1% of the total amount of liquid sucked up by CP in phase 1. In phase 3, the PP pushes liquid towards the body of the insect and the CP discharges the remaining fluid in both directions. The velocity of reverse flow inside the food canal in phase 3 is slightly smaller (-0.27 cm s^-1) than that in phase 2. This reverse flow from the CP to the food canal is generated because the negative suction pressure caused by the dilator muscles of the PP is not sufficient in phase 3. However, the amount of reverse flow in phase 3 is also very small, that is, ∼3% of the total amount of liquid sucked up by the CP in phase 1.

Fig. 5 compares the cyclic volume variation of the two pumps of a representative female mosquito with that of a typical male mosquito (supplementary material Fig. S5). The experimental data were statistically averaged from 10 samples for each species and the results are summarized in Table 2. The liquid-sucking period of male mosquitoes is ∼202 ms and that of female mosquitoes is 160 ms. The total amount of liquid sucked up by mosquitoes can be estimated by multiplying the volume of liquid drawn by the CP in one period (net flow rate; Q_net) by the number of pumpings (f×t_tot, where f is the pumping frequency and t_tot is the total feeding time for liquid food). Q_net is equal to the amount of liquid drawn by the CP in phase 1. The pumping frequency of female mosquitoes is 4.1±1.3 Hz (mean ± s.d.) and that of male mosquitoes is 2.9±0.8 Hz.

These results demonstrate that the pumping performance of female mosquitoes is better than that of male mosquitoes. Considering the duration of each phase, as summarized in Table 1, phase 1 is the major phase governing the pumping performance of mosquitoes. In the dominant phase 1, the PP is in the relaxed idle state, while the CP is in the expansion state. The CP has a short relaxation time as well. On average, phase 1 of male mosquitoes (266 ms) was 1.6 times longer than that of female mosquitoes (171 ms). For male mosquitoes, the idle state in phase 1 occupied 24% of the feeding period, whereas that of female mosquitoes was merely 6%.

In this study, we employed advanced flow visualization techniques to analyze the functional features of the two pumping organs in the head of female mosquitoes. By synchronizing the velocity signal of the flow inside the food canal, the volume variation of the two pumping organs was classified into three phases. To our best knowledge, this is the first study to observe in vivo the processes involved in female mosquitoes sucking viscous liquid during feeding.

There have been a few earlier studies on the blood-sucking phenomena of mosquitoes, but these were based on anatomic observation (Patton and Cragg, 1913; Schiemenz, 1957). Pappas analyzed the electromyography signals of food-sucking mosquitoes and reported that the CP initiates the pumping sequence earlier than the PP (Pappas, 1988). This report agrees well with the present results, even though there is no detail description on the blood-sucking mechanism. The present study made it possible to explain the detailed phasic variation of each pump of the two pumping organs and the functional characteristics of each pump, based on direct visualization of the pumping organs of blood-sucking female mosquitoes.

In the normal blood-sucking process of a mosquito, the unevenness and stoppage of flow due to the non-Newtonian aspect of blood, are disadvantageous because the blood viscosity increases as the flow speed decreases or under reversed flow conditions. The two pumps of a mosquito seem to contribute to maintaining the forward flow. As a result, unevenness and stoppage in the flow are minimized in the three phases of the systaltic motion of the two pumps. In phase 1, the PP sucks the liquid with forward inertia due to the earlier expansion with a time shift (β). The expansion of the PP contributes to reduce reverse flow, compared with the degree of reverse flow in phases 2 and 3. At the end of phase 3, the CP starts to expand while the PP contracts, accumulating blood in the CP and then transporting it to the PP.

Morphologically, a two-pump system is necessary. The CP is located at a curved in the path, where minor energy loss occurs. This curved flow path is inevitable for the connection of the proboscis and the digestive organs of the main body. If the flow path from the food canal to the main body was straight, a flying mosquito would encounter many difficulties in hunting and piercing the skin of a host to suck blood (Fig. 6). If mosquitoes were equipped with only one pump, it would be difficult to draw food from outside and to regulate the flow effectively.

In summary, the functional characteristics of the two-pump system of mosquitoes were analyzed in detail during the sucking process by estimating the volume variation of the two pumping organs and the velocity change of blood flow in the food canal. Mosquitoes suck viscous liquid very efficiently using the systematic operation of the two pumping organs (CP and PP) and a smart valve located between CP and PP. The systaltic motion of the two pumps has a phase shift, α and one food-feeding period is divided into three phases. It is noteworthy that the second pump (PP) starts to expand in advance, with a time-shift, β. This earlier expansion of the PP reduces the pumping period and saves energy used for regulating the flow through the two pump organs.

Based on statistical analysis, the pumping performance of blood-sucking female mosquitoes is superior to that of nectar-eating male mosquitoes. A major difference results from the sustained muscular motion of the CP of female mosquitoes during phase 1. In conclusion, the two pumps of female mosquitoes work systematically in a well-coordinated manner.