Tyrosine detoxification is essential in tsetse

Tyrosine catabolism is a highly conserved pathway (Fig 1A) in most eukaryote and prokaryote species with only a few exceptions such as the pea aphid, Acyrthosiphon pisum [14], and trypanosome parasites [15]. The genes encoding TAT and HPPD proteins were identified in five Glossina species [16], as well as in all hematophagous arthropod species with sequenced genomes (S1 Fig). RNA interference (RNAi), of either TAT or HPPD genes, was lethal to flies once they fed on blood. However, 100% mortality was not observed, which may be explained by incomplete knockdown (Fig 1B and S2 Fig). This lethality was further validated by feeding flies with blood supplemented with mesotrione, an HPPD inhibitor widely used as a selective herbicide on corn crops under the brand name Callistro, Syngenta (S3A Fig). The mesotrione concentration that killed 50% of the insects 24 hours after administration (LC₅₀) was 357.7 μM (95% CI: 222.5 to 512.4) (S3A and S3C Fig). This lethal concentration of mesotrione is approximately 30× higher than the drug concentration detected in human plasma (4 μg/ml (11.78 μM)) after volunteers received an oral dose of 4 mg/kg body weight [17]. No differences in susceptibility to mesotrione (or NTBC) were observed between fly sex (S4 Fig).

Fig 1

Inhibiting tyrosine catabolism is lethal for tsetse, but not for bees.

(A) Tyrosine catabolism pathway. (B) Survival of Glossina morsitans morsitans when injected with dsRNA to knockdown TAT or HPPD. Two independent experiments were performed, each with n = 10–13, 12–16, and 12–14 for dsGFP (CTRL), dsTAT, and dsHPPD, respectively. Five insects from each group were dissected three days after a bloodmeal (PBM) to assess gene knockdown efficiency (S2 Fig). (C) Survival of G. m. morsitans fed with horse blood supplemented with NTBC or PBS (Control. 9:1; v/v). Three to six independent experiments were performed for different doses, each with n = 10–25 insects. The total number of flies (male and female) used was 914 (All doses applied are shown in S3B Fig). (D) Dose-response curves for G. m. morsitans survival 24 hours PBM: NTBC feeding and topical application assays (mean ± SEM). (E) Male Glossina pallidipes survival after feeding on PBS- or NTBC-treated rats. Doses are presented as mg of NTBC per kg of rat body weight. Three independent experiments were performed per dose, each with n = 12–15 flies. The total number of male flies used was 86. (F) Survival of NTBC-treated (topically) G. m. morsitans after a bloodmeal. Three independent experiments were performed for each dose, each with n = 10–20 insects. The total number of male flies used was 449. (G) Bee survival when maintained on a 50% sugar solution supplemented with NTBC (or PBS) and pollen. Two independent experiments were performed, each with n = 25 insects. The total number of bees used was 200. The underlying data for this figure can be found in S1 Data. CTRL, control; dsGFP, double-stranded green fluorescence protein; dsRNA, double-stranded RNA; FAH, fumarylacetoacetase; HgD, homogentisate 1,2-dioxygenase; HPPD, 4-hydroxyphenylpyruvate dioxygenase; MAAI, maleylacetoacetate isomerase; NTBC, nitisinone; PAH, phenylalanine hydroxylase; PBM, post-blood meal; TAT, tyrosine aminotransferase.

