Tsetse gut response to infection with symbiotic Spiroplasma or parasitic trypanosomes

To assess the impact of Spiroplasma or trypanosome infection outcomes on midgut functions, we performed transcriptional comparisons between the Spi⁺ or Tbb⁺ groups relative to the uninfected Ctrl group. Infection with Spiroplasma resulted in a total of 69 differentially expressed (DE) genes, of which 49 were induced and 20 were suppressed (Fig 1A; S2 Table Sheet 3). In contrast, infection with Trypanosoma resulted in a total of 989 DE genes, of which 518 were induced and 471 suppressed (Fig 1A; S2 Table Sheet 4). Clustering based on the Euclidean distances between Ctrl and Tbb⁺ samples resulted in a separate cluster group while no cluster groups were observed between Ctrl and Spi⁺ samples (S1B Fig). These results indicate that infections with trypanosomes induced a more robust response in the gut in comparison to that elicited in the presence of Spiroplasma infections.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Tsetse Gut Tissue Responses to Spiroplasma and Trypanosome Infection.

A. Differentially expressed (DE) genes identified in the Spi⁺ and Tbb⁺ groups relative to uninfected Ctrl are presented as follows: The orange and red areas represent up-regulated genes in the Spi⁺ and Tbb⁺ state respectively, while light blue and dark blue indicate down-regulated genes in Spi⁺ and Tbb+ respectively, at a significance level of 10% adjusted p-value and log₂ fold change (log₂FC) ≥ 1. B. Venn diagram illustrates the number of DE genes that are unique to or shared between the Spi⁺ and Tbb⁺ states relative to uninfected controls (Ctrl). C. Heat maps display the DE genes with putative PM-associated and immunity-related functions that are shared between the Spi⁺ and Tbb⁺ datasets. Fold-change values are represented as a fraction of the average normalized gene expression levels from age-matched Spi⁺ or Tbb⁺ versus uninfected control flies (Ctrl). The heat maps (dendrograms) were clustered using Euclidean distance calculation methods in R-package software. The clusters were manually separated into various categories.

https://doi.org/10.1371/journal.ppat.1012692.g001

We next evaluated the DE genes from the Spi⁺ or Tbb⁺ groups to understand the functional impact of each microbe infection on host physiology. We began by investigating the potential functions of the 38 genes (Fig 1B), of which 33 were upregulated and five were downregulated in both Spi⁺ and Tbb⁺ groups compared to control (Ctrl) group. Of note, among the induced genes, eight encoded proteins associated with the function of the peritrophic matrix (PM) in tsetse flies. These included four peritrophins, chitin synthase 2 (chs2), a chitin binding protein (CG34282), chitinase 2, and chitin deacetylase 8 (Figs 1C and S2A). Tsetse fly PM is composed of a chitinous matrix embedded with glycoproteins, serving as a barrier that separates the gut lumen from the epithelia. This structure protects the gut cells from harmful compounds present in the bloodmeal as well as from ingested pathogens [53]. The eight induced putative products noted above have previously been described with functions related to PM structure and development in Gmm [24] as well as with chitin production or degradation processes [54–56]. Increased expression of genes encoding PM proteins, including Peritrophins (Pro1 and Pro3/trypsin), has also been observed in trypanosome infected Gpg [57]. We previously noted compromised PM function(s) in Gmm infected with trypanosomes in the gut and salivary glands as well as in newly eclosed young adults following a bloodmeal supplemented with trypanosomes [58, 59]. Because our results indicated variations in the expression of products involved in PM functions in Spi⁺ individuals relative to Ctrl, we investigated the status of PM integrity by using entomopathogen S. marcescens in a fly survival assay as we previously reported [24, 58–60]. Typically, flies with compromised PM survive longer as their gut epithelia can detect the presence of S. marcescens and elicit an immediate and robust response that eliminates the pathogen before it causes a fatal systemic infection [24]. We did not observe a statistically significant difference in host survival between Spi⁺ and Ctrl individuals (S2B Fig), suggesting that the presence of Spiroplasma does not influence the structural integrity of PM in the tsetse fly. It is possible that the increased expression of PM associated genes in Spi⁺ and Tbb⁺ individuals may represent a process of enhanced PM degradation in the presence of pathogens and hence the higher levels of PM associated gene expression to produce more PM proteins to accommodate this process.

In addition to PM-related functions, we detected immunity-related genes that were induced in both Spi⁺ and Tbb⁺ transcriptome datasets compared to Ctrl, including C-type lectin, mucin-5AC (GFUI18_012666), and nitric oxide synthase (NOS) (Figs 1C and 2SA). These products have been shown to be part of the immune response to trypanosome infections in Gmm [41], with the reactive oxygen species, nitric oxide (NO) generated by NOS, exhibiting potent trypanocidal activity [61, 62]. It is possible that the increased expression of these immune molecules in the presence of Spiroplasma may contribute to the enhanced parasite refractoriness observed in Spi⁺ individuals.

We also detected DE genes unique to each infection: 31 in the Spi⁺ group and 951 in the Tbb⁺ group (S2 Table Sheet 3 and 4). To gain a comprehensive knowledge on the biological process(es) affected by Spiroplasma or trypanosome infections, we subjected the microbe-specific DE genes (log₂FC ≥ 1) to gene ontology (GO) enrichment analysis using Blast2GO software. The GO enrichment analysis of downregulated putative products in both the Spi⁺ and Tbb⁺ datasets revealed a shared pathway linked to metabolic processes, particularly lipid biosynthesis,. This was supported by the presence of several genes encoding fatty acid synthases (FAS), fatty acyl CoA reductases, and Acyl-CoA binding protein (Figs 2A and S3A). Interestingly, the genes associated with the suppressed fatty acid biosynthesis pathway (GO:0006633) were among the most abundant and highly DE (normalized counts ≥ 700 and log₂FC ≥ 2) products in both the Spi⁺ and Tbb⁺ datasets (Figs 2B; S3A). A reduction of lipid availability in the gut could negatively impact host fitness and may also reduce pathogen survival. Despite the decreased expression of lipid-associated genes, we found three transcripts including carnitine O-palmitoyltransferase I, fatty acyl-CoA reductase wat like, and james bond that were upregulated in the Tbb⁺ dataset (S3B Fig). Additional biological processes significantly suppressed in the Tbb⁺ state included the acyl-CoA metabolic process, the tricarboxylic acid cycle, and amino acid metabolism (Fig 2A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Functional classification of DE-pathways and abundant DE genes unique in Spi⁺ or Tbb⁺ state.

