Superb feeding behavior of Aedes albopictus transmitting Zika virus

Disease-mediated mosquitoes have been receiving much attention, as the World Health Organization recently declared the Zika virus a global public health emergency. Mosquitoes transmit pathogens that cause various tropical diseases including malaria, dengue fever and yellow fever as well as Zika virus. The vector efficiency of mosquitoes depends on their blood-feeding characteristics and the mechanics of their blood-sucking pump system, but only a few studies have attempted to investigate these key issues. In this study, we demonstrate the rapid and gluttonous liquid-feeding characteristics of Ae. albopictus which transmits Zika virus can be explained by similar proportion of two blood-sucking pumps and accelerated liquid intake driven by fast expanding of pumps. Our results provide insight into the vector efficiency of Ae. albopictus in terms of feeding velocity, pumping frequency, liquid-intake rate, and wall shear stress.


Introduction
Mosquitoes transmit various pathogens that cause infectious tropical diseases including dengue fever, malaria, filariasis, and Zika virus [1], which was declared a public health emergency of international concern on Feb 2016 [2]. The World Health Organization estimates that there are currently 500,000 to 1.5 million cases of Zika in the Americas and that it is spreading explosively [3].
The Aedes albopictus mosquitoes, which are found in the USA, southern Europe, Asia, and South Korea, also transmit Zika virus [2,4]. This species has been known as a very aggressive biter [5]. It also transmits Japanese encephalitis, West Nile virus, yellow fever, and dengue virus [5][6]. Aedes togoi occurs from subarctic to subtropic area along the coasts of the East Asian contries [7]. It is mainly involved in the transmission of numerous species of filariae, yellow fever, and Japanese encephalitis [8][9][10].
The vector efficiency of mosquitoes could be related to their blood-feeding behaviors [10][11]. A previous study investigated nectar-feeding of Ae. albopictus and Ae. togoi mosquitoes [12]. This nectar-feeding was appeared to be gradually activated within days after emergence. The feeding rate of Ae. togoi reached a peak value of 97.9% on day 9 after emergence. On the other hand, both male and female Ae. albopictus reached 80-90% feeding rate while they had a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 been fed during 8hr after the nectar-feeding [12]. However, the mechanisms underlying these liquid-feeding behaviors remain unclear.
The blood-feeding behaviors are known to be controlled by two pump organs in the head, the cibarial dilator pump (CP) and the pharyngeal dilator pump (PP) [13]. The systaltic motions of the CP and PP generate high differential pressure for sucking highly viscous blood [14]. Therefore, the actions of these pumps combined with salivation directly affect the efficiency of pathogen delivery by mosquitoes [10,[15][16]. During the blood-feeding process, pathogen intrusion can be influenced by probing behavior, blood intake rate, feeding persistence, and locomotor activity [10][11][17][18][19]. In addition, wall shear stress (WSS) is an important pathogenic factor that affects the transport of blood through conduits and all substances through vessel walls located near the tip of the food canal [10]. The adhesion and movement of the pathogens are critical for reproduction [20]. In our previous study, Ae. togoi generates higher feeding pressure with bigger volume variation of pumping organs, compared with Ae. sinensis [10]. In addition, Ae. togoi has high WSS in the feeding process. The high WSS is beneficial for removing pathogen attached to the host's vessel wall [21][22]. However, the relationship between the high WSS of Ae. togoi and the mosquito's vectorial capacity has not been well known yet. Unlike Ae. togoi, mechanical characteristics of Ae. albopictus, which are closely related to the feeding behaviors and pathogen transmission, remain poorly understood.
In this study, the feeding pump organs and liquid-feeding mechanics of Ae. albopictus were experimentally investigated and compared with those of Ae. togoi. The 3D morphological structures of the pumping organs of Ae. albopictus and Ae. togoi were examined using synchrotron X-ray microscopic computed tomography (SR-μCT). The volumetric variations and proportions of the blood-feeding pumps of the two Aedes mosquitoes were quantitatively compared. Blood-feeding parameters, such as flow rate (Q), WSS, and pumping frequency were measured in the two Aedes mosquitoes using a micro-particle image velocimetry (micro-PIV) technique to distinguish their vector efficiencies with respect to the blood-feeding flow characteristics. In addition, 2D X-ray images of the systaltic movements of the two pump organs were also investigated to examine their phasic volumetric variations and reveal the feeding mechanism of Ae. albopictus mosquitoes. These experimental results will elucidate the bloodfeeding characteristics of Ae. albopictus in comparison with those of Ae. togoi.