Ingestion of NTBC is lethal to tsetse

HT-1 is a severe human genetic disease caused by a mutation in the gene encoding for the last enzyme of the tyrosine catabolism pathway, fumarylacetoacetase (FAH). This mutation causes the accumulation of toxic metabolites in blood and tissues. The only drug available to minimise the effect of HT-1 is the orphan drug NTBC. As an HPPD inhibitor, NTBC prevents the buildup of toxic products derived from fumarylacetoacetate accumulation [18]. NTBC is remarkably safe to use with few reported side effects in <1% of patients [19,20]. When NTBC was fed to tsetse flies in an artificial bloodmeal, it was approximately 173 times more potent than mesotrione with a LC₅₀ = 2.07 μM (95% CI: 0.709 to 4.136) (Fig 1C and 1D). This lethal concentration is approximately 12 times lower than the concentration of NTBC persisting in human plasma (8 μg/ml (24.3 μM)) after administering a standard therapeutic oral dose (1 mg/kg) [17]. The NTBC half-life in human plasma is 54 hours [17], so assuming linear drug degradation, human blood would remain toxic to tsetse for at least a week after a single therapeutic dose. These results suggest that NTBC used as endectocide (here defined as a drug or insecticide that is administered to humans or animals to control parasites) [21–23] could be used to decrease tsetse populations as part of a drug-based vector control strategy. Furthermore, it did not matter if the tsetse flies were infected with T. brucei; NTBC lethality remained unaltered in infected flies (S5 Fig) suggesting that, unlike proline metabolism [24], the tsetse tyrosine degradation pathway does not seem to be affected during trypanosome infection.

Tsetse fed on NTBC-treated mice die

To evaluate the potential of using NTBC as an endectocide, the in vivo efficacy of NTBC was assessed. Colony-reared tsetse were allowed to feed on rats that had been orally treated with different NTBC concentrations. We observed that after 26 hours of feeding on rats receiving an oral dose of NTBC equal to 1 mg/kg (therapeutic dose recommended for humans with HT- I), approximately 90% of the flies died, as reflected in our previous data (Fig 1E). Furthermore, compared to control flies, tsetse fed on NTBC-treated rats showed the same phenotypic characteristics (including darkness of the abdomen (S6 Fig)) as those fed on NTBC-blood in artificial membrane feedings (see below). Together, these data provide direct evidence that NTBC could be used as an endectocide for tsetse control.

NTBC ingestion causes insect paralysis and systemic tissue destruction

Treatment of tsetse with either mesotrione- or NTBC-supplemented blood presented the same unique physiological changes, which has two distinct phases we classified as “early stage” and “death” phenotype. The early stage phenotype was observed 8 to 10 hours after NTBC (or mesotrione) ingestion in the bloodmeal, i.e., the flies remained alive (evidenced by red eyes) but were unable to fly (often upside down; S1 Video, S7 Fig). Exposure to bright light or tarsal stimulation often produced only a short burst of leg waving. Small, white, rhomboidal crystals (previously identified as tyrosine crystals [11]) were observed on the outside surface of the midgut epithelium against the body wall cuticle. Also, dark brown melanin-like deposits had formed in different tissues, such as the fat body, salivary glands, and flight muscle (Fig 2A and 2B). Disorganisation of tissues in the digestive tract such as the proventriculus and midgut was evident (Fig 2C) and likely contributed to intestinal fragility, compromised digestion, and leaking of the midgut content into the haemocoel. Reproductive and other non-digestive abdominal tissues remained intact.

Fig 2

Tsetse phenotypes observed upon NTBC treatment.

(A) Fat body. Scale bars = 100 μm (control) and 80 μm (treated). (B) Flight muscle. Scale bars = 1 mm. (C) Bright field (left) and fluorescence 3D-reconstructed (middle) overviews of the anterior midgut (AM), proventriculus (PV), crop (C), and oesophagus (E) from tsetse fed either without (Control) or with NTBC. Fluorescence channels merge SiR-actin staining (yellow) and DAPI (magenta). Scale bars = 200 μm (100×). Detailed sections (right) of transversal muscular fibres from the outer muscular tissues in the anterior midgut region. Scale bars = 20 μm (630×). (D) and (E) Death phenotype. Scale bars = 1 mm. Flies are already dead and the abdomens are “liquified”, black and filled with undigested blood. BF, brightfield; NTBC, nitisinone.