A. Gene ontology (GO) enrichment analysis for biological processes was conducted with unique DE genes. B. Functional annotations are provided for some of the most abundant and highly DE genes unique to each microbidal infection state, defined by normalized counts ≥ 700 and (log₂FC) ≥ 2. In this represenetation, red dots indicate genes that are DE in the Tbb⁺dataset, while blue dots indicate genes that are DE in the Spi⁺ dataset.

https://doi.org/10.1371/journal.ppat.1012692.g002

Trypanosomes, but not Spiroplasma infections, induce canonical immune pathways in Gff

Among the genes induced by trypanosomes were those encoding products associated with immunity and adenylate cyclase signaling pathways (Fig 2A). In fact, some of the most abundant and highly DE genes (normalized counts ≥ 700, log₂FC ≥ 2) in the Tbb⁺ group included components of the Immune Deficiency (Imd) pathway, mucins, serine proteases, and redox balance associated proteins (Figs 2B and S3B). In contrast, these immunity related genes were not induced in the Spi⁺ group. The induced immunity-related genes in the Tbb⁺ group included two peptidoglycan recognition proteins (PGRPs) -PGRP-3 and PGRP-LA- along with various antimicrobial peptides (AMPs) such as Attacin A like, three Cecropins, Defensin A and Toll (Toll-like receptor 7), defense protein l(2) 34Fc, mucins, C-type lectin 37Db and two phenoloxidase 2. We also identified two genes encoding Homeobox family transcription factors, Wingless and Araucan, associated with the Wingless signaling pathway (S3B Fig). In Gmm, this pathway has been implicated in the regulation of expression of host microRNA-275, which in turn modulates PM-associated gene expression during the trypanosome colonization process [58]. Further gene families upregulated by trypanosomes included trypsins, six serine proteases (SPs), and two serine protease inhibitors (SPIs) linked with insect immunity (S3C and S3D Fig). SPs and SPIs play critical roles in modulating the expression of immune pathways (Toll and Imd) by regulating the activation of specific effectors following exposure to an infectious agent [63]. Such regulation ensures that the impact of protease-activated cascades remains localized in time and space [64]. Finally, we detected genes associated with detoxification processes in the Tbb⁺ group, including nine cytochrome P450 (CYPs) transcripts, six of which were induced and three reduced. In the Spi⁺group, we detected three CYPs that were DE, with two significantly induced and one reduced relative to controls (S3E Fig). The increased expression of CYPs in both the Spi⁺ and Tbb⁺ datasets may suggest a protective response to the heightened oxidative stress caused by the presence of trypanosomes and Spiroplasma.

Impact of Spiroplasma and/or trypanosome infection on Gff lipid metabolism

Our transcriptomic data indicate that infection with Spiroplasma and/or trypanosomes negatively impacts the expression of several genes associated with fatty acid biosynthesis, suggesting decreased synthesis and thus low levels of lipids critical for host physiology. We previously demonstrated that pregnant female Gff infected with Spiroplasma had significantly lower levels of circulating triacyl glyceride (TAG) compared to age-matched, pregnant Spiroplasma-negative flies [46]. These flies exhibited reduced fecundity [46], likely as the result of low levels of circulating TAG that make up an important component of tsetse milk [65]. Trypanosomes also scavenge nutrients from their environment, including lipids for metabolism and structural integrity [47, 66]. This nutrient competition may affect critical physiological processes of the host, such as reproductive fitness as previously reported in Gmm [48]. With this in mind, we repeated the experiment to investigate the impact of trypanosome infection, and Spiroplasma plus trypanosome infections (Spi⁺/Tbb⁺) on circulating TAG levels in virgin female Gff. We used virgin females in this analysis to exclude any effects on TAG levels that might be due to the varying stages of pregnancy. We compared TAG levels in the hemolymph of two-week-old virgin Spi⁺, Tbb⁺ and Spi⁺/Tbb⁺ females with their age-matched Ctrl females. Our results showed that virgin Ctrl females had higher levels of circulating TAG in their hemolymph (33.75±1.005 μg/μl) compared to their age-matched Tbb⁺ (27.72±1.438 μg/μl, p = 0.0001), Spi⁺ (24.90±1.316 μg/μl, p<0.0001) and Spi⁺/Tbb⁺ (19.59±1.438 μg/μl, p<0.0001) counterparts (Fig 3A). In addition, we also found that Spiroplasma-negative (Ctrl) Gff males present with significantly more (11.56±1.316 μg/μl) circulating TAG than do males infected with the endosymbiont (6.78±1.316 μg/μl, p = 0.0023) (Fig 3B). These findings indicate that the tsetse fly, Spiroplasma, and African trypanosomes compete for this metabolically critical TAG. This tripartite interaction likely underlies the reduced fecundity exhibited by Spi⁺ females [46], and may account at least in part for why flies that house this bacterium are more refractory to infection with trypanosomes.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Impact of Spiroplasma and/or Trypanosome Infection on Host Lipid Level.

A. Levels of triacylglyceride (TAG) circulating in the hemolymph of virgin female Gff were measure for Control (Ctrl), Spi⁺, Tbb⁺ and Spi⁺Tbb⁺ groups. Letters A, B and C indicate significant differences (p<0.05) in TAG levels, while the same letter indicates no significant difference (p>0.05). B. The amount of TAG in the hemolymph of Spi⁺ and Control (Ctrl) male Gff flies. Statistical significance was determined using ANOVA for virgin female flies and unpaired t-test for male flies.

https://doi.org/10.1371/journal.ppat.1012692.g003

A newly discovered AMP is expressed in the tsetse fly species in the Palpalis subgenera

