Symbiont-induced odorant binding proteins mediate insect host hematopoiesis

Abstract

Symbiotic bacteria assist in maintaining homeostasis of the animal immune system. However, the molecular mechanisms that underlie symbiont-mediated host immunity are largely unknown. Tsetse flies (Glossina spp.) house maternally transmitted symbionts that regulate the development and function of their host's immune system. Herein we demonstrate that the obligate mutualist, Wigglesworthia, up-regulates expression of odorant binding protein six in the gut of intrauterine tsetse larvae. This process is necessary and sufficient to induce systemic expression of the hematopoietic RUNX transcription factor lozenge and the subsequent production of crystal cells, which actuate the melanotic immune response in adult tsetse. Larval Drosophila's indigenous microbiota, which is acquired from the environment, regulates an orthologous hematopoietic pathway in their host. These findings provide insight into the molecular mechanisms that underlie enteric symbiont-stimulated systemic immune system development, and indicate that these processes are evolutionarily conserved despite the divergent nature of host-symbiont interactions in these model systems.

Keywords: D. melanogaster; crystal cell; developmental biology; hematopoiesis; immunology; odorant binding protein; stem cells; symbiont; tsetse fly.

Figure 1.. Symbiont-mediated differential expression of odorant binding protein 6 in tsetse larvae.

(A) Number of genes exhibiting significant differential expression, and a relative transcript abundance [in transcripts per million (TPM)] over 3, in Gmm^WT compared to Gmm^Apo larvae. Significance is based on a Baggerly’s test followed by a false detection rate correction (p<0.01). (B) Significantly enriched gene ontology categories, determined using a Fisher’s exact test. (C) Genes exhibiting significant differential expression (measure as fold-change in gene expression) in Gmm^WT compared to Gmm^Apo larvae, and a minimum TPM value of 1000. Significance was determined as in (A). (D) Enrichment analysis of odorant binding protein-encoding genes expressed in Gmm^WT adult males (purple) and females (yellow), and Gmm^WT larvae (green). (E) Relative expression (TPM) of tsetse odorant binding protein-encoding genes in Gmm^WT larvae, and their differential expression (measure as fold-change in gene expression) in Gmm^WT compared to Gmm^Apo larvae. Significance was determined as in (A). (F) Relative obp6 expression in Gmm^WT larvae, as well as larvae derived from symbiont-cured moms fed a diet supplemented with yeast and Wigglesworthia-containing bacteriome extracts (Gmm^bact/Wgm+), Wigglesworthia-free bacteriome extracts (Gmm^bact/Wgm-), Sodalis cell extracts (Gmm^Sgm+), and bacteriome extracts harvested from Gmm^Apo females (Gmm^bact/Apo). Gmm^WT and Gmm^Apo flies served as controls. n = 6 biological replicates for groups Gmm^WT, Gmm^bact/Wgm+ and Gmm^Sgm+ samples, and n = 5 biological replicates for Gmm^bact/Wgm-, Gmm^bact/Apo and Gmm^Apo samples. Replicates for all groups contain a mixture of four first and second instar larvae. Data are presented as mean ± SEM. Bars with different letters indicate a statistically significant difference (specific p values are listed in Figure 1—source data 1) between samples. Statistical analysis = ANOVA followed by Tukey’s HSD post-hoc analysis. DOI: http://dx.doi.org/10.7554/eLife.19535.003

Figure 1—figure supplement 1.. Developmental stage-specific enrichment analysis of tsetse orthologues that putatively cluster within the ‘hematopoiesis’ COG.

Ontology enrichment analyses were performed as described in the Materials and methods, under the ‘Transcriptomics’ sub-heading. Refer to Supplementary file 2 for a description of the enriched hematopoiesis-associated gene expressed specifically in Gmm^WT and Gmm^Apo tsetse larvae. DOI: http://dx.doi.org/10.7554/eLife.19535.005

Figure 2.. Tsetse odorant binding protein 6 does not mediate the development and function of phagocytic hemocytes.