The death phenotype was observed between 15 to 18 hours after treatment and was characterised by a prolonged shift in eye colour from red to golden brown (Fig 2D). NTBC-treated tsetse showed extreme cuticular thinning and complete loss of abdominal elasticity. The abdomen was strikingly distended, dark, and filled with blood that had leaked into the haemocoel due to damage of the gut epithelium (Fig 2E). The ingested blood was blackened as though it had been oxidised. Most of the internal tissues and organs (e.g., gut, testes, ovaries, salivary glands, and Malpighian tubules) were completely destroyed and could not even be recognised in the abdomen during dissection, suggesting extensive autolysis (S8 Fig); only the hindgut, rectum, and trachea remained identifiable. Furthermore, the fat body also disappeared and only lipid droplets could be seen floating in the haemocoel. From the outside, the thorax appeared normal, but upon dissection, we observed the flight muscles had detached from the thorax and become melanised, which explains why tsetse quickly lost the ability to fly.

NTBC leads to an accumulation of free tyrosine and 4-hydroxyphenyl lactic acid

To better understand the metabolic changes in tsetse exposed to NTBC, haemolymph from NTBC-treated (and control) flies was collected at different time points. Metabolomic analysis revealed increased levels of tyrosine and HPLA (Fig 3). The knockdown of TAT gene, which is expected to reduce levels of 4-hydroxyphenylpyruvate (HPPA) and, in consequence, produce less HPLA, caused the same lethal phenotype observed upon HPPD inhibition. Collectively, this suggests that tyrosine accumulation is the likely cause of the flies’ death. Furthermore, feeding the flies with blood supplemented with HPLA or the injection of HPLA into the fly haemocoel did not affect their survival (S1 Table). In contrast to tyrosine and phenylalanine, the level of all other detected amino acids was reduced (Fig 3C and 3D), which may reflect a lower digestion rate in NTBC-treated flies.

Fig 3

Metabolomic analysis of tsetse haemolymph upon NTBC treatment at different times post-bloodmeal (0, 5, and 10 hours).

Five samples (in each group and for each time point) collected from two independent experiments were processed for and analysed by mass spectrometry. (A) Schematic representation of the tyrosine catabolic pathway highlights the main metabolites accumulated and the decreased production of final products. (B) Peak area normalised to internal standards of differentially regulated metabolites in the tyrosine catabolism pathway. ns = P > 0.05, * = P ≤ 0.05, ** = P ≤ 0.01, *** = P ≤ 0.001, **** = P ≤ 0.0001. (C) A heat map graphically presents the metabolites identified. (D) Schematic representation of the metabolic pathways indicates metabolites that were differentially regulated in NTBC-treated flies. The underlying data for this figure can be found in S1 Data. HPLA, 4-hydroxyphenyl lactic acid; HPPD, 4-hydroxyphenylpyruvate dioxygenase; NTBC, nitisinone; TAT, tyrosine aminotransferase.

Toxicity associated to insect HPPD inhibition depends on the protein concentration in the bloodmeal and not haem

A potential cause of toxicity among tyrosine catabolism inhibition could be haem. Haem is a product of blood (haemoglobin) digestion that becomes toxic by amplifying reactive oxygen species [25]. To investigate if haem contributed to the lethal phenotype, flies were fed with horse serum (red blood cells were removed to reduce haemoglobin) supplemented with mesotrione. Fly mortality remained high even in the absence of haemoglobin, which suggested haem was not involved in the toxicity of HPPD inhibitors. Furthermore, this experiment demonstrated that the protein content in horse serum (51 to 72 mg/ml) [26] is high enough to produce toxic levels of tyrosine (Fig 4A).

Fig 4

NTBC-associated toxicity depends on the concentration of protein in the bloodmeal.

(A) Percent survival of flies fed with horse serum supplemented with PBS (control) or different mesotrione concentrations. Three independent experiments were performed: n = 10–25 tsetse per treatment (355 insects in total). (B) Tsetse mortality (percentage) was calculated after feeding flies with different concentrations of BSA as a protein source alongside a lethal concentration (0.001 mg/ml) of NTBC. A protein concentration of 12 to 15 mg/ml is required to induce NTBC lethality (dotted horizonal line, LC₅₀). Four independent experiments were performed; each BSA concentration tested used n = 10–80 flies (788 insects in total). Data are shown as mean ± SEM. The underlying data for this figure can be found in S1 Data. BSA, bovine serum albumin; NTBC, nitisinone; PBM, post-blood meal.