Mosquito rearing and sample preparation
Following established rearing procedures [23], mosquitoes (Aedes albopictus (Diptera: Culicidae) and Aedes togoi Theobald 1907) were fed a 10% sugar solution at 27˚C, 80% humidity and a 16 h: 8 h light/dark cycle. Mosquitoes were provided by Korea Centers for Disease Control and Prevention. Larvae were fed a slurry of ground fish food and baker's yeast once a day. Approximately 150 to 200 larvae were reared in a 2L plastic container before the beginning of pupation. After pupation, the mosquitoes were transferred to a cage and provided a 10% sucrose-soaked cotton rod. The time of pupation was approximately 2 days. Synchrotron X-ray microscopic computed tomography (SR-μCT) All SR-μCT experiments were conducted at the Biomedical Imaging beamline of Pohang Light Source-II in Korea. A white beam emitted from the multi-pole wiggler source of the 3-GeV storage ring was filtered through 1-mm-thick graphite after passing through a double multilayer monochromator that provided photon energies in the range of 10 KeV to 50 KeV. The optimal monochromatic X-ray energy for the tomographic scanning experiments in this study was experimentally determined to be approximately 24 KeV. The detector was positioned 20 cm downstream from the test sample to capture phase-contrast images. The monochromatic X-ray images transmitted through the test sample were consecutively recorded using an imaging detector (Andor Zyla, Andor, Oxford Instruments, Belfast, UK) with a spatial resolution of 2560 × 2160 pixels, upon which a YAG:Ce (30-μm thick) scintillation crystal was adhered. The captured X-ray images were converted on the scintillator into visible images. Each mosquito sample was immersed in ethanol and placed at the tip of a heat-sealed pipette made of polypropylene, which has a comparatively low X-ray absorption coefficient [24]. Parafilm was used to prevent the alcohol from evaporating. Each test sample was placed on a rotating stage, and 2D tomographic slice images of the samples were captured by rotating the stage at 0.5˚-intervals from 0˚to 180˚. The 3D morphological structure of each mosquito sample was reconstructed from the captured tomographic slice images using Octopus image-processing software. To obtain a 3D volumetric image of the sample, a filtered back-projection algorithm was applied to the images. Cross-sectional images were stacked to animate a 3D structural image. The stacked images were rendered to show the 3D morphological structure using Amira1 5.3.3 image-analysis software (Visualization Sciences Group, Burlington, MA, USA).

Scanning electron microscopy
Scanning electron microscopy (SEM) was employed to illustrate the morphological configurations of the mosquito samples tested in this study. Mosquito specimens were prepared by air-drying. The samples were then Ag-coated using a coater (Quorum Technology, SC7640 mode, East Sussex, United Kingdom) and examined by a field emission SEM (XL30S FEG, Philips Electron Optics B.V., the Netherlands) connected to an EDXS system at an acceleration voltage of 5 kV.

Velocity and flow measurements
To quantitatively visualize the flow in the food canal of the two mosquito species, a micro-PIV technique was used. More than 50 female mosquitoes of similar size were used to acquire experimental information for each mosquito species. To visualize feeding liquid-flow in the food canal of a female mosquito, the opaque cuticle was removed by microsurgery.
A 1% (w/w) sucrose solution (density; ρ = 1002.1kg/m 3 , dynamic viscosity; μ = 0.927 × 10 −6 m 2 /s) at 25˚C was used as the working flow. Kikuchi and Mochizuki measured blood flow rate (Q) in the food canal of a female mosquito and compared with their previous result acquired by using a sucrose solution. The velocity profiles in the food canal are almost parabolic, regardless of feeding liquid [25]. This implies that the blood-feeding flow in a female mosquito's food canal can be considered as a Hagen-Poiseuille flow.
To obtain fluorescent images of the feeding flow, 1.0 μm-diameter fluorescent particles (Molecular Probes, Eugene, OR, USA) were diluted with the working fluid. An Nd:YAG laser (λ = 532 nm, SLOC, Shanghai. China) was used as a light source for the PIV system to extract information related to the blood-sucking velocity field with high accuracy. The fluorescent image (λ = 554 nm) of the tracer particles was passed through an objective lens (M = 20, N.A. = 0.5), and after high-pass filtering with an optical filter (λ = 550 nm), it was attached to the front of the microscope (Eclipse 80i, Nikon, Tokyo, Japan). Flow images were consecutively recorded by a high-speed camera (Fastcam SA1.1, Photron, USA). We captured consecutive images at a frame rate of 10,000frames −1 . The field of view is 256μm × 128μm and the spatial resolution of the PIV system is about 1μm/pixel. The interrogation window size is 16 × 8pixels. Personal computer was used to store the captured flow images and process the data. To estimate whether the tracer particles reliably represented the flow, the Stokes number, St, (ρ p a 2 U mean /18μD h ) was calculated, where ρ p is the density of the tracer particles; a 2 is the diameter of the tracer particle; U mean is the mean velocity of the flow; μ is the dynamic viscosity of the working fluid; and D h is the internal diameter of the food canal. The Stokes number was estimated to be approximately 10 −3 , which is much smaller than 1. This estimate indicates that the tracer particles can be used as reliable flow tracers. All experiments were conducted under the same environmental conditions (25˚C, 45% RH).

