Activated hemocytes increase the expression of carbohydrate transporters and trehalase

Since Drosophila hemocytes increase their uptake of carbohydrates during parasitoid wasp infection [4], we analyzed the expression of carbohydrate transporters in the hemocytes. The SLC2 family of hexose sugar transporters in Drosophila comprises 31 genes (S1 File; FlyBase ID: FBgg0000691), most of which have not been functionally characterized. Our transcriptomics shows that the expression of 10 of these genes is greater than 3 transcripts per million (TPM) in hemocytes (S1 File and S1 Table) and 4 of these (Tret1-1, CG4607, sut1, and CG1208) become expressed in hemocytes during infection (Fig 1A). Tret1-1 has been characterized as a glucose and trehalose transporter [19] and CG4607 appears to be involved in lysosomal glucose metabolism [20]. Both Tret1-1 and CG4607 are weakly expressed in hemocytes in the uninfected state, but their expression strongly increases upon infection (Fig 1A and S1 File). Based on scRNAseq [9], both are expressed exclusively in lamellocytes (S1 File). There are 2 mRNA variants of Tret1-1, with Tret1-1-RA increasing 31-fold upon infection, whereas there is no increase in the Tret1-1-RB variant (S1 File). sut1 is highly expressed in most hemocyte types in both uninfected and infected states (Fig 1A and S1 File). MFS3, which belongs to the SLC17 family of organic anion transporters, has also been shown to transport glucose and trehalose [21] and its high expression in uninfected hemocytes decreases during infection (Fig 1A and S1 File). The as yet functionally uncharacterized CG1208 shows the strongest increase in expression upon infection, predominantly in lamellocytes (Fig 1A and S1 File). Thus, MFS3 and sut1 may provide basal carbohydrate transport in most hemocyte types in the uninfected state, whereas Tret1-1 and presumably CG4607 and CG1208 likely enhance carbohydrate transport in lamellocytes during infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Analysis of carbohydrate metabolism gene expression.

(A) Expression heat map (bulk RNAseq) of selected transporters and trehalase in circulating hemocytes, lymph gland and wing disc from uninfected (Uninf) and infected (INF) third-instar larvae collected 9 and 18 hpi (0 h = 72 h after egg laying). Mean values given in each cell are TPM—data in S1 Table. (B) Expression of the trehalase (Treh) gene (bulk RNAseq) in circulating hemocytes and lymph gland. Each dot represents a biological replicate in TPM, bars represent mean ± SEM. Samples were analyzed using DESeq2 in Geneious Prime (S1 Table), **** shows adjusted P < 0.0001, ns = not significant. (C) Transcript-specific analysis of Treh expression by RT-qPCR 18 hpi. cTreh, represents cytoplasmic trehalase (primers cTreh-F and Treh-R shown in (D)), increases 35-fold after infection. sTreh represents secreted trehalase (primers sTreh-F and Treh-R shown in (D)). Bars show fold change compared to uninfected cTreh samples (expression levels were normalized by RpL32 expression in each sample), each dot represents a biological replicate. An unpaired two-tailed Welch’s t test was used to compare uninfected and infected samples; ****P < 0.0001 (numerical values in S1 Data). (D) Map of the Treh gene with individual transcripts (RA-RG). Lines indicate introns, boxes represent exons with coding sequence in orange. Labeled arrows show primers used for RT-qPCR expression analysis. Schematic representation of a Gal4 knock-in into the first exon of Treh-RA, creating a 47-base deletion that removes both cTreh start codons, replaced by a cassette containing the Gal4 coding sequence and an RFP marker, expressed downstream of the P3 regulatory sequence in the fly eye. The resulting fly strain is Treh[RAΔG4]. (E, F) General Treh expression analysis of hemocytes from control and homozygous Treh[RAΔG4] mutant by RT-qPCR 18 hpi using primers common for all transcripts (Treh-F1 and Treh-R1 shown in (D)). (G) Treh-RB and RC transcript specific expression analysis of hemocytes from control and Treh[RAΔG4] mutant by RT-qPCR 18 hpi using primers specific to Treh-RB/RC (Treh-RB-F and Treh-R shown in (D)). (E–G) Bars show fold change compared to uninfected control samples (expression levels were normalized by RpL32 expression in each sample), each dot represents a biological replicate. An unpaired two-tailed Welch’s t test was used to compare samples; *P < 0.05, **P < 0.01, ****P < 0.0001 (numerical values in S1 Data). (H) Analysis of cTreh expression in hemocytes at 40 hpi. Treh[RAΔG4], expressing Gal4 in the cTreh expression pattern, drives UAS-mCherry expression (red) together with a lamellocyte-specific marker MSNF9-GFP (green) in differentiated lamellocytes but not in plasmatocytes. hpi, hours post infection; PPP, pentose phosphate pathway.

https://doi.org/10.1371/journal.pbio.3002299.g001

While sut1 is moderately expressed in the lymph gland, we did not detect increased expression of transporters other than CG1208 at 18 hours post infection (hpi) (Fig 1A). MFS3 and sut1 also appear to be the major carbohydrate transporters in the wing discs which, however, display no changes in carbohydrate transporter expression during infection (Fig 1A).

Notably, in addition to the transporter Tret1-1, expression of the enzyme Treh, which converts trehalose to 2 glucose molecules, is strongly up-regulated in hemocytes upon infection (Fig 1A and 1B), primarily in lamellocytes (S1 File). Since Treh occurs in 2 forms, cytoplasmic (cTreh) and secreted (sTreh; Fig 1D) [22], we used transcript-specific qPCR to distinguish between them. Multiple cTreh transcripts (RA, RD, RG, RE) increased 35-fold in hemocytes upon infection (Fig 1C), whereas those encoding sTreh (RC, RF) were induced by infection only slightly above the basal level of cTreh. Therefore, substantially more cTreh than sTreh is expressed in hemocytes during infection.

We generated a cTreh specific mutant, Treh[RAΔG4], by replacing 47 bp including the first 2 start codons with the Gal4 coding sequence, which can also serve as a cTreh expression reporter (Fig 1D). The homozygous Treh[RAΔG4] mutant shows a phenotype similar to the cTreh-specific mutant Treh[c1] [22], with 20% lethality during pupal development and two-thirds of adult flies becoming lethargic and dying within 3 days of eclosion. Using a heterozygous Treh[RAΔG4] line crossed with flies carrying UAS-GFP, we found lamellocyte-specific expression of cTreh (S1 Fig) consistent with scRNAseq data. cTreh was also expressed in other larval tissues, such as imaginal discs (S1 Fig) and brain (S2 Fig), with and without infection. To verify cTreh expression in lamellocytes, we combined Treh[RAΔG4] driven UAS-mCherry with a lamellocyte-specific marker MSNF9-GFP (Figs 1H and S3) and found that cTreh is indeed specifically expressed in fully differentiated lamellocytes. While the expression of MSNF9 is already visible at 18 hpi in not yet fully differentiated lamellocytes (S3 Fig), cTreh starts to appear at 24 hpi (S3 Fig) and is expressed in most lamellocytes at 40 hpi (Fig 1H), including in those encapsulating the parasitoid egg, which is already being melanized (S3 Fig).

Since the increased Treh expression during infection is mainly owing to the Treh-RA transcript, we expected that this increase would be abolished in the homozygous Treh[RAΔG4] mutant. The overall expression of Treh was reduced to one-third in uninfected Treh[RAΔG4] homozygotes (Fig 1E), but was still up-regulated nine-fold in the infected mutant (compared to a 25-fold increase in the wild type; Fig 1F). Using transcript-specific qPCR, we found that there was a compensatory expression from the transcription start site common to both Treh-RB (cTreh) and Treh-RC (sTreh) transcripts, which are weakly expressed in wild-type hemocytes (Fig 1G). Thus, while there is a strong increase in cTreh expression in lamellocytes during infection from the Treh-RA (or RD, RG, RE) transcripts, loss of this transcript is compensated by the expression of Treh-RB.