In order to assess the quantity of protein necessary to cause the lethal phenotype, flies were fed with PBS or sugar supplemented with a fixed concentration of NTBC (lethal when added to bloodmeal) and different concentrations of bovine serum albumin (BSA) as the sole protein source. As expected, flies that fed on NTBC-supplemented diluent (protein-free) did not die, whereas NTBC toxicity showed a clear dependence on the protein content of the meal. The observed LC₅₀ was 12.5 mg/ml of protein (95% CI: 11.09 to 13.67) (Fig 4B). Only a small percentage (approximately 6%) of flies died 24 hours after feeding with either PBS-BSA (34 mg/ml) or fructose-BSA without NTBC (S9 Fig), indicating that the 24-hour mortality was due to the addition of NTBC.

Topical application of NTBC kills tsetse

To investigate if NTBC-induced mortality was restricted to ingestion, we tested if the drug could be absorbed through the insect cuticle. Topically applying NTBC to the fly thorax caused high mortalities after tsetse took a bloodmeal, with an LD₅₀ of 39 picomoles/fly (95% CI: 9 to 90) (Fig 1D and 1F), 24 hours after bloodmeal ingestion. We compared this NTBC dose against a standard pyrethroid insecticide (deltamethrin) that tsetse are highly susceptible to, and the LD₅₀ was 0.116 picomoles/fly (95% CI: 0.093 to 0.145) (S10 Fig), which means that deltamethrin is approximately 336 times more potent than topically delivered NTBC.

In the field situation where HPPD inhibitors could be absorbed through the insect cuticle, fly contact with the drug could occur either before or after a bloodmeal. Thus, it is important to know how long NTBC activity persists in insects, particularly if tsetse feed after cuticular exposure. High fly mortality was observed even when a tsetse ingested a bloodmeal six days following a single topical application of NTBC (S11 Fig). To address the opposite situation where blood-feeding precedes topical exposure to NTBC, the drug was topically applied at different times post-bloodmeal. NTBC reduced tsetse survival up to 48 hours after a bloodmeal (S12 Fig), which implies that bloodmeal proteins were digested after two days (as predicted for tsetse) and excess tyrosine had been catabolised. In contrast to NTBC, topical application of mesotrione was not lethal to tsetse, which could be due to reduced penetration through the insect cuticle (S13 Fig).

NTBC is not metabolised by insect P450 enzymes

Metabolic resistance due to increased rates of insecticide metabolism by P450s can cause resistance liabilities for new compounds. However, NTBC appears to only be moderately metabolised by CYP3A4 in humans [27], with little oxidative metabolism by other liver CYP enzymes [28]. We have incubated NTBC with microsomes extracted from tsetse, Aedes, and Anopheles and failed to detect evidence of metabolism as measured by substrate depletion/turnover (S2 Table). Furthermore, we find no obvious metabolism by recombinant Anopheles gambiae CYP6P3, a P450 with broad substrate specificity similar to CYP3A4, and associated with cross-resistance to pyrethroids and other insecticides [29]. Overall, this suggests that NTBC may be a weak substrate for P450 metabolism in these insects. This does not preclude the evolution of metabolic resistance, but suggests it may be less likely to emerge rapidly.

NTBC is highly stable under environmental conditions

Considering uses for NTBC in the field, we examined several stressors that could degrade drug activity and reduce its toxicity. An NTBC solution was subjected to 10 freeze–thaw cycles, prolonged ambient temperature storage (5 weeks), or different exposures to light (S3 Table). These NTBC test samples (final concentration in blood 3 μM = 1 μg/ml) were screened for activity by adding them into a tsetse bloodmeal as a type of activity bioassay. In all cases, the fly mortality resulting from the stressed samples matched that of freshly made NTBC, thus indicating NTBC does not quickly degrade and is stable under simple, room temperate storage conditions. In agreement with our results, Barchanska and colleagues [30] recently demonstrated that NTBC shows considerable stability under different experimental conditions such as pH of solution, temperature, time of incubation, and ultraviolet radiation.