Statistical analysis
All data were statistically analyzed using t-tests with 95% confidence intervals in SPSS (IBM, Chicago, IL, USA).

Morphological differences of the proboscises and heads of two different female mosquitoes
A female mosquito sucks blood which contains pathogens through its proboscis by operating two pump organs located in the heads. We examined the morphological characteristics of the proboscises and heads of two female mosquitoes: Ae. albopictus and Ae. togoi (Fig 1). The length of the proboscis of Ae. albopictus is shorter than that of Ae. togoi and the head size of Ae. albopictus is smaller than that of Ae. togoi. The length of the proboscises of two mosquitoes are 1.19 ± 0.13 mm and 2.11 ± 0.05 mm, respectively. The height of their head parts are 0.51 ± 0.04 mm and 0.71 ± 0.05 mm, respectively.

Comparison of 3D morphological structures of the blood-feeding pump organs
The 3D morphological structures of the blood-feeding pump organs of two mosquito species under static state were investigated by using synchrotron X-ray computed tomography (Fig  2A to 2D). The blood-feeding pumps of mosquitoes constitute the CP and the PP [13]. The two pump organs are connected by a conduit (C-P) (Fig 2A and 2B). The CP positioned under the curved clypeus is directly linked to the proboscis (Fig 2C and 2D). The PP is located behind the CP and supported by dilator muscles.
The pump organs of Ae. albopictus are smaller than those of Ae. togoi at similar developmental stage. The total volume of the two pump organs of Ae. albopictus and Ae. togoi is 0.84 ± 0.26 × 10 −3 mm 3 and 1.96 ± 0.54 × 10 −3 mm 3 , respectively. The total volume of Ae. albopictus is about 43%, compared to that of Ae. togoi (p < 0.01) (Fig 2E). The volume of each feeding organ of Ae. albopictus is also smaller than that of Ae. togoi. However, the proportions of individual feeding pump organs in Ae. albopictus are quite different from those in Ae. togoi (Fig 2F). The proportions of the CP, C-P, and PP volumes are approximately 42.9%, 12.6%, and 44.5%, respectively, in Ae. albopictus, whereas those in Ae. togoi are 20.8%, 24.6%, and 54.6%. The volume of CP in Ae. albopictus took up approximately 40% of the total volume, which is similar to the proportion of PP. The proportion of CP in Ae. albopictus is relatively high, compared to Ae. togoi.

Contraction and expansion of CP during feeding process
The consecutive 2D X-ray images of the head parts of two mosquito species show the variations in the cross-sectional area of the CP (Fig 3). The boundaries of the CP are marked with dashed lines to distinguish its maximum and minimum cross-sectional areas. The expansion rate of CP area obtained by dividing the maximum area (A max ) by the minimum area (A mim ), are 195.6 ± 8.5% and 145.8 ± 5.4% for Ae. albopictus and Ae. togoi, respectively (p < 0.01). The volume expansion rates for two different species of mosquito were estimated under assumption of CP as a rotating body. The volumetric expansion rate of CP in Ae. albopictus is 386.6 ± 31.2% and 228.2 ± 23.5% in Ae. togoi. The overall expansion rate of CP is much higher in Ae. albopictus than in Ae. togoi (p < 0.01). Both mosquitoes exhibit periodic liquid suction and ejection characteristics. The start of feeding (t = 0) is determined to be the point when the sign of flow velocity changed from negative (-) to positive (+). The average feeding velocity of Ae. albopictus is 0.14 ± 0.12 m/s, which was much faster than that of Ae. togoi (0.04 ± 0.03 m/s) (p < 0.01). The pumping frequency of Ae. albopictus is 10.5 ± 0.2 Hz, which is also faster than in Ae. togoi (6.1 ± 0.2 Hz) (p < 0.01). Ae. albopictus exhibits fast and frequent liquid sucking, compared to Ae. togoi.