One of the abundant and DE genes in the Tbb⁺ dataset was annotated as Stomoxyn-like (GFUI18_001176 or GFUI020894-RA), hereinafter referred to as GffStomoxyn (Fig 2B). The ortholog of GffStomoxyn was identified in related Diptera, including S. calcitrans (stable fly, designated as ScalStomoxyn), Lucilia sericata (green blowfly) and Hermetia illucens (black soldier fly) [67–69]. The 207 bp GffStomoxyn gene encodes a 68 amino acid (aa) pre-pro-mature peptide, which is composed of a 25 aa signal peptide at the N-terminus, followed by pro-mature peptides of 43 aa, similar in organization to the previously described Stomoxyns (Fig 4A). BLASTP homology searches of the 43 aa GffStomoxyn mature peptide against proteome databases of S. calcitrans and Musca domestica (housefly) identified two orthologs in S. calcitrans, (SCAU016907; E-value = 8.70E−10 and SCAU016937; E-value = 0.000632 annotated as ScalStomoxyn 2 and ScalStomoxyn, respectively) and a single ortholog in M. domestica (MDOA008330; E-value = 1.42E-13 annotated as Stomoxyn-like but hereinafter referred to as MdomStomoxyn). We further searched for the stomoxyn locus in other tsetse species using the WGS data from Gpp in the Palpalis subgroup, Gmm and Gpd in the Morsitans subgroup, Gau in the Austenina subgroup, and Gbr in the Fusca subgroup. This search revealed a single ortholog in Gpp (GPPI027903; E-value = 1.48E−40), while no orthologs were identified in the other four tsetse species. Additionally, we searched the non-redundant (nr) protein sequence database [70] and relevant literature for putative Stomoxyns. This search identified one ortholog in the flesh fly Sarcophaga bullata (DOY81_004902), the previously reported one in Lucilia sericata (XP_037825072.1; [68]), and two in the Australian sheep blowfly L. cuprina (XP_023308701.2; KAI8119624.1) (Fig 4A). Interproscan analysis of all identified peptides confirmed their classification within the Stomoxyn protein family. Furthermore, our search also revealed Stomoxyn orthologs in several other Diptera, including Episyrphus balteatus (XP_055851874.1) and Eupeodes corollae (XP_055904620.1) hoverfly species and the black soldier fly (H. illucens) (XP_037911389.1; XP_037913598.1) [69].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Molecular characterization of GffStomoxyn.

A. Sequence alignment of putative Stomoxyn peptides was performed, including sequences from S. calcitrans (ScalStomoxyn; SCAU016937 and ScalStomoxyn 2; SCAU016907), Gff (GffStomoxyn-like; GFUI18_001176), Gpp (GppStomoxyn; GPPI027903), M. domestica (MdomStomoxyn-like; MDOA008330), L. cuprina (LcupStomoxyn; KAI8119624.1 and LcupStomoxyn-like; XP_023308701.2), S. bullata (SbulStomoxyn; DOY81_004902) and L. sericata (LserStomoxyn-like; XP_037825072.1). The proteolytic cleavage site is shown in bold letters. The alignment indicates the pre-pro-mature domains of the full predicted peptides, with identical residues indicated by and asterisk (*). Conservative substitutions are marked with a colon (:) and semi-conservative substitutions are marked by a period (.). B. Distance matrix analysis was conducted with the mature peptide domains of Stomoxyns from different Diptera.

https://doi.org/10.1371/journal.ppat.1012692.g004

Multiple sequence alignment of the putative Stomoxyn peptides revealed a conserved structure comprised of pre-pro-mature domains. It is interesting to note that the signal peptide domain of Glossina Stomoxyn consists of 25 amino acids, while all other Stomoxyns have a 24-aa signal peptide. This difference includes an additional valine residue and a substitution of alanine with serine at the predicted cleavage site (Fig 4A). Distance matrix analysis of the mature peptide domains showed high identity among the homologs, with LcupStomoxyn and LcupStomoxyn-like sharing 97% identify. In contrast, the correspondig domains in the S. domestica orthologs, ScalStomoxyn and ScalStomoxyn 2, exhibited only 81% identity. Orthologs within closely related species showed high amino acid identity, such as Gff and Gpp at 97.67% and L. sericata and L. cuprina at 93.02% (Fig 4B). The amino acid identity within the mature peptide domain of Stomoxyns between M. domestica and S. calcitrans was 72%, while GffStomoxyn and GppStomoxyn exhibited 46% identity with the orthologs from M. domestica and S. calcitrans (Fig 4B). Phylogenetic analysis of mature stomoxyns’ amino acid sequences indicate that the Glossina stomoxyn-like loci cluster together and are sister to the stomoxys clade, derived from other muscoid flies (S4A Fig). This stomoxyn phylogeny coincides with host nuclear phylogenies [71], suggesting that further sampling across Calyptrate taxa would yield additional stomoxyn-like peptides. I-TASSER prediction of the putative tertiary structure for mature GffStomoxyn and ScalStomoxyn 2 indicate that they are composed of double α-helices and β-folds (S4B Fig), similar to the structure reported for ScalStomoxyn [72].

We analyzed the genomic context flanking the Stomoxyn loci in Gff and Gpp and compared their synteny using WGS data available for other Glossina species, where no orthologs were identified (S4C Fig). This analysis revealed several syntenic loci located immediately upstream and downstream of the Stomoxyn locus in both Gff and Gpp. Notably, several loci GffStomoxyn were retained on the same genomic scaffolds in Gpd, Gau and Gbr, while such syntenic loci were largely absent from Gmm WGS data. Additionally, a BLASTP search of the pre-pro-mature GffStomoxyn sequence against the putative proteomes of different tsetse species did not yield any significant hits, further suggesting that this protein-coding sequence is absent in those species. To rule out the possibility that the absence of the Stomoxyn locus was due to inadequate genome annotation, we performed de novo transcript assembly from midgut RNA-seq datasets available for Gmm and Gpd [58, 73] using Trinity software [74]. We then performed BLASTN searches with the stomoxyn sequences collated from Gff, Gpp, and other Diptera against our transcriptome-based assembly. This search also did not yield significant hits, further confirming the absence of this locus in Gmm, Gpd, Gau and Gbr.

We next investigated the presence of the stomoxyn locus in several Gff individuals obtained from distinct populations in North-West Uganda, as well as in Gpg from West Africa, Gbr and Gau from South Africa and Gau and Gpd from Kenya. We also analyzed this locus in a colony of Gpp that was distinct from the one used for the WGS analysis. PCR-based amplification of the stomoxyn loci—which included 5’ and 3’ UTRs, two exons and one intron–followed by sequence analysis of the products revealed over 99% identity at the nucleotide level between field collections and flies from laboratory lines. This high level of identity was also observed among Gff, Gpp and Gpg indicating the close evolutionary relatedness of these fly species (S4D Fig).