NTBC is not toxic to the insect pollinator, Bombus terrestris

Bees are the world's most important pollinators of food crops [31], and several reports have shown that bee populations are directly and indirectly affected by insecticides [32]. To determine the environmental impact of NTBC on off-target species (non-bloodfeeders), we investigated the mortality of phytophagous pollinators when exposed to NTBC. Colony-reared Bombus terrestris (the buff-tailed bumblebee) were fed ad libitum with NTBC-supplemented sugar as a sole hydration source; bee mortality was assessed over 10 days. Mortality rates did not differ between NTBC-treated and control bees during this sustained exposure, despite feeding NTBC at doses as high as 50 μg/ml (152 μM) and providing pollen as a protein source (Fig 1G).

A mathematical model supports the use of NTBC for tsetse control

To illustrate the impact that NTBC may have in controlling African trypanosomiasis, we include human and livestock treatment in an epidemiological model and simulate the resulting reduction in transmission (details in S1 Methods). Fig 5 shows what NTBC treatment regime is required to interrupt transmission for systems with diverse vector biting ecologies. The temporal dynamics associated with control are shown in S14 Fig. Fig 5 is generated by averaging the disease control achieved over the time between doses. When both livestock and humans receive treatment (at 80% coverage), monthly dosing can control disease spread if at least 20% of bites are on humans. If fewer bites are on humans, this only becomes a feasible stand-alone control strategy when employed at higher frequencies (e.g., every 10 days). When wildlife (preferred host for some tsetse species) comprises a substantial blood source, treatment regimens become less tenable. At the current plasma efficacy half-life, NTBC will be most useful as part of an integrated disease control strategy. Simulations were also repeated for a transmission system in which animal reservoirs play a negligible role such as in many regions endemic for trypanosomiasis (T. b. gambiense) where humans are the primary reservoir. Here, absence of the animal reservoir facilitated disease control when humans comprised at least 20% of tsetse fly bloodmeals and NTBC dosing was monthly (S15 Fig).

Fig 5

Impact of NTBC application on the effective reproduction number of T. brucei, (R_e > 1 in brown and R_e < 1 in purple).

Top row: NTBC applied to livestock only; Middle row: Only humans are treated with NTBC; Bottom row: Simultaneous treatment of both livestock and humans with NTBC. Treatments are modelled with NTBC application every 10 days (left column), every 30 days (middle column), or every 90 days (right column). The axes denote the proportional split of bloodmeals taken by local tsetse populations when humans, livestock, and wildlife are potential hosts (where proportion of bites on wildlife are 1 - (bites on humans + bites on livestock)). NTBC, nitisinone.

Ethics statement

Male (4 to 6 weeks old, 180 to 250 g in weight) Wistar rats sourced from the Animal Rearing Unit and Containment Unit (ARCU) at International Centre of Insect Physiology and Ecology (icipe) were used. The animals were housed in standard plastic rodent cages (Thoren Caging Systems, Hazleton, United States) with wood shavings as bedding material. The rodents were maintained on commercial food pellets (Unga Kenya, Nairobi, Kenya), and water was provided ad libitum. Animal use and all accompanying procedures and protocols were in accordance with The Guide for the Care and Use of Laboratory Animals (Institute for Laboratory Animal Research, 2011). These procedures and protocols were reviewed and approved by Institutional Animal Care and Use Committee (IACUC) of icipe (Ref. No. IcipeACUC2018-003).

Tsetse

The G. m. morsitans Westwood colony (originally from Zimbabwe) was housed in an insectary at the Liverpool School of Tropical Medicine. Flies were kept at 26°C and 73% ± 5% relative humidity and a 12-hour light:dark cycle. The colony was maintained on defibrinated horse blood (TCS Biosciences, Buckingham, United Kingdom) and fed every 2 to 3 days using artificial silicon feeding membranes. Male and female experimental flies of different ages were separately caged and fed. The G. pallidipes (originally from Nguruman, Kenya) at the Insectary Unit at icipe were maintained at 26°C and 75% ± 4% relative humidity with 12-hour light:dark cycle. The flies were fed on defibrinated bovine blood using an in vitro silicon membrane system. Only teneral G. pallidipes were fed on NTBC-treated rats.