Our expression analysis showed that hemocytes express multiple carbohydrate transporters, some of which are dramatically induced during infection. This is consistent with our previous results showing enhanced sugar consumption under these conditions. The combination of highly elevated expression of trehalose transporters and of the cytoplasmic form of Treh exclusively in lamellocytes, together with the compensatory expression of cTreh, suggests the importance of trehalose metabolism in these cells.

Our bulk transcriptomic analysis showed that all enzymes associated with glycolysis and PPP are strongly expressed in hemocytes in both the uninfected and infected states (Fig 2 and S2 File and S1 Table). The levels of specific mRNAs in each pathway correspond well with scRNAseq analyses (S2 File; [8,9]). The same glycolytic and PPP genes are similarly expressed in the lymph gland and wing disc with or without infection (S2 File and S1 Table). Combining our bulk transcriptomics and scRNAseq data revealed a tendency for a slight decrease in the expression of glycolytic genes in most prohemocytes and plasmatocyte-like cells during infection, while the expression shifted towards lamellocytes and crystal cells (S2 File). Expression of PPP genes was generally unchanged during infection. Phosphofructokinase (Pfk) showed the lowest expression of all glycolytic genes and was further reduced in all hemocyte types during infection, suggesting a shift away from glycolysis to PPP and back to glycolysis at the glyceraldehyde-3P level. Overall, the expression analysis does not indicate specific changes in these metabolic pathways in either of the hemocyte types, or during infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Expression of glycolytic and pentose phosphate pathway enzymes in hemocytes.

Diagram showing metabolic pathways; metabolites and enzymes are abbreviated with Drosophila gene names; strongly expressed genes in hemocytes are in large font in bold, moderately expressed genes in large font, and insignificantly expressed genes in small font. Trehalose metabolism is shown in gray, glycolysis in black, pentose phosphate pathway in blue, purine metabolism in orange, and mitochondrial metabolism in green. Tables show glycolytic and PPP gene expression based on bulk RNAseq in TPM in circulating hemocytes at 18 hpi and from the uninfected (18 h Uninf) larvae; the shading highlights the strength of expression—white the weakest, red the strongest—data in S1 Table. hpi, hours post infection; PPP, pentose phosphate pathway; TPM, transcripts per million.

https://doi.org/10.1371/journal.pbio.3002299.g002

Whereas in our previous study, we detected slightly increased expression (1.5- to 2-fold) of most glycolytic genes after infection using qPCR analysis [4], here we detected decreased expression of most of these genes using bulk RNAseq analysis (Fig 2). This can be explained by increased overall gene expression in activated hemocytes after infection, where normalization of RNAseq data to one million transcripts (TPM) versus normalization to a selected housekeeping gene may lead to such differences. This difference is evident, for example, when comparing the change in trehalase expression, where RNAseq shows only an 11.5-fold increase (7.2 versus 83 TPM, Fig 1B), whereas qPCR shows a 25-fold increase (Fig 1E).

Activated hemocytes increase uptake and metabolism of glucose and trehalose via glycolysis and pentose phosphate pathway

In order to monitor changes in hemocyte metabolism during infection, we took 2 approaches based on tracing of metabolites labeled with the stable ¹³C isotope: in vivo and ex vivo. In the former, we fed D-glucose-¹³C₆ (fully labeled, i.e., on all 6 carbon atoms) to larvae at 16 hpi for 6 h before collecting hemocytes at 22 hpi. In addition to following metabolism in activated plasmatocytes, we assessed the last 6 h of differentiation of the first lamellocytes, which appeared complete by the end of the experiment. In the ex vivo approach, we incubated hemocytes in media containing either fully labeled D-glucose-¹³C₆ or α,α−trehalose-¹³C₁₂. Hemocytes collected by bleeding larvae 22 hpi were immediately incubated for 40 min with the labeled sugars. About half of the cells have already differentiated into lamellocytes. By examining only 1 time point, we are unable to determine metabolic fluxes, but we could still observe metabolic changes. It is important to note that in both approaches we worked with a heterogeneous population of hemocytes and thus observed either averaged metabolic changes that occurred across different types of hemocytes or a metabolism that occurred only in a specific hemocyte type.

After 6 h of feeding D-glucose-¹³C₆, 21% of the glucose in the larval hemolymph was labeled (S4 Fig). Some of the labeled glucose was converted to trehalose in the fat body, resulting in 16% of the circulating trehalose being labeled at one of the 2 glucose monomers (S4 Fig). Both glucose and trehalose in the hemolymph increased during infection and the increase was mainly due to the unlabeled sugars (S4 Fig), suggesting a release from stores consistent with our previous results [4]. Hemocytes can directly uptake labeled dietary glucose or convert labeled trehalose to glucose. Five-fold increased content of ¹³C-labeled and unlabeled glucose in infection-activated relative to control hemocytes indicated that hemocytes increased sugar uptake upon infection (S4 Fig), again in agreement with our previous results [4]. A direct comparison of downstream metabolite labeling between uninfected and infected larvae is complicated by the different fraction of labeled glucose in hemocytes of uninfected and infected larvae—15% versus 7%, respectively (S4 Fig), where the increased amount of labeled glucose in infected larvae is diluted by unlabeled glucose, also taken up by the hemocytes. Nevertheless, while the m+6 fraction (with all 6 carbons labeled) of intracellular glucose after infection is half that of the uninfected fraction, the m+6 fractions of glucose-6-phosphate (G6P) are comparable and in addition there are m+2 and m+3 fractions (i.e., with 2 or 3 labeled carbons) in infected samples that are not present in the uninfected samples (S4 Fig). Thus, the overall increase in labeled G6P indicates more intense glucose metabolism of hemocytes after infection.

Our ex vivo approach with comparable starting amount of labeled carbohydrates showed a significant decrease in the fraction of unlabeled and a significant increase in labeled glycolytic and PPP metabolites during infection, indicating increased metabolism in these pathways (Fig 3). Hemocytes of uninfected larvae metabolized glucose (Fig 3A, 3B, 3G, 3I and 3K) but little trehalose, which became metabolized substantially only during infection (e.g., Fig 3D, 3E, 3F, 3H, 3J and 3K). This is consistent with the fact that the cytoplasmic form of Treh is expressed only in lamellocytes. Therefore, most of the ¹³C incorporation from labeled trehalose can be attributed to lamellocyte metabolism. Fully labeled G6P-¹³C₆ (m+6 in Fig 3D) shows increased levels upon infection, but a prominent increase was also detected in a fraction of partially labeled G6P-¹³C₃ (m+3). Partially labeled G6P can arise from fully labeled glucose via cyclic PPP (Fig 3L), in which pentoses formed by oxidative PPP are converted back to G6P by the action of transketolase/transaldolase in the nonoxidative PPP and the reverse action of G6P isomerase (Figs 2, 3L, and S5; [23,24]). Since labeled pentoses represented a minor fraction (Fig 3B and 3E), they largely combined with unlabeled pentoses to form partially labeled G6P (S5 Fig). This m+3 fraction was particularly pronounced when labeled trehalose was used as the source (Fig 3D), demonstrating that cyclic PPP is active in lamellocytes. The increased m+3 fraction was also detected when labeled glucose was used, both with and without infection (Fig 3A), indicating a basal level of cyclic PPP activity in hemocytes. The m+3 fraction was also abundant in fructose-6P (Fig 3G and 3H), where it even exceeded the m+6 fraction (but note that fructose-6P could not be distinguished from glucose-1P). While m+6 fructose-6P most likely was a product of glycolysis, m+3 arose via cyclic PPP, suggesting stimulation of cyclic PPP under infection. The enhancement of cyclic PPP during infection is further supported by the in vivo experiment. Besides the expected m+6 fraction of G6P, we also detected m+2 and m+3 fractions during infection but not in uninfected larvae (S4 Fig). Before increasing amounts of labeled xylulose-5P and ribose-5P entered the transketolase/transaldolase conversion of pentoses to hexoses, G6P m+3 (pink in S5 Fig) was the main product, being formed by coupling labeled glyceraldehyde-3P with unlabeled sedoheptulose-7P. This was evident after a short ex vivo incubation (Fig 3A and 3D). Once the labeled pentoses entered the transketolase/transaldolase conversion to hexoses, G6P m+2 began to dominate (yellow in S5 Fig), as was observed after prolonged exposure to ¹³C in vivo (S4 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Analysis of hemocyte metabolism by stable ¹³C isotope tracing ex vivo.

