Edin expression in the fat body is important for the encapsulation

The major defense mechanism against wasp parasitism in Drosophila is the encapsulation of the invading wasp egg by hemocytes. In a successful encapsulation reaction, the wasp egg becomes surrounded by a melanized capsule and is killed inside it. The encapsulation response requires the concerted action of all of the different hemocyte types. Especially the lamellocytes are considered as the hallmark cells of the encapsulation response, as their formation is induced by a wasp infection.

To study the functional role of Edin in the encapsulation response, we infected fruit fly larvae with wasps and monitored the formation of a melanotic capsule around the wasp eggs 27-29 h after wasp infection, at which point, the capsule is already melanized, but the wasp larva has not yet hatched from the egg. In our control larvae, 45-66% of the infected larvae contained a melanized capsule, which was then set as the normal encapsulation rate for our experiments (six left-most columns in Figure 11C).

Because hemocytes are important players in the encapsulation response, we first knocked down edin in hemocytes using the Hml^∆;He-GAL4 driver. Both edin RNAi lines (edin¹⁰⁹⁵²⁸and edin¹⁴²⁸⁹) were analyzed and were found to have encapsulation rates similar to those of the controls (68% and 60%), indicating that the expression of edin in hemocytes is not necessary for the encapsulation response (Figure 11C).

This finding led us to hypothesize that the expression of edin might be needed in the fat body. This idea was further supported by the qRT-PCR data that showed that edin was highly induced in the fat body upon a wasp infection (Figure 11B).

Therefore, we next used the C564-GAL4 driver to drive the expression of the edin RNAi constructs in the fat body as well as in some other tissues (Harrison et al., 1995). This resulted in significantly reduced encapsulation rates of 6% and 14% for edin¹⁹⁰⁵²⁸ and edin¹⁴²⁸⁹, respectively (Figure 11C). Because the decrease in the encapsulation rate was observed with both RNAi constructs, it was unlikely that the encapsulation phenotype would have been caused by an artefact of the genetic background of the flies or simply by the presence of the RNAi construct.

As is discussed above, the C564-GAL4 driver is expressed in many tissues including the fat body, and in order to confirm the role of the fat-body dependent expression of edin in the encapsulation response, we next used a fat-body specific driver, Fb-GAL4. Again, both edin RNAi constructs were crossed with the driver line and the encapsulation response was analyzed 27-29 h after the wasp infection. As a result, edin knock down with the Fb-GAL4 driver caused a significant decrease in the

encapsulation rates in both edin¹⁴²⁸⁹ and edin¹⁰⁹⁵²⁸ RNAi larvae (7% and 8%) indicating that Edin is required in the fat body, but not in hemocytes, for a normal encapsulation response upon a wasp infection (Figure 11C).

Even if the fruit fly larva manages to successfully encapsulate a wasp egg, it might not be able to kill the wasp inside the capsule. In order to investigate the role of Edin in the killing of the wasp larva, we repeated the experimental setting of the encapsulation assay, but scored for the presence of dead wasp larvae 48-50 h post wasp infection. The wasp larva was considered dead if an intact melanized wasp egg was present in the fruit fly hemolymph without the presence of a living wasp larva.

If the fly larva had not managed to kill the wasp, there was either a living wasp larva with remnants of a melanized capsule or only a living wasp larva present in the hemocoel. Knocking down edin with the C564-GAL4 driver led to a decrease in the killing ability of the fruit fly larvae compared to the controls. With the edin¹⁴²⁸⁹ construct, the proportion of dead wasp larvae was reduced to 8%, when in controls the proportion of killed wasps was 22-38% of all infected larvae (Figure 11D).

However, although the expression of edin¹⁰⁹⁵²⁸ also led to a decrease in the amount of dead wasp larvae when driven with the C564-GAL4 driver (9% of wasps killed), this reduction was not however statistically significant (Figure 11D). Additionally, when edin was knocked down in the fat body with the Fb-GAL4 driver, only the edin¹⁴²⁸⁹ RNAi construct produced a reduction in the amount of killed wasp larvae (8% of wasps killed), whereas the edin¹⁰⁹⁵²⁸ construct did not affect the killing ability of the fruit fly larvae (21% killed wasps) (Figure 11D). Nevertheless, together these results show that the expression of edin in the fat body is required for a normal encapsulation response against wasps.

5.3.2 The role of Edin in the regulation of hemocyte activation upon wasp infection (III)

Drosophila hemocytes, and especially lamellocytes, are key players in the encapsulation response of wasp eggs. To study whether edin played a role in the formation of lamellocytes, we carried out a flow-cytometry based assay to quantify the amount of hemocytes in wasp-infected edin RNAi larvae using the eaterGFP and msnCherry reporter lines. With the reporter lines, plasmatocytes express mostly eaterGFP (green), whereas lamellocytes express msnCherry (red) and can be effectively observed only in the infected larvae (Figure 3A-B’ in III). Again, the expression of edin was knocked down in the fat body using the Fb-GAL4 driver, and the hemocytes from

the infected larvae were collected and analyzed 27-29 h after the wasp infection. Our flow cytometry data showed that edin RNAi larvae had a normal amount of lamellocytes (Figure 12A) and a normal distribution of different hemocyte populations (Figure 12B) when compared with controls, indicating that the impaired encapsulation response was not due to the loss of lamellocytes in the edin RNAi larvae. In the controls, however, the number of plasmatocytes increased significantly in the infected larvae compared to the uninfected larvae, but this effect was not observed in the wasp-infected larvae, where edin had been knocked down in the fat body (Figure 12A). This result supports the idea that the amount of circulating hemocytes does not increase when edin is knocked down in the fat body, most likely because of an insufficient release of hemocytes from the sessile compartment. In fact, when the larvae were visualized under a fluorescence microscope, the banded pattern of sessile hemocytes was lost upon infection in the control larvae, but not in the edin RNAi larvae suggesting that Edin production in the fat body was necessary for the activation of the sessile hemocytes upon a wasp infection (Figure 12C-D’;

Figure 4 in III). These results suggest that the production of Edin in the fat body is required for the efficient release of sessile hemocytes into the circulation and for a normal encapsulation response.

Figure 12. Production of Edin in the fat body is needed for the activation of the sessile hemocyte compartment upon a wasp infection. (A) The number of hemocytes per larva was quantified using flow cytometry and eaterGFP (green) and msnCherry (red) reporter lines 27-29 h after a wasp infection.

Green bars represent plasmatocytes, red bars lamellocytes and black bars non-fluorescent cells. Data were collected 27-29 h post-infection. (B) Proportion of different hemocyte types based on Figure 12A.

(C-D’) Live imaging of Drosophila 3^rd instar larvae of the indicated phenotype. (C-D) Uninfected larvae showing the characteristic striped pattern of sessile hemocytes. (C’-D’) Infected larvae 27-29 h after wasp infection. A wasp infection triggers the formation of lamellocytes and the release of the sessile hemocytes in the control but not in the edin RNAi larvae (D’). - = no infection, + = infection, green = plasmatocytes, red= lamellocytes. (Modified from original publication III)