Bumblebee rearing

Hives of Bombus terrestris audax ordered from Agralan (Wiltshire, United Kingdom) were kept at 27°C under constant red light and fed ad libitum with pollen and 50% sugar water (Ambrosia syrup, EH Thorne). Only worker bees were used in this study. Five bees from five different colonies were placed into an acrylic box in five biological replicates, totalling 25 bumblebees per each experimental group. NTBC was diluted in sugar solution and placed into the box 24 hours after the bees acclimatised to laboratory conditions. Bees were allowed to continually drink NTBC-supplemented water ad libitum for 10 days. Each group received PBS or a dose of NTBC at concentrations shown to be lethal to tsetse (0.05 mg/ml, 0.005 mg/ml, or 0.0005 mg/ml). Solutions were changed every five days. Worker bees were also provided with pollen (protein source) according to their rate of consumption. Bee mortality was scored daily for 10 days.

Phylogenetic analysis

Dendrograms showing the phylogenetic relationships of TAT and HPPD proteins among insects were created according to the maximum likelihood method. Confidence values for each branch were determined through bootstrapping at 100. Analysis was performed on the web-based interface on Phylogeny.fr [57,58], using T-Coffee for alignments under the BLOSUM62 scoring matrix and optimised with Gblocks. Dendrograms were constructed with PhyML 3.0 [59]. The TAT and HPPD protein sequences used in the phylogenetic analyses were taken from either Vectorbase or NCBI using the most recent gene sets available.

Synthesis of double-stranded RNA (dsRNA)

Specific primers for G. m. morsitans TAT and HPPD genes were designed using primer-blast software (NCBI, https://www.ncbi.nlm.nih.gov/tools/primer-blast) (Table 1). These primers contained the T7 polymerase binding sequences required for dsRNA synthesis at 5′ end. The green fluorescence protein (GFP) gene amplified from the peGFP-N1 plasmid (NCBI ID: U55762.1) was used as a control dsRNA to assess in vivo off-target effects. PCR products were sequenced to confirm identity. MEGAscript High Yield T7 Transcription kit (Ambion, Austin, United States) was used according to the manufacturer's instructions to synthesise dsRNA. After synthesis, dsRNAs were precipitated by adding an equal volume of ice-cold isopropanol, and the resulting pellets were washed with ethanol, air dried, and then the pellet resuspended in ultrapure water. The dsRNAs concentrations were determined spectrophotometrically at 260 nm on a NanoVue Plus Spectrophotometer (GE Healthcare, Buckinghamshire, United Kingdom) and visualised in an agarose gel (1.5% w/v) to determine dsRNA size, integrity, and purity. An ISS110 speedvac concentrator (Thermo Scientific, Waltham, United States) was used to dry the dsRNAs and samples were adjusted to a final concentration of 5 μg/μl in sterile nuclease-free water (NFW).

Table 1

T7 promoter sequences that were necessary for transcription are underlined.

Sequences of the primers used to amplify target genes for RNAi experiments.

Gene	NCBI/Vector Base ID	Forward primer	Reverse primer
Green fluorescence protein	U55762.1	TAATACGACTCACTATAGGGACGTAAACGGCCACAAGTTC	TAATACGACTCACTATAGGGCTTGTACAGCTCGTCCATGCC
Tyrosine aminotransferase	GMOY012088	TAATACGACTCACTATAGGGAGACGAAGTGACTGCCGGTCTACG	TAATACGACTCACTATAGGGAGATTCACGAGGCACTGTTAGCAC
Hydroxyphenylpyruvate dioxygenase	GMOY012145	TAATACGACTCACTATAGGGAGAATCGCAGCCAATATCGTGGTG	TAATACGACTCACTATAGGGAGACTTTAATTTTGGTGCGGCTGTGC