(A–K) ¹³C labeling of glycolytic and pentose phosphate pathway metabolites in hemocytes from uninfected (light gray/blue bars) and infected (dark gray/blue bars) larvae. Hemocytes were incubated ex vivo for 40 min in media containing 0.5 mM D-glucose-¹³C₆ and 5 mM unlabeled trehalose (gray) or 5 mM α,α−trehalose-¹³C₁₂ and 0.5 mM unlabeled glucose (blue). m+0 indicates the fraction of the compound with unlabeled molecular mass; m+1 indicates the fraction of the compound with one ¹³C-labeled carbon, and so on. Bars represent means of 5 biological replicates with 95% CI, each dot represents 1 biological replicate; asterisks represent significant differences between uninfected and infected fractions tested by unpaired two-tailed Welch’s t test (numerical values in S1 Data). Note that ribulose-5P and ribose-5P (B, E) and fructose-6P and glucose-1P (G, H) were indistinguishable. For lactate (K), only the m+3 fraction is shown, as m+1 and m+2 were not detected above naturally occurring levels (m+0 and m+3 make up 100%). Schematic representation (L) of the cyclic PPP in blue and nonoxidative PPP-producing pentoses for de novo purine/pyrimidine synthesis in orange. PPP, pentose phosphate pathway.

https://doi.org/10.1371/journal.pbio.3002299.g003

Increased labeling of ribulose-5P/ribose-5P (Fig 3B and 3E) and sedoheptulose-7P (Fig 3C and 3F), the products of PPP and transketolase/transaldolase activity, respectively, further supports enhanced PPP in hemocytes during infection; we were unable to distinguish between ribulose-5P and ribose-5P (S2 Table) and therefore present them together. Ribulose-5P and ribose-5P can arise either from oxidative or nonoxidative PPP. The opposite direction of transketolase/transaldolase activity in the latter generates pentoses without NADPH (scheme in Figs 2, 3J, and S6). While m+3 and m+5 ribulose-5P and m+5 sedoheptulose-7P can be generated by oxidative PPP, m+2 ribulose-5P, and m+4 sedoheptulose-7P are rather markers of nonoxidative PPP (Figs 3B, 3C, 3E, 3F, S4, S5 and S6). The detected mixtures of these forms indicate that hemocytes use both oxidative/cyclic PPP and nonoxidative PPP. Different hemocyte types in our heterogenous cell population likely engage these 2 branches of PPP differently.

To summarize this part, hemocytes from both uninfected and infected larvae use nonoxidative PPP to generate pentoses as well as cyclic PPP to generate NADPH. Cyclic PPP increases in hemocytes during infection and particularly in lamellocytes as evidenced by trehalose metabolism.

The m+3 fraction of glyceraldehyde-3P, which is a product of both glycolysis and PPP, increased during infection from both labeled glucose and trehalose (Fig 3G and 3H). Pyruvate was more difficult to detect, and the labeled pyruvate reached detectable levels in the ex vivo experiment only after infection (S2 Table). We have previously shown that hemocytes release more lactate during infection [25], which is confirmed here by higher levels in hemolymph during infection (S4 Fig). The increased incorporation of ¹³C into lactate, from both glucose and trehalose (Fig 3I), demonstrates its increased production during infection.

Since the incorporation of ¹³C into tricarboxylic acid (TCA) cycle metabolites was very low (<1%) after 40 min of ex vivo incubation, we compared the incorporation into citrate from the in vivo experiment with feeding ¹³C-labeled glucose for 6 h. As we detected only the m+3 fraction of pyruvate (S7 Fig), we expected either m+2 (via pyruvate dehydrogenase) or m+3 (via pyruvate carboxylase) fractions of citrate. However, we detected an m+1 fraction as the highest (S7 Fig), which we also found in other TCA metabolites (S7 Fig). Therefore, we checked the ¹³C labeling of glutamine and glutamate as a potential source and found that it matched the fractional pattern of the TCA metabolites (S7 Fig). Since we fed larvae with labeled glucose for 6 h, some of the glucose was metabolized to amino acids; the labeled amino acids were released into the hemolymph and taken up by hemocytes as a source for mitochondrial metabolism or, via citrate, potentially for lipid synthesis. Pyruvate derived from glucose could still be used in mitochondria and contribute to the m+2/m+3 fractions, but the amino acids seem to be an important source for mitochondrial metabolism in hemocytes, both without and during infection. Nevertheless, changes in mitochondrial metabolism need to be further investigated.

Ribose-5P can be used for de novo purine/pyrimidine synthesis. This is indicated by the ¹³C labeling of AMP in vivo (S8 Fig). While all ribulose-5P/ribose-5P fractions in infected larvae were significantly lower (due to liberation of unlabeled endogenous glucose) than in uninfected larvae, the AMP fractions were similar, indicating increased purine production during infection. This is supported by the overall increase of AMP/ADP/ATP levels during infection (S8 Fig). Interestingly, the m+1 fraction of AMP was markedly higher compared to ribulose-5P/ribose-5P. In addition to m+1 ribose-5P as a source of m+1 AMP, labeled glycine (produced from labeled glucose similarly to glutamine; S8 Fig), added to unlabeled ribose-5P during purine synthesis, could increase the m+1 fraction of AMP. The carboxylation step utilizing labeled CO₂ released during oxidative PPP could be another source for the increased m+1 fraction of AMP [26].

In summary, metabolic changes occurring in hemocytes during infection include increased glucose uptake and, in lamellocytes, glucose production from trehalose. The prevailing fate of glucose during infection appears to be in cyclic PPP, where G6P is oxidized in multiple rounds, as evidenced by elevated partial G6P labeling. Increased glycolysis correlates with higher lactate release from hemocytes. A substantial portion of glucose is used for de novo purine synthesis. While the degree of glucose utilization in mitochondrial metabolism remains unclear, amino acids appear to be the source for the TCA cycle and potentially for lipid synthesis.

Oxidative PPP is required for proper immune response

Our metabolomics analysis shows that carbohydrate uptake and oxidative PPP increase in hemocytes upon infection. To test the importance of oxidative PPP during immune response, we used a double mutant in G6P dehydrogenase (Zwischenferment or Zw) and 6-phosphogluconate dehydrogenase (Pgd), the first and the third enzymes in oxidative PPP, respectively (Fig 4A). The Pgd[n39] pn[1] Zw[lo2a] mutant flies are fully viable [27]. We also used a hemocyte-specific RNAi of Zw (Fig 4A and 4B) using Srp-Gal4 and the Zw RNAi construct under the UAS promoter (P{TRiP.HMC03068}attP2). The number of plasmatocytes was altered neither in the double mutant nor by hemocyte-specific Zw knockdown, suggesting that basal hematopoiesis was unaffected by these manipulations (Fig 4C). However, both the double mutant and knockdown displayed decreased lamellocyte number (Fig 4C) and larval survival upon infection (Fig 4D), indicating that oxidative PPP is important for effective lamellocyte differentiation and parasitoid killing.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Oxidative pentose phosphate pathway is important for lamellocyte differentiation and resistance.