Analysis of stomoxyn expression profile and regulation

We profiled stomoxyn expression across various tissues, including cardia, midgut, fat body, female ovary and male testes. Our results indicate that stomoxyn is preferentially expressed in the cardia, with significantly lower expression detected in the midgut and other tissues (Fig 5A). Stomoxyn has been similarly reported to be preferentially expressed in the anterior midgut region of S. calcitrans [67]. Next, we examined the temporal expression profile of stomoxyn and found that gut transcript levels increased in adults post-eclosion, peaking at 72 h, when newly eclosed flies are mature enough to imbibe their first bloodmeal. Gut Stomoxyn levels measured 72 h after the first bloodmeal and following multiple bloodmeals remained equally high (Fig 5B). We also evaluated the expression of three antimicrobial peptides (AMPs), stomoxyn, cecropin and attacin, in the cardia of adult flies 72 h after their last bloodmeal. Similarly, we evaluated the expression of these AMPs in the cardia following systemic stimulation by E. coli or per os stimulation by a BSF Tbb-containing bloodmeal. Our results revealed significantly higher levels of stomoxyn expression in the cardia compared to the other AMPs and its expression remained high but unresponsive to immune stimuli (Fig 5C). The levels of both cecropin and attacin showed a significant increase following E. coli challenge, while only cecropin was significantly increased following trypanosome challenge (Fig 5C). We also evaluated the inducible nature of AMP expression in the gut (midgut and cardia) following immune stimulation of teneral Gff adults by per os challenge with E. coli, S. marcescens, and BSF Tbb. While expression of cecropin and attacin was significantly induced by E. coli and S. marcescens, stomoxyn expression remained high but unchanged in the gut (Fig 5D), similar to our findings in the cardia tissue. To see if Spiroplasma infection prevalence influenced stomoxyn levels, we quantified the stomoxyn expression in the gut of 15-day old adults infected with either Spiroplasma (Spi⁺) or Trypanosomes (Tbb⁺) and compared it to that found in uninfected control (Ctrl) guts. Results indicated that stomoxyn expression measured in the gut was not influenced by either Spiroplasma or trypanosome infection status of the host (Fig 5E). Additionally, using Gpg, another member of the Palpalis subgroup of tsetse flies, we measured the expression level of stomoxyn in the cardia of adult flies and compared its transcript levels to attacin and cecropin following immune challenge with a Tbb-supplemented bloodmeal (Fig 5F). We found that stomoxyn transcript levels were comparable between Gff and Gpg cardia tissues, and were significantly higher than those of the other two AMPs in the unchallenged state (Control). Similarly, we found that neither stomoxyn nor attacin were induced upon trypanosome challenge in the cardia in Gpg, while cecropin expression increased upon trypanosome challenge similar to our findings in Gff. Collectively, our results indicate that stomoxyn is expressed at high levels predominantly in the cardia tissue of both teneral and mature adult tsetse. This finding is consistent with previous reports of stomoxyn expression in S. calcitrans, where it was shown to localize to the anterior gut [67]. Unlike canonical AMPs, such as attacin and cecropin, which are induced in response to pathogen exposure, stomoxyn is constitutively expressed in the cardia, and is not responsive to microbial presence, similar to the expression pattern reported in S. calcitrans [67].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Spatial and Immune Responsive Expression of GffStomoxyn.

A. The qRT-PCR expression profile of stomoxyn was analyzed from the cardia, midgut, fat body, female ovary, and male testes tissues. Eight biological replicates were used, each consisting of three individual fly tissues collected from 10-day old adults. Results are presented relative to gapdh expression. Letters A, B and C above each bar represent significant differences in stomoxyn expression (p<0.05), while the same letter indicates no significant difference (p>0.05). B. Temporal expression of stomoxyn was measured from late pupa and whole gut tissue at multiple time points: 24 and 72 h post-eclosion, 72 h after the first bloodmeal (designated as 72hr*), and 15 day-old adult flies analyzed 72 h after last bloodmeal. Six to ten biological replicates, each comprised of individual midguts, were used in the experiment. Results are presented relative to gapdh expression. Letters A, B and C on top of each bar indicate significant differences (p<0.05) in stomoxyn expression, while the same letter indicates no significant difference (p>0.05).

https://doi.org/10.1371/journal.ppat.1012692.g005

C. The expression levels of stomoxyn, attacin and cecropin were analyzed in the cardia tissue of 8-day old adult Gff flies, 72 h after their last bloodmeal. The last bloodmeal for the E. coli and Tbb groups was supplemented with E. coli and Trypanosome (Tbb), respectively, while the Control group received a normal bloodmeal. This experiment included five biological replicates, each comprising three cardia tissues dissected from individual flies. Tbb indicates trypanosome challenge. *indicates significantly different gene expression (p<0.05) compared to the unchallenged control. D. The expression levels of stomoxyn, attacin and cecropin in the gut of newly eclosed flies were analyzed 24 h following a bloodmeal supplemented with E. coli, Tbb or Serratia, and compared to controls that received a normal bloodmeal. This experiment included nine to twelve biological replicates, each consisting of individual flies. Tbb indicate trypanosome challenge. *indicates significantly different gene expression (p<0.05) relative to the unchallenged control. E. The relative expression of stomoxyn was measured in the guts of 15-day old Spi⁺ or Tbb⁺ flies, and compared to age-matched uninfected controls. This experiment included eight to twelve biological replicates, each comprised of individual fly guts dissected 72 h post-bloodmeal. Trypanosome or Spiroplasma infection status of each fly was determined as described in the Materials and Methods section. Tbb⁺ indicates trypanosome infected Gff (Spiroplasma uninfected); Spi⁺ indicates Spiroplasma infected (Trypanosome uninfected). P-values for each comparison are shown above the corresponding bar. F. The relative expression levels of stomoxyn, attacin and cecropin were determined in the cardia tissue of 8-day old Gpg flies, 72 h after their third bloodmeal supplemented with trypanosomes. This experiment included five biological replicates, each comprising three cardia from individual flies. Tbb indicate trypanosome challenge. *indicates significantly different gene expression (p<0.05) relative to the control.