Gene silencing

The protocols described by Walshe and colleagues [60] were followed for RNAi knockdown in tsetse. Briefly, male G. m. morsitans were injected in the thorax with 10 μg of each target gene dsRNA (2 μl of 5 μg/μl stock). Tsetse controls were injected with 10 μg of GFP dsRNA. Injection needles were handmade using pulled micro-haematocrit glass capillary tubes (2.00 mm outside diameter) (Globe Scientific, Mahwah, United States) mounted inside 200 μl yellow micropipette tips and sealed with Araldite epoxy glue. Flies were injected 24 hours after receiving a bloodmeal to increase survival rates. Flies were chilled for 10 minutes on ice to immobilise them, and 2 μl of dsRNA was injected into the dorsolateral surface of the thorax. Injected flies were allowed to rest for 24 hours before the next bloodmeal. Following injections, flies were fed every 2 to 3 days on sterile, defibrinated horse blood. Three days after dsRNA injection, the digestive systems were excised to determine the efficacy of gene-silencing as evaluated by quantitative PCR (qPCR).

RNA isolation and cDNA synthesis

G. m. morsitans flies were immobilised on ice, and the digestive system was dissected into ice-cold PBS. The total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, United States), according to the manufacturer's instructions. Following treatment of the extracted mRNA with Invitrogen Ambion TURBO DNase (Thermo Fisher Scientific, Loughborough, UK), first-strand cDNA synthesis was performed using 1 μg total RNA with “Superscript III First-strand Synthesis System for RT-PCR Kit” (Thermo Fisher Scientific, Loughborough, UK) and poly-T primer, according to manufacturer instructions. The cDNAs were stored at −80°C until use.

Quantitative polymerase chain reaction (qPCR)

Gene-specific qPCR primers were designed to amplify a different region from that amplified by the RNAi primers to prevent dsRNA amplification. They were designed to span different exons to easily discern contaminating genomic DNA amplification. Primer efficiency was experimentally tested (Table 2). Glossina α-tubulin and β-tubulin genes were used as references (housekeeping) genes. Quantitative PCR was performed using a MxPro – Mx3005P thermocycler (Agilent Technologies, Santa Clara, United States) with Brilliant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent Technologies, Santa Clara, CA, USA) under the following conditions: 95°C for 15 minutes, followed by 40 cycles of 95°C for 15 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, and a final extension of 72°C for 10 minutes. Target gene expression levels were assessed as 2e^−ΔCT values (ΔC_T = C_T gene of interest – C_T housekeeping gene) and were used to evaluate the mRNA levels of the genes in the different experimental groups (dsTAT- or dsHPPD-injected flies and dsGFP-injected flies) [61].

Table 2

Sequence of the primers used to amplify target genes by real-time PCR.

Gene	Vector Base ID	Forward primer	Reverse primer	% Efficiency
Alpha tubulin	GMOY004645	TGTATGTTGTATCGTGGTGATGT	GAATTGGATGGTGCGTTTAGTTT	100.6
Beta tubulin	GMOY000148	CCATTCCCACGTCTTCAGTT	GACCATGACGTGGATCACAG	96.6
Tyrosine amino transferase	GMOY012088	CCTAGCAATCCGTGTGGTAGTG	TAACCGCTATGTGCTGCGAAC	103
Hydroxyphenyl pyruvate dioxygenase	GMOY 012145	CTAAAGAACGTGGAGCAACTGTG	TCCACAAAAGTGTGAGTCGT	106.5