(A) Schematic representation of pentose phosphate pathway with metabolites in black and Drosophila genes encoding enzymes in blue. Genetic manipulations include hemocyte-specific RNAi of Zw induced by Srp-Gal4 (red lightning) and double null mutant in Zw and Pgd (Pgd[n39] pn[1] Zw[lo2a]; red crosses). (B) Efficiency of hemocyte-specific RNAi of Zw tested by RT-qPCR in hemocytes from uninfected (Uninf) and infected (INF) larvae 18 hpi. Bars represent mean values of fold change compared to uninfected control samples (Zw expression levels normalized by RpL32 expression in each sample); dots show individual samples, error bars represent ± SEM, the control (Srp> P{y[+t7.7] = CaryP}attP2) was compared to RNAi (Srp>P{TRiP.HMC03068}attP2) using unpaired t test, asterisks indicate p values (* P < 0.05, ** P < 0.01). (C) Number of hemocytes 22 hpi in control (Srp> P{y[+t7.7] = CaryP}attP2), Zw Pgd null mutant (Pgd[n39] pn[1] Zw[lo2a]), and hemocyte-specific RNAi of Zw. Each dot represents number of hemocytes in one larva, lines represent the mean, samples were compared by unpaired t test, asterisks indicate p value (* P < 0.05, **** P < 0.0001). (D) Survival of parasitoid wasp infection in control, null Zw Pgd mutant and hemocyte-specific RNAi of Zw. Bars show percentages of surviving flies (white), developing parasitoids (gray) and dead pupae when neither fly nor parasitoid survived (dark). Bars represent mean values; dots represent biological replicates, error bars represent ± SEM; survival rates were compared using ordinary one-way ANOVA with multiple comparisons; asterisks indicate p values (** P < 0.01, *** P < 0.001; numerical values in S1 Data). hpi, hours post infection.

https://doi.org/10.1371/journal.pbio.3002299.g004

Oxidative PPP generates NADPH, which is important for lipid synthesis. While lipid metabolism is beyond the scope of our study, we found a rather striking difference in lipid droplets between hemocytes from control larvae and larvae with hemocyte-specific RNAi of Zw during infection (S9 Fig). This indicates that properly functioning oxidative PPP in hemocytes is important for their lipid metabolism during infection.

Systemic trehalose metabolism and carbohydrates supply to hemocytes are required for efficient immune response

Metabolomics with ¹³C-labeled trehalose shows that trehalose is metabolized by infection-activated hemocytes. Therefore, we tested the importance of trehalose for the immune response. We first tested the hypomorphic mutation in trehalose-synthesizing gene Tps1 (Mi{y[+mDint2] = MIC}Tps1[MI03087] / Tps1[d2]). Tps1 mutant larvae develop with normal levels of glucose, glycogen, and triglycerides but their trehalose levels are only 20% those of controls [7]. Infected Tps1 mutants differentiated significantly fewer lamellocytes than control larvae (Fig 5A). Similarly, fewer lamellocytes differentiated in null Treh[cs1] mutants (Fig 5B) that cannot convert accumulating trehalose to glucose [22]. We verified that hemocytes from the Treh[cs1] mutant were unable to metabolize ¹³C-labeled trehalose supplied ex vivo (Fig 5C and S2 Table). These results demonstrate that trehalose is important for efficient lamellocyte differentiation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Effects of systemic carbohydrate metabolism and carbohydrate supply to hemocytes on lamellocyte differentiation.

(A) Number of lamellocytes 22 hpi in control (heterozygous Mi{y[+mDint2] = MIC}Tps1[MI03087] / +) and hypomorphic Tps1 mutant (Mi{y[+mDint2] = MIC}Tps1[MI03087] / Tps1[d2]). Each dot represents number of hemocytes per larva, lines represent mean, samples were compared by unpaired t test, asterisks indicate p value (**** P < 0.0001). (B) Number of lamellocytes 22 hpi in control (heterozygous Treh[cs1] / CyO Ubi-GFP) and null Treh[cs1] homozygotes. (C) Heat map of ¹³C-labeled fraction of metabolites from control and Treh[cs1] hemocytes in uninfected (Uninf) and infected (INF) conditions incubated ex vivo with α,α−trehalose-¹³C₁₂ (data in S2 Table). (D) Efficiency of hemocyte-specific RNAi of Tret1-1 tested by RT-qPCR in hemocytes from uninfected and infected larvae 18 hpi. Box plots (median, 75th and 25th percentiles) show fold change compared to uninfected Tret1-1-RA control samples (expression levels were normalized by RpL32 expression in each sample), each dot represents a biological replicate. Gray boxes represent control (Srp> P{y[+t7.7] = CaryP}attP2) and red boxes Tret1-1 RNAi (Srp>P{TRiP.HMS02573}attP2). Unpaired two-tailed t test was used to compare control with RNAi; asterisks indicate p values (* P < 0.05, ** P < 0.01, *** P < 0.001). (E) Fraction (%) of G6P-¹³C_n in hemocytes from uninfected and infected larvae, incubated ex vivo with labeled glucose (left bars) or trehalose (right bars), respectively. Gray bars represent mean in control (Srp> P{y[+t7.7] = CaryP}attP2) and red bars mean in hemocyte-specific Tret1-1 RNAi (Srp>P{TRiP.HMS02573}attP2). Unpaired two-tailed t test was used to compare control with RNAi; asterisks indicate p value (** P < 0.01). (F) Number of lamellocytes 22 hpi in control (Srp> P{y[+t7.7] = CaryP}attP2) and in hemocyte-specific Tret1-1 RNAi (Srp>P{TRiP.HMS02573}attP2). Each dot represents number of lamellocytes per larva, lines represent mean, samples were compared by unpaired t test, asterisks indicate p value (*** P < 0.001). (G) Number of lamellocytes 22 hpi in control (MFS3[CRISPR] Tret1-1[XCVI] / Cyo Ubi-GFP heterozygotes) and in MFS3 Tret1-1 double-null mutants (MFS3[CRISPR] Tret1-1[XCVI]). Each dot represents number of lamellocytes per larva, lines represent mean, samples were compared by unpaired t test, asterisks indicate p value (** P < 0.01). Numerical values for A, B, D, E, F, and G are available in S1 Data. hpi, hours post infection.

https://doi.org/10.1371/journal.pbio.3002299.g005

Next, we knocked down the Tret1-1 transporter specifically in hemocytes (Srp>Tret1-1-RNAi—P{TRiP.HMS02573}attP2), which resulted in a reduction of Tret1-1 expression upon infection to one-third compared to control (Fig 5D). Because the knockdown did not prevent the increase in Tret1-1 expression during infection (it only partially suppressed it), ex vivo metabolism of ¹³C-labeled trehalose, indicated by lower incorporation into G6P (Fig 5E and S2 Table) as well as the number of lamellocytes (Fig 5F) were only slightly reduced. Since Tret1-1 knockdown did not substantially suppress trehalose uptake, we used a null mutant of Tret1-1 together with a null mutation of MFS3. The double mutant MFS3[CRISPR] Tret1-1[XCVI] significantly reduced the number of lamellocytes, albeit again only slightly (Fig 5G), suggesting that other transporters ensure carbohydrate supply.

Cell-autonomous role of trehalose in hemocytes

Null Treh mutation impairs systemic trehalose metabolism. Since lamellocytes specifically increase the expression of the cytoplasmic Treh, we used the specific mutations in cTreh to test a cell-autonomous role of trehalose metabolism in hemocytes. The Treh[c1] mutation (S10 Fig) should block the utilization of trehalose inside the cells but not trehalose-glucose metabolism in the circulation [22]. However, when we incubated Treh[c1] mutant hemocytes with ¹³C-labeled trehalose ex vivo, we observed no effect on trehalose metabolism in hemocytes (S10 Fig and S2 Table). The failure of the Treh[c1] mutation to block intracellular trehalose metabolism could be due to a second start codon in the cTreh-RA/RD/RG/RE transcripts (S10 Fig). However, when both start codons were removed in our Treh[RAΔG4] mutant (Figs 1 and S10), the homozygous mutant hemocytes still metabolized labeled trehalose almost normally (S10 Fig and S2 Table), and the mutation did not affect the number of lamellocytes (S10 Fig). This is likely due to the compensatory expression occurring at the alternative transcription start sites common to both Treh-RB (cTreh) and Treh-RC (sTreh) transcripts (Fig 1).