Feeding characteristics of the two mosquito species
The liquid flow in the food canal of mosquito during liquid-feeding is analogous to a simple internal flow in a circular tube. The Reynolds number (Re) of the flow inside the food canal is

Feeding behavior of Aedes albopictus
To investigate the liquid-feeding characteristics of the two mosquito species, phasic variations of flow rate Q(t) were obtained from the liquid-feeding velocity information in one pumping period (T 0 ) (Fig 5). Assuming the liquid flow through food canal as the Hagen-Poiseuille flow, the flow rate (Q) can be estimated as follows: where L is the length of the food canal; Δp is the differential pressure generated by the two and Ae. togoi (d). The cross-sectional CP area was estimated from the X-ray images captured by the synchrotron X-ray micro-imaging technique. A diluted iodine solution was used as the feeding fluid to enhance image contrast. Interestingly, the liquid-feeding features at the ejection phases of the two mosquito species are quite different. The ejection phase start when 76.7% and 86.8% of the time has elapsed from the start of one cycle for Ae. albopictus and Ae. togoi, respectively (p < 0.01). However, the average flow rates in the ejection period are quite similar for the two mosquito species; the ejection flow rate of Ae. albopictus is -6.3 ± 1.0 nl/s and that of Ae. togoi is -7.2 ± 0.5 nl/s (p < 0.01).
We compared the WSS of the two species of mosquito exerted on the wall of food canals during the liquid-feeding process. Assuming that the Hagen-Poiseuille flow holds, WSS (τ w ) was calculated as follows: The WSS increases with increasing flow rate Q(t) and decreasing food canal diameter (D). The probability density function (PDF) distribution of WSS are compared for two different species of mosquito during their intake period (Fig 6). The PDF distribution for Ae. togoi has the local maximum peak at a relatively low WSS of 8.6 Pa. However, the PDF of WSS for Ae. albopictus is expressed as Gaussian distribution with a peak at approximately 21.7 Pa. During the liquid intake process, the average WSS of Ae. albopictus is 21.2 ± 0.4 Pa, which is approximately 2.5 times higher than that of Ae. togoi (8.2 ± 1.2 Pa) (p < 0.01). The WSS of Ae. albopictus during the entire feeding period (14.9 ± 1.0 Pa) is also much greater than that of Ae. togoi (6.6 ± 0.9 Pa) (p < 0.01). Feeding behavior of Aedes albopictus