Antimicrobial activity spectrum of GffStomoxyn

We next investigated whether GffStomoxyn exhibits trypanocidal activity similar to that reported for ScalStomoxyn [67]. We commercially generated synthetic mature GffStomoxyn, and mature ScalStomoxyn [67] as well as its ortholog, ScalStomoxyn 2. We assessed the microbicidal activity of these three synthetic peptides against E. coli, Sodalis (the tsetse fly commensal symbiont) and Tbb (both BSF and PCF forms) in vitro using minimum inhibition concentration (MIC) assays. The MIC values for ScalStomoxyn, ScalStomoxyn 2 and GffStomoxyn against E. coli were 10 μM, 5 μM and 5 μM, respectively. None of the three peptides exhibited activity against Sodalis at concentrations up to the 100 μM (Fig 6A), consistent with the previously published results for ScalStomoxyn [67, 75]. In contrast, the minimum median inhibitory concentration (MIC₅₀) values for ScalStomoxyn, ScalStomoxyn 2 and GffStomoxyn against BSF trypanosomes were 18.4, 2.01 and 2.51 μM, respectively (Fig 6B and replicated in S5 Fig). These results indicate that ScalStomoxyn 2 exhibits significantly stronger killing activity against BSF trypanosomes compared to previously identified ScalStomoxyn (at 2.01 and 18.4 μM, respectively). GffStomoxyn also demonstrated a high level of trypanolytic activity, with an MIC₅₀ of 2.51 μM (Fig 6B and replicated in S5 Fig). However, all three peptides were ineffective against PCF trypanosomes with the peptides being able to kill less than 10% of the parasites at concentrations up to 100 μM (Fig 6B and replicated in S5 Fig). Our results indicate that GffStomoxyn exhibits strong antimicrobial activity against both gram-negative E. coli and BSF trypanosomes in vitro. Conversely, this peptide is not effective against the endosymbiotic Sodalis and insect stage PCF trypanosomes, both of which have coadapted to survive in the insect midgut environment.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. In vitro and in vivo Bioactivity of Stomoxyn.

A. In vitro antibacterial activity: The antibacterial activity of mature ScalStomoxyn, ScalStomoxyn 2 and GffStomoxyn peptides was assessed against E. coli and Sodalis bacteria. B. In vitro bioactivity against trypanosomes: The trypanocidal activity of mature ScalStomoxyn, ScalStomoxyn 2 and GffStomoxyn peptides was assessed against mammalian bloodstream forms (BSF) and insect stage procyclic form (PCF) trypanosomes. The blue lines indicate BSF parasite inhibition and the red lines show PCF parasite inhibion over the range of peptide concentrations tested. Different recPeptides are shown by varying symbols. C. Gene-silencing validation: The relative expression of stomoxyn was measured from dsGFP and dsStomoxyn-treated Gff flies, showing a significant decrease in expression (p<0.0001) in the case of dsStomoxyn-treated group. Ten biological replicates, each comprising individual flies, were used to validate silencing efficacy. D. Prevalence of trypanosome infections: The prevalence of midgut trypanosome infections in dsRNA-treated Gff flies was microscopically analyzed 15 days post-parasite acquisition, with significant difference observed (p<0.0011). The total number of flies (N) used in each experimental condition is shown. The experiment included a mixture of male and female flies across two biological replicates, with no significant difference in infection rates between the dsRNA-treated males and females.

https://doi.org/10.1371/journal.ppat.1012692.g006

Stomoxyn decreases susceptibility of tsetse fly to trypanosome infections

We employed an RNAi-based silencing approach to investigate the influence of Stomoxyn peptide on parasite establishment in Gff. Double-stranded RNAs, including dsGFP and dsStomoxyn, were injected into the thoracic haemocoel of teneral adult flies. Fourty-eight hours after treatment, we confirmed that GffStomoxyn expression in dsStomoxyn-treated flies was significantly reduced compared to the dsGFP treated group (p<0.0001; Fig 6C). At this time point, both dsGFP and dsStomoxyn flies were offered bloodmeal containing 1x10⁶ BSF trypanosomes and subsequently maintained on a normal blood diet. This experiment was replicated twice. On day 15, we analyzed gut parasite infection prevalence from two replicates and combined the for subsequent analysis. Chi-square analysis revealed that infection rates in the dsStomoxyn group were significantly higher (82/127; 64.56%) compared to the dsGFP -treated flies (64/153; 41.83%) at df = 1, X² = 10.63, p = 0.0011 statistical level of significance (Fig 6D). These findings indicate that Stomoxyn exhibits trypanocidal activity against BSF trypanosomes in vivo in the gut, similar to the effects observed with synthetic peptides in vitro.

Biological material

Glossina fuscipes fuscipes (Gff) pupae were obtained from the Joint FAO/IAEA IPCL insectary in Seibersdorf, Austria and reared at Yale University insectary at 26°C with 70–80% relative humidity and a 12 hour light:dark photo phase. This Gff line carries the heterogenous symbiotic infection, with about 50% of flies being negative (uninfected) for Spiroplasma. Newly eclosed teneral Gff females were challenged with 2x10⁶ Trypanosoma brucei brucei RUMP 503 strain BSF parasites (referred here as Tbb) per ml of blood in their first bloodmeal and were hereafter maintained on defibrinated bovine blood every 48 h through an artificial membrane feeding system [96] for 15 days. Another group of the Gff flies received normal bloodmeals, were maintained under the same conditions and used for uninfected controls. Whole gut (comprising cardia and midgut) were dissected 48 h after the last bloodmeal in Phosphate-buffered-Saline-Glucose (PSG) buffer (pH 8.0) and trypanosome infection status was microscopically determined using a Zeiss Axiostar Plus light microscope (Carl Zeiss Microscopy GmbH, Jana, Germany). All dissected tissues were stored individually at -80°C for RNA extractions. Legs corresponding to individual flies were also collected and kept separately at -80°C for genomic DNA (gDNA) production and determination of Spiroplasma infection status.