Since we could not test the importance of trehalose metabolism in hemocytes with cTreh mutations, we decided to generate mitotic recombination clones [28] in the hematopoietic lineage with the Treh[cs1] mutation, which we verified as blocking trehalose metabolism (Fig 5C). We recombined flippase (Flp) target site FRT42D, Treh[cs1] mutation and RFP marker on the second chromosome and crossed this line to flies with FRT42D and the GFP marker to generate heterozygous FRT Treh[cs1] RFP / FRT GFP flies (Fig 6A). To induce mitotic recombination in the hematopoietic lineage of the progeny, the parental flies also carried Srp-Gal4 and UAS-Flp on their third chromosomes. Mitotic recombination resulted in RFP-labeled hemocytes with a Treh[cs1] null mutation and their GFP-labeled wild-type siblings (Fig 6A and 6B). When mitotic clones were first induced with only RFP and GFP markers without any mutation, the expected equal ratio of RFP to GFP sibling hemocytes was detected (Fig 6C), approximately 40% each, indicating high recombination efficacy (only the remaining 20% were non-recombined GFP/RFP hemocytes). A similar result was obtained in uninfected larvae with the Treh[cs1] mutation, although there was a minor increase in RFP-labeled Treh[cs1]^-/- hemocytes compared to the GFP-labeled wild-type hemocytes (49% versus 43%; Fig 6D). The difference is more likely to result from different genetic backgrounds of the chromosomal arms that become homozygous after recombination than from the lack of trehalase activity. A comparable result was obtained with lamellocytes from infected animals, where we found similar proportions of red and green cells—the small difference was consistent with the difference observed in uninfected larvae (Fig 6E). Generating clones with the Treh[cs1] null mutation did not change the total number of lamellocytes compared to larvae without induced recombination (Fig 6F). These results indicate that trehalose metabolism is dispensable for lamellocyte differentiation in a cell-autonomous manner, which is consistent with cytoplasmic trehalase not being expressed until in fully differentiated lamellocytes (Fig 1H).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Cell-autonomous role of trehalose metabolism in hemocytes.

(A) Generation of hemocyte clones with null Treh[cs1] mutation by Srp-Gal4 UAS-Flp induced mitotic recombination. Genotypes of parental and daughter cells after mitotic recombination are color marked red: RFP-marked Treh[cs1] mutant cells; green: GFP-marked wild-type sister clone; yellow: RFP/GFP nonrecombinant heterozygous cells. (B) Red-, green-, and yellow-marked hemocytes (both plasmatocytes and lamellocytes) upon mitotic recombination from infected larvae in circulation (left) and attached to a wasp egg (right). (C) Percentage of wild-type recombinant and nonrecombinant hemocytes in uninfected larvae without any mutation. (D) Percentage of Treh[cs1] homozygous, sister wild-type and nonrecombinant heterozygous hemocytes in uninfected larvae. (E) Percentage of Treh[cs1] homozygous, sister wild-type and nonrecombinant heterozygous lamellocytes in infected larvae. (C–E) Bars represent mean percentage, dots represent counts from single larvae, error bars represent ± SEM; samples were compared using unpaired two-tailed t test; asterisks indicate p value (* P < 0.05; ns = not significant). (F–I) Controls without recombination (gray) compared to larvae with Treh[cs1] mutant hemocyte clones (FRT42D GFP / FRT42D Treh[cs1] RFP; Srp-Gal4 / UAS-Flp; purple). (F) Number of lamellocytes 22 hpi. Each dot represents number of lamellocytes per larva, lines represent mean, samples were compared by unpaired t test, no significant difference. (G) Percentages of surviving adult flies, wasp larvae, dead pupae (neither fly nor wasp survived), and adult parasitoid wasps. Bars represent mean percentage, dots represent biological replicates (individual infection experiments), error bars represent ± SEM, samples were compared by ordinary one-way ANOVA with Šídák’s test for multiple comparisons, asterisks indicate p values (* P < 0.05; ** P < 0.01, ***P < 0.001; ns = not significant). (H) H₂O₂ levels in the hemolymph of infected control larvae (gray) and larvae with Treh[cs1] mutant clones (purple) at 30 hpi. Dots represent paired biological replicates (infection performed and H₂O₂ measured in parallel for 1 control and 1 mutant sample); data were compared using two-tailed paired t test, asterisks indicate P < 0.01. (I) The average number of progeny produced by an uninfected female and a female surviving the larval infection. Control (gray) and females bearing Treh[cs1] mutant clones (purple) were compared by two-tailed Welch’s t test; asterisk indicates P < 0.05. Bars represent the average numbers of progeny produced per female, each dot represents 1 infection experiment with at least 10 surviving females, error bars represent ± SEM. (J–M) Levels of unlabeled NADPH (J), GSH (K), GSSG (L), and taurine (M) in hemocytes from experiments ex vivo or in vivo. The bars show the mean metabolite amounts expressed by the normalized peak area, samples were obtained from hemocytes from uninfected (Uninf) or infected (INF) larvae, and were compared by two-tailed Welch’s t test; asterisks indicate p values (** P < 0.01, ***P < 0.001, ****P < 0.001; ns = not significant). (N) Model for the role of trehalose in producing antioxidants during response to parasitoid infection. Numerical values for C–M are available in S1 Data. hpi, hours post infection.

https://doi.org/10.1371/journal.pbio.3002299.g006

To test the effect on resistance, we used the larvae bearing approximately 40% of lamellocytes mutant for Treh. If trehalose metabolism in the lamellocytes is necessary for parasitoid killing, we would expect lower resistance. However, induction of Treh[cs1] mutant clones resulted in the opposite effect, with increased resistance compared to animals with the same genetic background in which clones were not induced (S11 Fig). Although a relatively large variability was observed, the average percentage of surviving adult flies in controls was 18% and never exceeded 30%, whereas in flies with induced Treh^-/- clones the average was 35% and reached up to 64% (Fig 6G). Subsequently, the percentage of surviving parasitoid wasps decreased from 65% in controls to 35% when Treh^-/- clones were induced (Fig 6G).

To kill the wasp egg, the larval immune cells generate ROS [29]. However ROS are potentially harmful also to the host, which prevents damage to its own tissues by producing antioxidants [13]. Flies use the thioredoxin system instead of glutathione reductase to reduce GSSG [30], and hemocytes highly express thioredoxin reductase Trxr1 and thioredoxin Trx2 with greater expression in lamellocytes (S12 Fig). Trx2 is also a substrate for peroxiredoxins [31], which are highly expressed in hemocytes as well (S12 Fig). The thioredoxin system requires NADPH (S12 Fig), which is produced by cyclic PPP. We found increased cyclic PPP activity and NADPH levels in hemocytes, including lamellocytes, during infection (Fig 6J). We also observed increased levels of the reduced glutathione form (GSH, Fig 6K) in hemocytes from infected larvae, whereas the oxidized GSSG form remained at comparable levels (Fig 6L). Besides GSH, infection elevated another antioxidant, taurine, in hemocytes (Fig 6M). Therefore, besides eliminating the pathogen, hemocytes seem to protect the host by mounting an antioxidant response.