(A) Survival following systemic challenge of siOBP6 and siGFP adults with 5 × 10² CFU of E. coli K12. Fly survival was monitored every other day for the duration of the 14 day experimental period. Survival assays were performed in triplicate, using 25 flies per replicate. Red curve depicts a statistically significant difference in infection outcome (p<0.0001, log-rank test). (B) Hemocyte abundance in siOBP6 and siGFP adults was quantified microscopically using a hemocytometer (Figure 2—source data 1). (C) A representative micrograph of hemocyte-engulfed recE. coli_GFP from siOBP6, siGFP and siOBP6^R adults. Experiment was performed using hemolymph collected from four distinct flies per (Figure 2—source data 2). Hemolymph was collected 12 hpc and fixed on glass slides using 2% paraformaldehyde. Magnification is x400. (D) E. coli densities (CFU/μl of hemolymph) in the hemolymph of siOBP6, siGFP and siOBP6^R adults at 2 and 6 dpc (Figure 2—source data 3). In (B) and (D), symbols represent one hemolymph sample per group, and bars represent the median hemocyte quantity (B) or bacterial density (D) per sample. Statistical analysis = ANOVA followed by Tukey’s HSD post-hoc analysis. DOI: http://dx.doi.org/10.7554/eLife.19535.006

Figure 3.. Obp6 mediates the melanization cascade in adult tsetse.

(A) Survival following administration of clean wounds to the thoracic cuticle of siOBP6, siGFP and siOBP6^R adults. Survival assays were performed in triplicate, using 25 flies per replicate. Red curve depicts a statistically significant difference in infection outcome (p<0.0001, log-rank test). (B) A representative micrograph of the cuticle of siRNA treated adults 3 hr post-wounding (hpw) with a clean needle. Melanin deposited at the wound site of siGFP and siOBP6^R controls, and hemolymph exudate from a siOBP6 treatment individual, are identified by black and red arrowheads, respectively. Scale bar = 500 μm. Experiment was performed using four distinct flies per group (Figure 3—source data 1). (C) Quantitation of PPO1 and PPO2 in the hemolymph of siOBP6, siGFP and siOBP6^R adults three hpw with a clean needle. Shown is a representative Western blot analysis using Drosophila anti-PPO1 and anti-PPO2 antibodies. 8 μl of pooled hemolymph was run per gel lane. Hemolymph was collected and pooled from four individuals from each group. Western blots were repeated in triplicate [Figure 3—source data 2 (for PPO1 westerns) and Figure 3—source data 3 (for PPO2 westerns)]. (D) PO activity in the hemolymph of siOBP6, siGFP and siOBP6^R adults at 0 and 3 hpw with a clean needle. n = 5 biological replicates per group per time point for pre-wound readings, and n = 8 biological replicates per group per time point for post-wound readings. Data are presented as mean ± SEM. Bars with different letters indicate a statistically significant difference between pre- and post-wound values (specific p values are listed in the Figure 3—source data 4). Statistical test = 2 way ANOVA followed by Tukey’s HSD post-hoc analysis. DOI: http://dx.doi.org/10.7554/eLife.19535.010

Figure 4.. Obp6 expression in the gut of larval tsetse is an integral component of the systemic pathway that actuates crystal cell production.

(A) Representative micrograph depicting spontaneous PPO activation in early third instar siGFP, siOBP6 and siOBP6^R tsetse larvae following subjection to a 10 min heat shock at 65°C. Experiment was repeated using one larvae from five distinct moms from each group. Melanotic spots were quantitated microscopically. Statistical analysis = Kruskal-Wallis test followed by Dunn’s post-hoc analysis (Figure 4—source data 1). (B) RT-qPCR analysis of obp6, serpent and lozenge expression in embryos prior to maternal treatment with siRNA, and in siOBP6, siGFP and siOBP6^R tsetse larvae from siRNA treated moms. Embryo replicates (n = 5) contain three embryos, larval replicates (n = 7 for siOBP6, n = 5 for siGFP and n = 6 for siOBP6^R) contain a mixture of four first and second instar larvae. ND, not detectable. Data are presented as mean ± SEM. Bars with different letters indicate a statistically significant difference between samples (specific p values for larval samples are listed in the Figure 4—source data 2). Statistical analysis = 2 way ANOVA followed by Tukey’s HSD post-hoc analysis. (C) Representative image of obp6 and lozenge spatial expression patterns, determined using semi-quantitative RT-PCR, in the gut and carcass of second instar Gmm^WT larvae. Experiment was repeated using guts and carcasses from five distinct individuals (Figure 4—source data 3). DOI: http://dx.doi.org/10.7554/eLife.19535.015

Figure 5.. Drosophila’s indigenous microbiota actuates larval hematopoietic pathways and thus functionality of the adult melanization response.