Genomic DNA extraction and Spiroplasma infection determination

The legs obtained above from individual flies were used to extract gDNA using the Monarch Genomic DNA purification kit following manufacturer’s protocol (New England BioLabs, MA, USA). A Gff specific alpha-tubulin primer set was used for gDNA quality control, and Spiroplasma-specific 16s rDNA primer set was used to determine the infection status of the bacterium in each fly. The Spiroplasma locus was amplified using touchdown PCR as previously described [45]. Based on the PCR results for Spiroplasma infection status, the samples were pooled for downstream analysis as either Spiroplasma uninfected (Ctrl: control negative for Spiroplasma) or Spiroplasma infected (Spi⁺; positive only for Spiroplasma). Flies that were provided a bloodmeal supplemented with Tbb and microscopically found to be infected as described above were similarly screened for the presence of Spiroplasma. Of the trypanosome positive flies, only those that were Spiroplasma negative were used for downstream analyses (Tbb⁺, positive only for trypanosomes). Primer sequences and PCR-amplification conditions used as in S1 Table.

RNA extraction, RNA-seq library preparation and Bioinformatics analysis

Total RNA was extracted from the gut of all adult individuals using TRIzol and subsequently treated with Turbo-DNase to eliminate contaminating DNA following the protocol described by the manufacturer (Thermo Fisher Scientific Inc., CA, USA). Elimination of DNA from the RNA was confirmed by PCR amplification using Gff-specific alpha-tubulin and glyceraldehyde-3-phosphate dehydrogenase (gapdh) primer sets (S1 Table). RNA quantity and quality was determined using an Agilent 2100 Bioanalyzer RNA Nano chip (Agilent, Palo Alto, CA, USA). RNA from two individuals was pooled per biological replicate and four, five and five biological replicates representing Ctrl, Spi⁺, and Tbb⁺ groups, respectively were obtained.

RNA-seq libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England BioLabs, Inc., MA, USA) according to the manufacturer’s protocol. The individual libraries were barcoded for Illumina HiSeq 2500 sequencing system (Illumina, Inc., CA, USA) and paired-end sequenced (100 bases) at Yale Center for Genome Analysis (YCGA, New Haven, CT). Read files are deposited in the National Center for Biotechnology Information (NCBI) archive, BioProject ID PRJNA1112339.

Raw RNA-seq reads were checked for quality using FASTQC and parsed through Trimmomatic for quality trimming [97]. For analysis, the Gff 2018 reference genome version 63 was obtained from VectorBase [98] (https://www.vectorbase.org/). The quality filtered reads were aligned using STAR (v2.3) [99] and ‘htseq-count’ function in the HTSeq (v0.11.2) [100] was used to count the number of reads mapped per gene with the intersection-nonempty mode. Using counts data from HTSeq, correlation between replicates within each condition was evaluated by calculating Pearson’s correlation coefficient (r) value. Only genes that had ≥ 10 reads mapping in at least 50% of the biological replicates for each experimental condition were used for downstream analysis. DESeq2 [101] was then used to determine differentially expressed (DE) genes between the control and the infected groups. Genes were considered significantly DE if they exhibited log₂ fold change either ≥ 1 or ≤ -1 at a p value < 0.05 and an adjusted p value < 10% [84, 102]. Gene ontology (GO) enrichment analysis of the DE genes was performed using the Blast2GO software [103] using the blastx [104] algorithm at significance threshold of 1×10⁻³ to search against non-redundant (NR) NCBI protein database. An enrichment analysis via Fisher’s exact test at an FDR, p-value ≤ 0.05 in Blast2GO was conducted to determine overexpressed GO terms, relative to the entire Gff transcriptome.

Investigation for Peritrophic Matrix integrity

To assess the effect of Spiroplasma infection on the PM integrity, we made use of a host survival assay following infection of the Gff line described above with Serratia marcescens strain db11 [24, 58]. Briefly, newly eclosed teneral flies were provided a bloodmeal supplemented with S. marcescens (1×10³ CFU/mL). All flies were then maintained on normal bloodmeals and the number of fly deaths were recorded every 48h. Dead flies at each collection date were kept individually at -80°C for subsequent assessment of Spiroplasma infection status via the PCR amplification assay described above. This assay was repeated twice.

Hemolymph Triacylglycerol (TAG) assay in Spiroplasma of Trypanosome infected Gff

Hemolymph (3 μl/fly) was collected as described in [105] from two-week-old Ctrl, Spi⁺, Tbb⁺ and Spi⁺/Tbb⁺ virgin female Gff, centrifuged (4°C, 3000xg for 5 minutes) to remove bacterial cells, diluted 1:10 in PBS containing 1.2 μl/ml of 0.2% phenylthiourea (to prevent hemolymph coagulation) and immediately flash frozen in liquid nitrogen [46]. These experiments used two-week-old flies, as this is the typical timepoint for assessing trypanosome infection status by microscopy. Unmated virgin flies were chosen for this analysis because triacylglycerol (TAG) is produced in tsetse’s fat body and exported into the hemolymph, where it is taken up by the milk gland and incorporated into milk secretions. Since TAG is the most abundant lipid in tsetse milk, even small differences in pregnancy stage (i.e., several hours) can significantly affect circulating TAG levels. By using virgin flies, were were able to eliminate pregnancy-related variations and focus on the effects of Spiroplasma or trypanosome infection. Hemolymph TAG levels were quantified colorimetrically by heating samples to 70°C for 5 min followed by a 10 min centrifugation 16,000xg. Five μl of the supernatant was added to 100 μl of Infinity Triglycerides Reagent (Thermo Scientific) and samples were incubated at 37°C for 10 min. Absorbance was measured at 540nm using a BioTek Synergy HT plate reader as described [46, 106]. Male Gff flies Ctrl and Spi⁺ were also included in the experiment and treated the same way. All Gff sample spectra data were compared to that generated from a triolein standard curve (0–50 μg, 10 μg increments). Analysis of Variance (ANOVA) was employed in the analysis of the amount of TAG circulating in the hemolymph of Ctrl, Spi⁺, Tbb⁺ and Spi⁺/Tbb⁺ female and t-test for circulating TAG in Gff male flies’ hemolymph using GraphPad Prism software v.7 (GraphPad Software, La Jolla CA, USA).