Since trehalose metabolism in hemocytes did not appear to be important for resistance, we speculated that it may rather mediate the antioxidant host protection against own immune response. This would be consistent with trehalose being metabolized by cyclic PPP to produce NADPH, which is required for antioxidant production. Removing this activity from larval lamellocytes could decrease their ROS scavenging capacity and thus increase toxicity and resistance towards the pathogen. Therefore, we analyzed the effect of trehalase mutant hemocytes on the level of hydrogen peroxide in the hemolymph. The melanization response is generally associated with strong catalase activity in insect hemolymph, making it difficult to detect increased ROS during infection without catalase inhibitors [32]. A small increase in H₂O₂ was detected only at 30 hpi, but not at 20 or 40 hpi during the Drosophila response to parasitoid [33]. We did not use antioxidant inhibitors because we were interested in differences in antioxidant activities, and this is likely why we failed to detect a difference in hemolymph H₂O₂ between infected and uninfected larvae (S13 Fig). Induction of mitotic Treh[cs1] mutant clones did not alter H₂O₂ levels in uninfected larvae (S13 Fig). However, we detected a consistent, significant increase in H₂O₂ levels in the hemolymph of larvae with induced clones at 30 hpi (Fig 6H). This result supports the notion that trehalose may play a role in mechanisms that protect the host from oxidative stress. Down-regulation of trehalose metabolism in hemocytes leading to increased ROS may explain the observed increased resistance (Fig 6N).

There is a small, statistically insignificant increase in pupal lethality of infected larvae with induced Treh^-/- clones (Fig 6G), suggesting that the increased ROS levels were not lethal to the host. To explore further, we looked at the lifespan of adult flies that survived the infection. Male lifespan was comparable (S13 Fig); however, a greater number of females with induced clones died earlier than controls, with the median survival significantly reduced from 48 to 34 days, although the maximum lifespan was comparable (S13 Fig). A more pronounced effect was observed in the production of viable offspring: females with induced clones that survived infection produced less than half of progeny of controls, whereas no such difference was observed in the absence of infection (Fig 6I). Thus, induction of Treh[cs1] mutant clones in hemocytes during parasitoid infection of larvae results in reduced fitness of survivors. Whether there is a relationship between increased hemolymph H₂O₂ and the viability of future offspring of the infected larvae remains to be tested.

We attempted to verify the effect of Treh[cs1] hemocytes mutant clones via another trehalase manipulation. The Treh[cs1] null mutation affects systemic trehalose/glucose metabolism and homozygous cTreh-specific Treh[RAΔG4] loss is compensated by expression of alternative transcripts. Therefore, we used the heterozygous Treh[RAΔG4] allele that (1) partially reduces cTreh expression (S14 Fig); and (2) expresses Gal4 in the cTreh pattern, including a strong up-regulation in lamellocytes. To suppress Treh in the cTreh expression pattern, we combined the Gal4 driver within Treh[RAΔG4] with either of 2 UAS-Treh-RNAi constructs, Treh^HMC03381 or P{GD5118}v30731 (S14 Fig). Both RNAi lines recapitulated the cTreh null mutant phenotype (10% lethality in pupae and lethargic adults compared to 20% lethality in pupae and dying young adults in the null mutant). In both cases, larvae with Treh RNAi showed significantly increased resistance by killing parasitoids more often than control larvae (S14 Fig), reproducing the results obtained with Treh^-/- mutant clones. However, in the case of RNAi, the increased resistance did not result in improved survival, but was associated with increased host lethality (S14 Fig). The effect of RNAi might be more severe as, unlike in Treh^-/- hemocyte clones, Treh depletion was not limited to hemocytes but affected all tissues where cTreh is expressed and the Treh[RAΔG4] driver is active. Nevertheless, the increased pathogen killing at the expense of own survival suggests that trehalose may indeed play a role in host protection.

In summary, we used Treh^-/- mitotic clones to test for a cell-autonomous role of trehalose metabolism in hemocytes and found that it is not required for lamellocyte differentiation or host resistance. Elimination of the ability to metabolize trehalose in 40% of lamellocytes enhanced resistance, which was associated with increased hemolymph ROS levels and reduced fitness of survivors. A similar increased in resistance was achieved with 2 independent Treh RNAi lines, but this was associated with higher host lethality. Since hemocytes metabolize trehalose via cyclic PPP, generating NADPH and antioxidants, these results suggest that trehalose may play a role in protecting the host from ROS that serve to kill the pathogen.

Fly strains and cultivation

Drosophila melanogaster strain w¹¹¹⁸ (FBal0018186) in Canton S genetic background (FBst0064349) was used as a control line unless otherwise stated. Strains Pgdⁿ³⁹ pn¹ Zw^lo2a (FBst0006033), UAS-Zw-RNAi Zw^HMC03068 (FBal0292280), UAS-Tret1-1-RNAi Tret1-1^HMS02573 (FBal0281575), UAS-Treh-RNAi Treh^HMC03381 (FBal0292531) and control lines for RNAi y¹ v¹; P{CaryP}attP2 (FBst0036303), y¹ v¹; P{CaryP}Msp300attP40 (FBst0036304) and P{VALIUM20-EGFP.RNAi.1}attP40 (FBst0041555) were obtained from the Bloomington Drosophila Stock Center and UAS-Treh-RNAi P{GD5118}v30731 (FBst0458630) from Vienna Drosophila Resource Center. Strains Treh^c1 (FBal0321693), Treh^cs1 (FBal0321690), Tps1^MI03087 (FBal0260512) and Tps1^d2 (FBal0302039) were obtained from T. Nishimura, MFS3^CRISPR (FBal0366542) and Tret1-1^XCVI (FBal0319692) were obtained from S. Schirmeier. The MSNF9-GFP strain (P{msn-GFP.F9}; FBal0256873) was obtained from Viktor Honti. The SrpD-Gal4 strain (FBtp0020112) was obtained from M. Crozatier, backcrossed into the w¹¹¹⁸ background, and recombined with P{tubP-GAL80ts}2 (FBti0027797), which was also backcrossed into w¹¹¹⁸ background, to generate the w¹¹¹⁸; +/+; SrpD-Gal4 P{tubP-GAL80ts}2 line with Gal4 expression in all hemocytes but very low expression in the fat body at 25°C (expression in the fat body is only present at 29°C in this line). Line w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, Treh^cs1, P{w[+mC] = Ubi-mRFP.nls}2R / CyO; SrpD-Gal4 / TM6B was generated by recombination of P{ry[+t7.2] = neoFRT}42D P{w[+mC] = Ubi-mRFP.nls}2R (FBst0035496) with Treh^cs1 and by crossing to SrpD-Gal4. Line w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, P{w[+mC] = Ubi-GFP.nls}2R1 P{Ubi-GFP.nls}2R2 / CyO; P{y[+t7.7] w[+mC] = 20XUAS-FLPD5.PEST}attP2 / TM6B was generated by recombination of P{FRT(whs)}G13 P{Ubi-GFP.nls}2R1 P{Ubi-GFP.nls}2R2 (FBst0005826) with P{ry[+t7.2] = neoFRT}42D (FBti0141188) and by crossing to P{y[+t7.7] w[+mC] = 20XUAS-FLPD5.PEST}attP2 (FBti0161054). All flies were grown on cornmeal medium (8% cornmeal, 5% glucose, 4% yeast, 1% agar, 0.16% methylparaben) at 25°C.

Generation of Treh[RAΔG4] mutant

CRISPR-mediated mutagenesis was performed by WellGenetics using modified methods of Kondo and Ueda [45]. In brief, the gRNA sequence TGATTGCTCGATGGATTCGC[TGG] was cloned into U6 promoter plasmid. Cassette attP-Gal4-3xP3-RFP, which contains attP, Gal4, RBS, and a floxed 3xP3-RFP, and 2 homology arms were cloned into pUC57-Kan as donor template for repair. Treh-targeting gRNAs and hs-Cas9 were supplied in DNA plasmids, together with donor plasmid for microinjection into embryos of control strain w¹¹¹⁸. F1 flies carrying selection marker of 3xP3-RFP were further validated by genomic PCR and sequencing. CRISPR generates a 47-bp deletion allele of Treh and is replaced by cassette attP-Gal4-3xP3-RFP (Fig 1). The line is depicted here as Treh^RAΔG4. Treh^RAΔG4 was 10 times backcrossed to our control w¹¹¹⁸ genetic background.