Discussion
In this study, we analyzed the liquid-feeding mechanics of the two pumping organs in the head of female mosquitoes: Ae. albopictus and Ae. togoi. In particular, we demonstrated that Ae. albopictus, the specific species of mosquito which is notorious for mediating the transmission of Zika virus, is a fast and gluttonous feeder. Sant'Anna et al reported that the number of non-ingestive events on the artificial feeder was significantly different from that of observation from live hosts [26]. Thus, it needs to pay attention when we draw conclusions about feeding characteristics from the results acquired by using an artificial meal. In addition, in nematocera diptera, carbohydrate diets are directed to Feeding behavior of Aedes albopictus the diverticulum which is attached to the anterior gut rather than to the midgut as it occurs with the blood [27]. However, diet destination can be affected by midgut valve and chemosensory-based control of the crop [28].
Ae. albopictus rapidly feeds the liquid (1% (w/w) sucrose solution) in spite of a relatively small stroke volume of small sized two-pumps. Kim et al investigated the effect of fluid viscosity on the liquid-feeding of female mosquitoes by measuring the intake flow rate and stroke volume with varying sucrose concentration. They reported that there is no significant difference in the ranges of sucrose concentration between 1% and 30% and 30% and 50% [13]. The total volume of the two pump organs of Ae. albopictus is found to be approximately half the volume of Ae. togoi. The stroke volume of Ae. albopictus is also 73% smaller than that of Ae. togoi. However, the pumping frequency of Ae. albopictus is about 1.7 times higher than that of Ae. togoi. High feeding frequency of Ae. albopictus means that the species of mosquito can complete the expansion and contraction processes of the pumping organs to fill the pump chamber with liquid in a very short time, compared with Ae. togoi. This trait may significantly affect the blood feeding and fast pathogen transmission via the blood feeding. Ae. albopictus is likely to be a more proficient vector because it feeds more rapidly than Ae. togoi.
The initial instantaneous liquid sucking of Ae. albopictus is significantly high at the intake phase. The incremental flow rate of Ae. albopictus during the first 5% of the initial period of one cycle is 15.0 nl/s, which is approximately 300% higher than that of Ae. togoi (4.5 nl/s). The fast liquid-feeding of Ae. albopictus at the start of liquid-intake phase may be closely related to the large proportion of the CP to the total pumping system. The volume of CP accounts for a larger proportion in Ae. albopictus, which means that the variations in volume with the systaltic motion of the CP in Ae. albopictus is significantly higher than that of Ae. togoi. The expansion rate of CP area in Ae. albopictus is approximately 1.3 times larger than that in Ae. togoi. The faster incremental flow rate of Ae. albopictus at the start of the intake phase leads to the rapid filling of the CP volume.
Ae. albopictus has small two-pump chambers, rapidly feeding liquid with small stroke volume, but the accumulated amount of liquid feeding of Ae. albopictus is 122% larger than that of Ae. togoi during the same time. The effective liquid-feeding of Ae. albopictus may be attributed to the similar proportions of its CP and PP. Because both pumps in the two-pump system have the similar size, the mosquito can easily feed and transmit liquid using the same operating frequency [29]. In addition, large expansion rate of CP may contribute to effective pumping performance of Ae. albopictus. Large displacement of pump chamber can generate the large volume flow rate per stroke [29]. The high WSS of Ae. albopictus can contribute to vector competence, which represents the infection and transmission abilities of pathogens through blood flow [30]. Adhesive force of pathogen reflects the extraction of elongated tethers by the macrophage membrane. The macrophage membrane immerses the pathogen quickly at the end of initial contact [31]. The shear-induced mechanical drag force in the cell surface cause a highaffinity conformation of the adhesion and decrease the bond off-rate [32]. If the WSS is higher than the adhesive force of the pathogen on the vessel wall, the pathogen can be detached and easily transmitted from hosts to mosquitoes [33]. The extensive WSS might lead to the spread of pathogen by transforming and eventually destroying infected red blood cells [33]. Due to the distinctive WSS difference of Ae. albopictus, these two Aedes mosquitoes potentially have different pathogen transmission efficiencies during sucking.
Mosquito-mediated transmission of Zika virus begins then a Aedes mosquito inoculate the virus to the permissive cells of host. Zika virus replication activates an antiviral innate immune response and produces the type I interferons in infected cells [34]. The mechanism underlying specific Zika pathogen adhesions will be helpful to understand the adhesive force with WSS of Ae. albopictus. However, Zika has no vaccines and drugs to treat the infection [35]. Further study of biological factors would be helpful to further understand the pathogen transmission characteristics of Ae. albopictus.
In this study, the liquid-feeding characteristics and the morphological change of the twopump system of Ae. albopictus during liquid-feeding were experimentally investigated and compared with those of Ae. togoi. This study demonstrates the rapid and gluttonous liquidfeeding characteristics of Ae. albopictus. Effective liquid feeding of Ae. albopictus could be explained by structural characteristics and operating characteristics of the pumping organs: similar proportion of CP and PP organs and accelerated liquid intake via fast expanding CP. These results also provide insight into the vector efficiency of Ae. albopictus in terms of feeding velocity, pumping frequency, intake rate, and WSS.
Supporting information S1 File. Data set about velocity field information. Liquid-feeding flow velocity signals in the food canals (Fig 4) is based on this data set. (XLSX) S2 File. Data set about phase-averaged flow rate. Phasic variations in the flow-rate profiles of the food canal during one pumping cycle (Fig 5) is based on this data set. (XLSX) S3 File. Data set about cumulative wall shear stress (WSS). Probability density function (PDF) distribution of WSS during the intake period (Fig 6) is based on this data set. (XLSX)