Genomic context and structural analysis of Stomoxyn peptides

To confirm the VectorBaseDB [98] annotation of Gffstomoxyn-like gene (GFUI18_001176 or GFUI020894-RA), Blastp [104] homology searches were performed using the putative proteomes for Stomoxys calcitrans and Musca domestica [92, 107]. This analysis revealed one ortholog in Glossina palpalis palpalis (GppStomoxyn, GPPI027903), the previously reported Stomoxyn peptide in S. calcitrans annotated as ScalStomoxyn (SCAU016907) [67] as well as an additional ortholog (annotated as ScalStomoxyn 2, SCAU016907) and a single ortholog in M. domestica (annotated as Stomoxyn-like MDOA008330). Search of published literature and analysis of NR database identified the presence of other orthologs in related Diptera, including Sarcophaga bullata (DOY81_004902), Lucilia sericata (XP_037825072.1), two in Lucilia cuprina (XP_023308701.2 and KAI8119624.1), Episyrphus balteatus (XP_055851874.1), Eupeodes corollae (XP_055904620.1) and two in Hermetia illucens (XP_037911389.1; XP_037913598.1). Full length protein sequence alignment of all significant Blast hits using clustalW [108] and InterProScan analysis [109] were performed to determine conserved Stomoxyn structural and functional domains. RAxML-NG with LG+G8+F model, with 25 parsimony, 25 random starting trees and 1000 bootstraps [110] was used to develop phylogenic tree based on MAFFT [111] sequence alignement of the mature Stomoxyn domain. The resulting tree was visualized with FigTree (http://tree.bio.ed.ac.uk/software/figtree/) and comparatively analyzed with that present in the Gff and Gpp proteome. The tertiary structure of the Stomoxyn peptides were predicted using the web based I-TASSER program Full length protein sequence alignment of all significant Blast hits using clustalW [108] and InterProScan analysis [109].

In order to understand the evolutionary history of this locus across Glossina, we obtained the WGS data available for the different tsetse species in VectorBaseDB, and compared the regions flanking the Gff and Gpp Stomoxyn loci with the same regions from Gmm, Glossina pallidipes (Gpd), Glossina austeni (Gau) and Glossina brevipalpis (Gbr), which lack GffStomoxyn ortholog. To do this, genes sequences flanking the GffStomoxyn (GFUI18_001176) in Gff supercontig JACGUE010000004 were compared with those in Gpp KQ080227 supercontig that contains GppStomoxyn via Blastp searches [104]. Similarly, Blastp was used to query the putative proteomes of other Glossina species for the presence of GffStomoxyn orthologs and the supercontigs where the ortholog exist in the genome.

The presence of the stomoxyn locus was also investigated from natural fly populations, Genomic DNA was prepared from Gff individuals collected in 2018 from the Albert Nile river drainage in Northwest Uganda (Amuru district: Gorodona (GOR; 3°15’57.6"N, 32°12’28.8"E), Okidi (OKS; 3°15’36.0"N, 32°13’26.4"E), and Toloyang (TOL; 3°15’25.2"N, 32°13’08.4"E)), Glossina palpalis gambienses (Gpg) collected from Bado Souleymane Mali (15°10’0”N, 7°31’0”W), Gbr and Gau from Hells Gate South Africa, Gau and Gpd from Shimba Hills Kenya. Genomic DNA was also prepared from a colony of Gpp maintained in Burkina Faso that is distinct from the colony in Seirbersdorf Vienna used for the WGS analysis and Gff colony from Seirbersdorf Vienna. The gDNA PCR primers (S1 Table) were designed to amplify the entire genomic sequence/region of stomoxyn in these species. When PCR products were obtained, they were analyzed by Sanger sequence and aligned using CLC Main Workbench (CLC bio, Cambridge, MA) against Gff and Gpp stomoxyn-like genomic loci from VectorBase.

Synthesis of Stomoxyn peptides

The predicted mature sequences of three Stomoxyn peptides were commercially synthesized and procured from ABClonal Science (500 West Cummings Park, Woburn, MA). These included the two homologs of Stomoxyn peptide from S. calcitrans, the previously described Stomoxyn (ScalStomoxyn) [67] and Stomoxyn 2 (ScalStomoxyn 2) discovered here from the S. calcitrans WGS data, and GffStomoxyn described here. The synthesized peptide sequences were RSLRKRLKKGVKNLRNTLKKTNNALKDAAGIAAGGAALGAAFG (GffStomoxyn), RSFRKRFNRFIKKIKHTISETAHVAKDAAVIAGSGAAVVAAAG (ScalStomoxyn 2) and RGFRKHFNKLVKKVKHTISETAHVAKDTAVIAGSGAAVVAATG (ScalStomoxyn). Purity of these synthetic peptides were 85.181%, 89.177% and 90.635%, for ScalStomoxyn, ScalStomoxyn 2 and GffStomoxyn respectively.

In vitro antibacterial and anti-trypanosomal activity of Stomoxyn peptides

The antimicrobial activity of the synthetic Stomoxyn peptides was tested against E. coli, tsetse endosymbiont Sodalis, and BSF and PCF trypanosomes. For antimicrobial activity, an overnight E. coli culture was inoculated into fresh medium and grown to logarithmic phase (OD₆₀₀ 0.3–0.4 at 37°C to yield about 1.5x10⁸ cells/ml). A 1:10 dilution of this culture was used to test the killing activity of different Stomoxyn peptide concentrations. Stock solutions (1 M) of synthetic peptides were prepared in water and 2-fold serial dilutions were made corresponding to 10 μM down to 1.25 μM. The E. coli culture was inoculated with synthetic peptides (0, 1.25, 2.5, 5 and 10μM concentrations) for testing the minimal inhibitory concentration (MIC). Cultures were grown for 1 h at 37°C and plated in duplicate on LB plates which were incubated overnight to measure colony forming units. For Sodalis killing assays, the bacteria were cultured on brain-heart infusion agar supplemented with 10% bovine blood as previously described [112] and grown in liquid cultures to OD₆₀₀ 0.3–0.4 at 25°C_. The synthetic peptides (0μM, 10μM, 20μM, and 100μM concentrations) were prepared and applied as described above for E. coli. The experiment was repeated three times with each synthetic peptide and microorganism.