Parasitoid wasp infection

Parasitoid wasps Leptopilina boulardi were reared on sugar agar medium (6% sucrose, 1.5% agar, 0.75% methylparaben) and grown by infection of wild-type Drosophila larvae. Early third instar larvae (72 h after egg laying) were infected with parasitoid wasps (= time point 0 hours). Weak infection (1 to 2 eggs per larva) was used for resistance and survival analysis. Strong infection (4 to 8 eggs per larva) was used for the rest of the experiments to obtain a strong and more uniform immune response. Infections were performed on 60-mm Petri dishes with standard cornmeal medium for 15 min with periodic interruption of infecting wasps for weak infection and 45 min for strong infection.

Hemocyte counting

Hemocytes were obtained from larvae by cuticle tearing of 1 larva in 15 μl PBS and counted based on morphology in Neubauer hemocytometer (Brand GMBH) using differential interference contrast microscopy.

Resistance, survival, and fitness analysis

To determine survival and parasitoid resistance rates, infected/control larvae were placed in fresh vials (typical 1 experiment = 30 larvae/vial and 3 vials/genotype in 3 independent biological replicates of infection). To determine resistance, pupae were dissected 4 days after infection to count melanized wasp eggs (winning host kills pathogen) or surviving wasp larvae (winning parasitoid). For the survival experiment, emerged adult flies were counted as survivors of infection excluding flies without melanized capsule (if no melanized capsule was visible in the abdomen, the fly was dissected). Adult wasps that emerged from pupae were counted as adult winning parasitoids. The lifespan of surviving infected flies was determined by transferring flies to a fresh vial (10 female flies per vial) every 2 to 3 days and counting the days until death. Fitness was determined by leaving at least 5 wild-type uninfected males with at least 5 females that survived the infection (maximum 10 females per vial) in a vial, transferring them to a fresh vial every 2 to 3 days, and counting the number of offspring pupae throughout their lifetime.

Gene expression analysis

Hemolymph was collected by rupturing 50 larvae in 2 μl of PBS on a microscope slide on ice, the hemolymph was transferred into 200 μl of Trizol reagent (Ambion), homogenized using a plastic motorized pestle, incubated for 15 min at room temperature and either frozen at −80°C or directly followed by RNA isolation using a Direct-zol RNA microprep kit (Zymo Research) according to the manufacturer’s protocol. Reverse transcription was performed using PrimeScript reverse transcriptase (Takara) and gene expression was analyzed using the TP SYBR 2x mastermix (TopBio) with the primers listed in Table 1 on a CFX 1000 Touch real time cycler (BioRad). Expression of a specific gene in each sample was normalized to expression of RpL32 (FBgn0002626).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Primers.

https://doi.org/10.1371/journal.pbio.3002299.t001

Bulk RNAseq analysis

RNA was extracted from circulating hemocytes (72 h after egg laying = time of infection = 0 h, 81 h after egg laying = 9 h post infection/hpi and 90 h after egg laying = 18 hpi), from lymph glands (9 and 18 hpi) and from wing discs (9 hpi) of uninfected and infected third instar w¹¹¹⁸ larvae. Circulating hemocytes were obtained by ripping 100 larvae in ice-cold PBS directly into 1.5 ml centrifuge Eppendorf tubes, centrifuging 5 min at 360xg, removing the supernatant, and isolating RNA using Trizol reagent (200 μl) (Ambion) according to the manufacturer’s protocol. Lymph glands and wing discs were dissected from larvae in ice-cold PBS, transferred to 1.5 ml centrifuge Eppendorf tubes with Trizol (200 μl) reagent (Ambion), homogenized using a plastic motorized pestle, followed by Direct-zol RNA microprep kit (Zymo Research) according to protocol. Frozen total RNA samples were sent to the Genomics Core Facility (EMBL Heidelberg) for preparation of barcoded 3′-end seq forward libraries, followed by deep uni-directional sequencing of 75-base long reads using Illumina NextSeq. Raw data are available at The European Nucleotide Archive under study accession number: PRJEB74490 (secondary acc: ERP159178) (https://www.ebi.ac.uk/ena/browser/view/PRJEB74490). Trimmed reads in Fastq files were mapped to the BDGP Drosophila melanogaster Release 6.29 genomic sequence using the Mapper for RNA Seq in Geneious prime software (Biomatters). Normalized counts of reads mapped to each gene annotation were calculated as TPM, expression levels were compared using the DESeq2 method in Geneious prime software, and data were exported to an Excel file (S1 Table).

Metabolomics and stable ¹³C isotope tracing

¹³C-labeled glucose feeding. ¹³C-labeled glucose (D-Glucose-¹³C₆ isotope purity ≥99 atom % ¹³C - Sigma-Aldrich) was added to the semi-defined diet (per 100 ml: 0.62 g of agar, 8 g of brewer’s yeast, 2 g of yeast extract, 2 g of peptone, 3 g of sucrose, 3 g of unlabeled glucose, 0.05 g of MgSO₄ × 6H₂O, 0.05 g of CaCl₂ × 2H₂O, 600 μl of propionic acid, 1 ml of 10% p-hydroxy-benzoic acid methyl ester in 95% ethanol). The diet mixture was brought to a boil, then cooled to 50 to 60°C with stirring and p-hydroxy-benzoic acid and propionic acid were added, and 2 ml of medium was taken into a Falcon tube and 100 μl of ¹³C-labeled glucose (600 mg/ml) was added (50% of the glucose in the diet was labeled). Approximately 1 ml of diet was poured into a glass vial, and 100 uninfected or infected larvae (16 h after the start of infection) were placed in the vial for 6 h. The larvae were then removed from the vial, washed twice in water and once in PBS and placed on a microscope slide covered with parafilm, and the PBS residue was removed with filter paper, and 4 μl of PBS was added to the larvae and each larva was ruptured, 20 μl of hemolymph was collected and transferred to sterile 1.5 ml Eppendorf polypropylene centrifuge tubes with 60 μl of PBS. Larvae were washed with an additional 25 μl of fresh PBS and 20 μl was recovered in the same tubes. Samples were centrifuged for 5 min at 360xg, and 95 μl of supernatant (with extracellular metabolites) was removed, 380 μl of cold acetonitrile-methanol (1:1) extraction buffer with addition of the internal standard 4-fluorophenylalanine (50 ng/sample) was added to control the extraction efficiency and mass spectrometer responses and samples were placed in liquid nitrogen. To collect metabolites from pelleted hemocytes, 50 μl of water was added, frozen in liquid nitrogen/thawed at 37°C 3 times (to disrupt the rigid cells of hemocytes), then 200 μl of cold acetonitrile-methanol (1:1) with addition of the internal standard 4-fluorophenylalanine (50 ng/sample) was added and stored at −80°C until LC-HRMS analysis.

Ex vivo hemocyte incubation with ¹³C-labeled glucose and trehalose. Larvae were washed first with distilled water and then with PBS to reduce contamination. Larval hemolymph was collected by carefully tearing the larvae on a glass microscope slide covered with parafilm. Hemolymph from 50 larvae was immediately collected into sterile 1.5 ml Eppendorf polypropylene centrifuge tubes prefilled with 100 μl PBS and centrifuged for 5 min at 25°C, 360xg. The supernatant was then removed and the cells were mixed with labeled medium (5 mM ¹³C₁₂ labeled or unlabeled trehalose, 0.5 mM ¹³C₆ labeled or unlabeled glucose, 5 mM proline, 0.3 mM methionine, and 5 mM glutamine, all reagents from Sigma/Merck) supplemented with gentamicin (10 mg/ml; Gibco), amphotericin B (250 μg/ml; Gibco), and 0.1 mM phenylthiourea (PTU; Sigma/Merck) to prevent melanization. The hemocytes were then incubated for 40 min at 25°C and 80% to 90% humidity. The cells were then centrifuged for 5 min at 25°C, 360xg, the supernatant was removed, the cells were mixed with 50 μl of cold PBS and frozen in liquid nitrogen/thawed at 37°C 3 times (to disrupt the rigid hemocyte cells). Finally, 200 μl of cold acetonitrile-methanol (1:1) was added, in addition, the internal standard 4-fluorophenylalanine (50 ng/sample) was added to control the extraction efficiency and mass spectrometer responses and samples were stored at −80°C until LC-HRMS analysis.