(A) Relative expression of obp28a and lozenge in conventionally reared (CR) and axenic (AX) Oregon-R and w¹¹¹⁸ Drosophila larvae. n = 9 (Oregon-R) and 6 (w¹¹¹⁸) biological replicates per group, each containing a mixture of thirty second and early-3^rd instar larvae. Data are presented as mean ± SEM. Asterisks indicate statistical significance (specific p values are listed in the Figure 5—source data 1). Statistical analysis = unpaired t-tests, corrected for multiple comparisons using the Holm-Sidak method. (B) AX larvae house fewer sessile crystal cells, and produce less PPO, than do and CR individuals. Top panels are representative micrographs depicting spontaneous PPO activation in AX and CR w¹¹¹⁸ and Oregon-R larvae following subjection to a 10 min heat shock at 65°C. n = 5 larvae per group [Figure 5—source data 2 (for Oregon-R larvae) and Figure 5—source data 3 (for w¹¹¹⁸ larvae)]. Bottom panels are representative Western blots using Drosophila anti-PPO1 antibodies. 8 μl of pooled hemolymph was run per gel lane. Hemolymph was collected and pooled from 15 individuals from each group. Western blots were repeated in triplicate (Figure 5—source data 4). (C) Survival of obp28a RNAi knockdown Drosophila (tub-GAL4/UAS-obp28a RNAi) and knockout (Obp28a^-) adults compared to controls (tub-GAL4, UAS-obp28a RNAi and wCs) following wounding with a clean needle. Experiment was performed in triplicate, n = 50 (RNAi) and n = 45 (knockout) per group per replicate. Red curve depicts a statistically significant difference in infection outcomes [p<0.0001 (RNAi) and p=0.0002 (knockout), log-rank test]. (D) A representative micrograph of the cuticle of obp28a knockdown, knockout and control Drosophila adults six hpw with a clean needle. Experiment was performed using two distinct experimental fly cohorts [n = 4 flies per group per experiment; Figure 5—source data 5 (for RNAi flies) and Figure 5—source data 6 (for deletion mutants)]. Wounds on the cuticle of control (melanized) and obp28a knockdown and knockout individuals (not melanized) are identified with black and red ovals, respectively. DOI: http://dx.doi.org/10.7554/eLife.19535.019

Figure 5—figure supplement 1.. Axenic and conventionally reared w¹¹¹⁸ and Oregon-R larvae following subjection to a 10 min heat shock at 65°C.

White ellipses demarcate an area of the larval cuticle that contains a high density of ruptured crystal cells. DOI: http://dx.doi.org/10.7554/eLife.19535.026

Figure 6.. Model illustrating the functional relationship between maternally-transmitted enteric symbionts and melanization in tsetse.

Gmm^WT larvae imbibe enteric symbiotic-containing milk gland secretions throughout their intrauterine developmental program. These bacteria colonize larval gut-associated tissues, including the bacteriome, and in doing so, induce the expression of obp6. OBP6 is either secreted directly into the hemolymph, or acts locally to induce expression of another unknown, (also secreted) protein. One of these molecules then acts systemically in the larval hematopoietic niche (hn) to stimulate lozenge (lz) expression in a small proportion of serpent (srp) expressing prohemocytes. These cells then become PPO-producing crystal cells [remaining prohemocytes become phagocytes after expressing glial cells missing (gcm)]. Finally, crystal cells are expelled from the hn, where they circulate in the hemolymph and are available to produce wound-healing melanin. Larvae that develop in the absence of symbiotic bacteria (Gmm^Apo) fail to produce any hemocytes, while those that develop in the presence of reduced obp6 transcript abundance (Gmm^OBP6-) fail to express lozenge and thus likely fail to generate crystal cells. dv, dorsal vessel; hc, hemocoel; w, wound; ep, epithelial cells of midgut; bc, bacteriome; pm, peritrophic matrix; gl, gut lumen. DOI: http://dx.doi.org/10.7554/eLife.19535.027