Kinetics of Plasmodium midgut invasion in Anopheles mosquitoes

Malaria-causing Plasmodium parasites traverse the mosquito midgut cells to establish infection at the basal side of the midgut. This dynamic process is a determinant of mosquito vector competence, yet the kinetics of the parasite migration is not well understood. Here we used transgenic mosquitoes of two Anopheles species and a Plasmodium berghei fluorescence reporter line to track parasite passage through the mosquito tissues at high spatial resolution. We provide new quantitative insight into malaria parasite invasion in African and Indian Anopheles species and propose that the mosquito complement-like system contributes to the species-specific dynamics of Plasmodium invasion.


6
[20] (S1a Fig). We first made sure that expression of the reporters did not 121 interfere with Plasmodium infection. As expected, a significant difference was 122 observed in infection intensity between As and Ag. Regardless of the infection 123 levels, As developed significantly higher oocysts numbers than Ag (S1b Fig). Since the mosquito immune system targets the ookinetes at the basal side of 256 the midgut [24], we examined whether the observed decrease in the proportion 257 of basally located ookinetes was rescued by TEP1 knockdown. TEP1 silencing 258 eliminated the decrease in the basally located ookinetes observed in As and 259 Ag mosquitoes and at the same time increased the proportion of parasites 260 within the cellular layer (Fig 3a). These results suggest that the first wave of 261 invading ookinetes is rapidly killed and lysed by the mosquito immune system. 262 As the parasites that reach the basal lamina at later time points do accumulate, 263 it is possible that asynchronous midgut invasion by Pb exhausts the 264 components of the mosquito immune system and, thereby, benefits the 265 establishment of infection by the second wave of the parasites. These results 266 may also explain why not all parasites are recognized and killed by TEP1 at the 267 basal lamina. We suggest that early crossing parasites may serve as pioneers 268 that attract and locally deplete TEP1, allowing later-coming parasites to survive 269 the immune attack. 270 To better understand Pb invasion dynamics, we measured ookinete motility in 12 20 to 120 min. In line with the previous work [14], we observed four distinct 274 ookinete motility modes: (i) passive floating within the blood bolus (guid 2107, 275 guid 1615, S1 Table), (ii) gliding within the cellular layer (guid 1628, S1 Table)  276 (iii) spiraling in the blood meal and within the cellular layer (guid 1622, guid 277 1624, S2 Table) and (iv) stationary rotation without translocation within the 278 cellular layer (guid 2115, S1 Table). Some ookinetes were observed within a 279 midgut cell for more than one hour, suggesting that the parasites may remain 280 intracellular for relatively long periods of time without inducing cellular 281 apoptosis. By measuring the parasite speed in the blood meal, cellular layer, 282 and at the basal lamina, we found that the speed of ookinetes carried by the 283 bolus content was the highest as compared to other locations (Fig 3b). 284 Interestingly, the speed of the ookinetes in the blood bolus differed between As 285 (8.2 µm/min) and Ag (3.4 µm/min) midguts, suggesting some differences in the 286 blood bolus environment. The ookinete spiraling motility in the cellular layer was 287 much slower in both mosquito species, namely 0.36 µm/min in As and 1.78 288 µm/min in Ag. The slowest stationary rotation movement of parasites was 289 observed at the basal lamina (in As, average speed 0.28 µm/min, guid 2113, 290 S1 Table, in Ag, average speed 0.54 µm/min, guid 1622, S2 Table). We noted 291 that the speed of ookinetes within the cellular layer and at the basal lamina was 292 faster in Ag mosquitoes than in As mosquitoes. This observation indicates 293 important differences in the cellular organization of midguts of the closely 294 related mosquito species. 295 developed an algorithm that classified intracellular, extracellular and 299 intercellular parasites based on the score of their 3D distance to the four 300 nearest neighboring nuclei of the midgut cells. The score was calculated for 301 each parasite (Fig 4a,b). The parasites with the score between 0 -0.45 were 302 defined as extracellular, 0.45-0.55 -as intercellular, and higher than 0.55 -as 303 intracellular. We noticed a proportion of parasites that was extracellular at all 304 time points in both species (Fig 4c). 305 S8-S9). Furthermore, we observed the parasites with very low levels of 344 fluorescence that appeared as a black hole on the background of the midgut 345 cells expressing GFP reporter in Ag mosquitoes (Fig 5a)

365
On average, 10-15% of all recognized parasites had no fluorescence and were 367 classified as dead (Fig 5b). Differences in distribution were observed for live 368 and dead parasites within the cellular layer. More dead parasites were found to 369 be located extracellular or intercellular (compare Fig 5c and Fig 4d). This 370 observation points to more efficient parasite killing of extracellular parasites. 371 Interestingly, we hardly detected any dead parasites in Ag TEP1KD mosquitoes, 372 suggesting that TEP1 may be involved in killing of parasites within the cellular 373 layer. 374 375

Cell damage caused by parasite passage 376
Midgut regeneration is a natural process of epithelia renovation after a blood 377 feeding, whether infective or not [26]. Blood meal generates a stressful 378 environment as it contains bacteria, reactive oxygen species and digestive 379 enzymes that may cause damage to the midgut cells. It has been previously 380 suggested that invaded midgut cells die after invasion and are expelled into the 381 midgut lumen [27] resulting in accumulation of hundreds of cells in highly 382 infected midguts. However, we only once observed GFP positive midgut cells 383 in the midgut lumen. This result indicates that either upon expel dead midgut 384 cells rapidly lose their GFP fluorescence, or that only few midgut cells are 385 expelled after invasion. To resolve these conjectures, we investigated the 386 integrity of the cell layer using high molecular weight Texas-Red conjugated 387 dextran which is trapped inside damaged cells [28]. In these experiments, the 388 fluorescent dextran was delivered into the midgut by blood feeding mosquitoes 389 on mice injected intravenously with fluorescent dextran several minutes before 390 As were predominantly detected in the cellular layer. In contrast in Ag, dextran 394 filled cells were observed both in the cellular layer and in the midgut lumen (Fig  395   6b). As many as 30% of dextran filled cells in Ag, were found in the midgut 396 lumen. Out of these, 50% contained a parasite (S10 Table). In contrast, we 397 found only one (5% of total) dextran filled cell the midgut lumen of As

412
Statistical analysis was performed by a Mann-Whitney non-parametric t-test.

414
Interestingly in Ag, the distance between the dextran-positive cell and the 415 nearest parasite significantly increased at 24-25 hpi compared to the earlier 416 time intervals (Fig 6c). Furthermore, no dextran-positive cells were found in As 417 at the late time interval after infection (24-25 hpi). Taken together, these results 418 suggest that in Ag mosquitoes, damaged cells are readily extruded into the 419 midgut lumen with or without the parasites. It is important to note that while 420 some dextran filled cells contained a parasite, most midgut cells that we 421 observed to host a parasite were dextran-negative, indicating that ookinete 422 invasion damaged and killed only a small proportion of midgut cells. 423 Interestingly, in both As and Ag, we never observed more than one parasite in 424 statistics are performed with MATLAB. 508 The data set was collected from 110 experiments including a total of 2,557 509 parasites (As -45 midguts, 1,068 parasites; Ag -34 midguts, 796 parasites; 510 and Ag Tep1KD -31 midguts, 693 parasites). There was no bias in the number of 511 parasites per midgut across different time points and mosquito species (S2 512