The anti-trypanosomal activity of Stomoxyn was determined using Alamar Blue assay as described [67, 113] with slight modifications. Briefly, 500 BSF Antat 1.1 90:13 T. b. brucei cells in 50 μl were added to 96-well microtiter plates containing 50 μl of culture medium with a 2-fold serially diluted Stomoxyn peptide [114]. Top and low peptide concentration were evaluated at 100 and 3.125 μM respectively. The test for each concentration was performed in triplicates. As controls, culture media was included for baseline fluorescence as well as parasites without the Stomoxyn peptide. After 72 h of incubation at 37°C, 10 μl of Alamar Blue was added to each well, and the plates were incubated for another four hours. The plates were then read using BioTek cytation1 imaging reader (Agilent Technologies Inc., Santa Clara, CA, USA) at an excitation wavelength of 530 nm and an emission wavelength of 590 nm [67]. The acquired data was analyzed to produce sigmoidal inhibition curves and to determine IC₅₀ (median drug concentration inhibiting 50% of fluorescence development) values. The Stomoxyn percentage inhibition was calculated using the formula: % inhibition = 100[1-(X-MIN)/(MAX-MIN)] where: X was fluorescence of the sample, MIN fluorescence of the control (media without parasites) and MAX fluorescence of the positive control (culture parasites without drug) [115]. Stomoxyn activity for the PCF Tbb was determined following the same protocol above using cells grown in SDM-79 medium at 28°C [116].

Analysis of GffStomoxyn expression

Analysis of spatial expression of GffStomoxyn was performed on the cardia, midgut, fat body and ovary tissues from females, and testes tissue from males dissected from 10-day old Gff. For temporal expression analysis, whole guts, including cardia and midgut, were dissected from three week-old pupae, teneral adults 24 and 72 h post eclosion, adults 72 h after their first bloodmeal and 15-day old adults that have taken multiple bloodmeals were included in the experiment.

We exposed 8-day old Gff adults that have had two normal bloodmeals to microbial challenge. The first two groups (Control and E. coli) were provided a normal bloodmeal, while the third (Tbb) received a bloodmeal supplemented with 1×10⁶ BSF Tbb per ml of blood. Eight hours after the last bloodmeal, the flies in the E. coli group were exposed to 1000 CFU E. coli by intrathoracic microinjection. The cardia of all flies were dissected 72 h after the last bloodmeal, and five biological replicates, each containing three cardia, were obtained for downstream RNA extractions.

To determine the immune responsive profile of GffStomoxyn expression, three experimental groups of teneral flies (48 h post eclosure) were provided with bloodmeals supplemented with 1,000 colony forming units (CFU) of E. coli, or S. marcescens strain db11, or 1×10⁶ BSF Tbb per ml of blood, respectively. A fourth control group was similarly established that received a normal bloodmeal. All flies that did not take the bloodmeals were removed from the experiment and the guts of remaining flies from all four experimental groups were dissected 24 h after exposure for RNA extraction.

For analysis of trypanosome infected adults, Gff were provided with 1×10⁶ BSF Tbb per ml of blood in their first bloodmeal and subsequently maintained on normal uninfected bloodmeal for 15 days. Forty-eight hours post last bloodmeal, the midgut of these flies dissected and the presence of trypanosome parasites in the gut microscopically determined using a Zeiss Axiostar Plus light microscope (Carl Zeiss Microscopy GmbH, Jana, Germany). Trypanosome infected guts were individually stored at 80°C until RNA extraction. Another group of uninfected control Gff flies were similarly treated but maintained on uninfected bloodmeals. Legs corresponding to individual flies were collected and frozen separately at -80°C for gDNA extraction and determination of Spiroplasma infection status as described above. Only flies that were determined to be Spiroplasma negative (Ctrl), or Spiroplasma infected (Spi⁺) or trypanosome infected but Spiroplasma negative (Tbb⁺) were used for downstream RNA extractions, subsequent cDNA synthesis and gene expression analysis.

We provided 8-day old Gpg adults that have had two normal bloodmeals a third bloodmeal supplemented with 1×10⁶ BSF Tbb per ml blood and established a control age-matched group that received a normal bloodmeal. The cardia of all flies were dissected 72 h after the last bloodmeal, and five biological replicates, each containing three cardia, were obtained for downstream RNA extractions.

Total RNA was prepared (and DNase treated) from these samples as described above. cDNA was synthesized with oligo-dT primers and random hexamers using the iScript cDNA synthesis reaction kit (Bio-Rad, Catalog No. 170–8891) according to the manufacturer’s protocol. Real time quantitative PCR (qRT-PCR) GffStomoxyn and other genes (S1 Table) were performed in technical duplicate for each sample. The expression level of GffStomoxyn was evaluated between tissues and/or experimental conditions by qRT-PCR analysis using Gff gapdh as internal control. All qRT-PCR results were thus normalized to tsetse gapdh, quantified from each biological replicate in the tissues or experimental condition. The qRT-PCR data was analyzed using Relative Expression Software Tool (REST)-384 version 2 software [117].

RNAi gene silencing and trypanosome infection prevalence

Green fluorescent protein (gfp) and GffStomoxyn specific dsRNAs were prepared using the MEGAscript High Yield T7 transcription kit (Ambion, Huntingdon, UK) and gene specific dsRNA primers (S1 Table). The PCR products dsGFP and dsGffStomoxyn were sequenced to confirm their specificity for the gene of interest. To test the efficacy of gene silencing, groups of teneral male and female flies 48 h post eclosion were intrathoracically microinjected with 5 μg dsGffStomoxyn or dsGFP in 2 μl nuclease-free water and the midguts were dissected 48 h post treatment. RNA was extracted from the dissected guts and analyzed by qRT-PCR amplification for GffStomoxyn expression from each treatment group as described above. For trypanosome infection effects, two groups of teneral flies treated with dsGffStomoxyn or dsGFP as above were provided 1×10⁶/ml BSF Tbb in their first bloodmeal administered 48 h post dsRNA treatments. Flies that did not feed were discarded, and all remaining flies were subsequently maintained on normal diets. Fifteen days post-trypanosome challenge, all surviving flies were dissected, and their midguts microscopically examined for the presence of parasite infections. Chi-square was used to compare proportions of trypanosome infections between dsGFP and dsGffStomoxyn treatment groups. Gene silencing and trypanosome infection experiment was done twice.