Frozen samples were melted on ice, then samples were homogenized using a TissueLyser LT (Qiagen, Hilden, Germany) set to 50 Hz for 5 min (with a rotor pre-chilled to −20°C). Homogenization and centrifugation (at 20,000×g for 5 min at 4°C) was repeated twice and the 2 supernatants were combined. Post-centrifugation, the supernatant was collected and filtered using a 0.2 μm PVDF mini-spin filter (sourced from HPST, Prague, Czech Republic). This filtration process was carried out at 7,000 RPM and a temperature of 5°C for 5 min. The filtered supernatant was then completely dried using a refrigerated vacuum concentrator (Jouan RC 10.10 and RCT 60, Saint Herblain, France). The dried aliquots were reconstituted with 50% acetonitrile (50 μl) and thoroughly mixed for 30 s before being placed in the ultrasonic bath for an additional 5 min. The prepared sample was then measured using LC-HRMS.

LC/HRMS condition and data mining

In this study, the flux metabolomic analysis utilized the high-resolution mass spectrometer Q Exactive Plus combined with Dionex Ultimate 3000 and Dionex open autosampler from Thermo Fisher Scientific, San Jose, California, United States of America. The mass spectrometer was operated in negative (NESI) ion mode. Full MS scan mode was employed for detection, with a mass range of 70 to 1,050 Da. The Q-Exactive settings included a 70,000 resolving power (scan rate at ±3Hz), 3 × 10⁶ automatic gain control (AGC) target, and a maximum ion injection time (IT) of 100 ms. The ionization parameters were set as follows: (+/−) 3,000 kV spray voltage, 350°C capillary temperature, sheath gas at 60 arbitrary units (au), aux gas at 20 au, spare gas at 1 au, probe temperature 350°C, and S-Lens level at 60 au. For accurate mass, lock masses of 301.9981 were utilized.

Chromatographic separation occurred on the SeQuant ZIC-pHILIC column (150 mm × 4.6 mm i.d., 5 μm, Merck KGaA, Darmstadt, Germany) with a flow rate of 450 μl/min, an injection volume of 5 μl, a column temperature of 35°C, and a mobile phase gradient of 20% B at 0 min, 80% B at 20 min, 95% B at 20.1 min, holding at 95% B until 23.3 min, returning to 20% B at 23.4 min, and maintaining 20% B until 30.0 min. Here, A represented acetonitrile and B represented 20 mmol/L aqueous ammonium carbonate (pH = 9.2; adjusted by NH₄OH). The data were acquired and processed using homemade software MetaboliteMapper and Xcalibur software version 2.1 from Thermo Fisher Scientific, San Jose, California, USA. Raw data are available at figshare (https://doi.org/10.6084/m9.figshare.25525657.v1) [46]. All reported metabolites were identified using standards, except FGAR, UMP, UDP-galactose, and 2PG, which were identified using online sources based on accurate mass and MS spectra. All satellites M+1, M+2, and M+3 was compensated for their natural representation. Full details on the identification and compensation of the natural occurrence can be found in S2 Table. For peak area analysis, the data were normalized to the total content of all screened unlabeled metabolites—the peak area of the metabolite in a particular sample was divided by the peak area of the same metabolite of the selected reference sample and this procedure was repeated for each individual unlabeled metabolite. These ratios of all metabolites in one particular sample were averaged to determine a normalization factor. We then divided the measured peak area by the normalization factor for that sample to obtain the normalized peak area values (normalization factors and normalized values are reported in S2 Table).

Generation of hemocyte mutant clones by mitotic recombination

The FRT RFP; SrpD-Gal4 control line (w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, P{w[+mC] = Ubi-mRFP.nls}2R / CyO; SrpD-Gal4 / TM6B) or the FRT Treh^cs1 RFP; SrpD-Gal4 mutant line (w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, Treh^cs1, P{w[+mC] = Ubi-mRFP.nls}2R / CyO; SrpD-Gal4 / TM6B) were crossed with either flippase-free FRT GFP line (w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, P{w[+mC] = Ubi-GFP.nls}2R1 P{Ubi-GFP.nls}2R2 / CyO) as control without clone induction or with FRT GFP; UAS-Flp line (w¹¹¹⁸; P{ry[+t7.2] = neoFRT}42D, P{w[+mC] = Ubi-GFP.nls}2R1 P{Ubi-GFP.nls}2R2 / CyO; P{y[+t7.7] w[+mC] = 20XUAS-FLPD5.PEST}attP2 / TM6B) to induce mitotic clonal recombination in hemocytes by expressing flippase using SrpD-Gal4 driver. Larvae with ubiquitous red and green fluorescence, i.e., without balancers, were selected and dissected in PBS on a microscope slide to obtain hemocytes. Images of hemocytes were taken using red and green fluorescence and differential interference contrast microscopy. Merged images were used to count green, red, and heterozygous yellow hemocytes.

Hemolymph hydrogen peroxide measurement

Early third instar larvae were selected and transferred to fresh medium in a Petri dish and were infected with parasitoid wasps. Larvae were collected from the Petri dish at 30 hpi, washed twice in water and once in PBS, placed on a microscope slide covered with parafilm, and the PBS residue was removed with filter paper. Hemolymph was obtained from the larvae by tearing the cuticle from 50 larvae and the hemolymph was collected in a sterile Eppendorf tube. The sample volume was adjusted to 50 μl with PBS. Pairs of samples were processed at the same time and the amount of H₂O₂ was immediately assessed using a colorimetric Hydrogen Peroxide Assay Kit (Abcam, Ab102500) according to the original protocol. The samples were not deproteinized. A fresh set of standards was used for each measurement.

Immunohistochemistry

The central nervous system of infected and noninfected third instar larvae were dissected and stained according to standard protocols. The following primary antibodies were used: GFP anti-chicken (Abcam, 1:1,000), Rabbit anti-Rumpel (1:500; [47]), Elav anti-rat and Repo anti-mouse (Developmental Studies Hybridoma Bank, 1:5), Tret1-1 anti-guinea pig [1:50; [48]). All secondary antibodies conjugated to Alexa Fluor 488, Alexa Fluor 568, or Alexa Fluor 647 were used at a ratio of 1:1,000 (Thermo Fisher Scientific). Confocal images were obtained using a Zeiss LSM 880 (Zeiss, Oberkochen, Germany) and analyzed using Fiji [49].

Lipid droplet staining

Three third instar larvae were cleaned in PBS and placed in a 5 μl drop of PBS on a coverslip. Forceps were used to rip the cuticle, allowing hemolymph to flow out. Carcasses were discarded and hemolymph and left for 10 min to adhere to the coverslip. Cells were washed 3 × 5 min in 1× PBS and fixed in 4% paraformaldehyde (PFA) for 15 min. PFA was removed and cells were washed 3 × 5 min in 1× PBS then incubated with BODOPY 493/503 for 1 h at RT. Cells underwent a final washing step of 3 × 5 min in 1× PBS. A drop of Vectashield (vector) was added to the coverslip before being placed on a microscope slide. Edges were sealed with nail polish and images were captured on a FV3000 Olympus Confocal Microscope, using a 100× objective and 2× digital zoom.

Data analysis

Data were analyzed and graphed using GraphPad Prism (GraphPad Software), with specific statistical tests shown in the legend of each figure.