Oral dosing of mesotrione and NTBC

Mesotrione (2-[4-(Methylsulfonyl)-2-nitrobenzoyl]cyclohexane-1,3-dione; PESTANAL analytical standard; Sigma-Aldrich, St. Louis, United States) and NTBC (2-[2-nitro-4-(trifluoromethyl)benzoyl]cyclohexane-1,3-dione; PESTANAL, analytical standard; Merck Life Science UK Limited, Gillingham, UK) were solubilised in sterile PBS (NaCl 0.15 M, Na-phosphate 10 mM (pH 7.0)) to different final concentrations (pH was readjusted to 7.0 with 1 M NaOH). One volume of the drugs, or PBS as control, was then mixed with nine volumes of sterile defibrinated horse blood, and these bloodmeals were fed to male and female flies. Only flies with visible redness in the abdomen were selected, and unfed flies were discarded. Final mesotrione concentrations used in blood were: 2, 1.5, 1, 0.75, 0.5, 0.25, 0.1, 0.05, and 0.01 mg/ml. Final NTBC concentrations in blood were: 0.05, 0.025, 0.01, 0.005, 0.0025, 0.00075, 0.0005, 0.00025, and 0.0001 mg/ml).

Confocal microscopy

Teneral (<24 hours old post emergence) male tsetse were fed on horse blood (TCS Biosciences, Buckingham, United Kingdom) and three days later were offered a second meal containing horse serum (to increase tissue visibility by removing red blood cell interference) with or without 500 ng/mL NTBC. Fifteen hours post-feeding, tsetse were anaesthetized on ice, and the midgut tissue (with attached proventriculi) was dissected into ice-cold PBS and immediately fixed in fresh 4% paraformaldehyde for 1 hour at room temperature (RT). Tsetse group survival rates were determined 72 hours post-NTBC administration. Fixed tissues were washed in PBS (and stained with SiR-actin (1,100 dilution, Cytoskeleton Inc.) for 3 hours at RT and posteriorly incubated in 500 ng/mL DAPI for 10 minutes at RT. Tissues were finally suspended in 1% (w/v) low-melting agarose at approximately 40°C mixed with Slowfade Diamond oil (Molecular Probes, Eugene, United States) on a slide. Slides were imaged using a Zeiss LSM-880 confocal laser scanning microscope and tissues were 3D-reconstructed from a series of z-stacks at intervals of 1.3 μm (Zeiss Company, Oberkochen, Germany).

NTBC treatment of trypanosome-infected flies

Newly emerged (24 to 48 hours post-eclosion), teneral male G. m. morsitans were fed an infected bloodmeal containing fly-infective T. b. brucei (strain Antat 1.1 90:13) [62] at 10⁶ parasites per ml of blood. Newly emerged colony-reared flies are highly susceptible to trypanosome infections. Nine days after infection, when procyclic trypanosomes should be established in the midgut, the flies were fed with blood supplemented with NTBC as indicated above.

Oral dosing of mesotrione in serum

Horse red blood cells were removed from defibrinated horse blood by centrifugation at 1,000 × g for 5 minutes. The horse serum was collected, supplemented with mesotrione as indicated above and subsequently fed to female and male G. m. morsitans.

Feeding fructose supplemented with BSA and NTBC

Teneral, male G. m. morsitans were sorted into 10 cages and starved for 48 hours. BSA fraction V (stock concentration at 200 mg/ml (w/v) in PBS) was serially diluted into sterile 0.1% (w/v) fructose in PBS to create final concentrations of 40, 34, 30, 25, 20, 18, 16, 15, 12, 10, 5 and 0 mg/ml BSA. Control flies were fed with either diluent alone (0.1% fructose) or diluent plus NTBC. All experimental groups were fed a NTBC dose of 0.001 mg/ml (lethal when delivered with blood or serum). Flies were offered their first meal at 72 hours post-emergence, and mortality was tracked daily for one week.

Administering HPLA to tsetse by feeding or injection

Newly emerged, male G. m. morsitans were sorted into 10 cages containing 25 flies/cage. At 48 hours post-emergence, each cage was fed on defibrinated horse blood supplemented with distilled water (control) or a serial dilution of DL-p-hydroxyphenyllactic acid (HPLA: Aldrich-Sigma, St. Louis, United States). All dilutions were made with the same ratio: 25 μl of additive to 1.975 ml of blood. The concentrations of HPLA tested were 0.05, 0.025, 0.01, 0.005, 0.0025, 0.0075, 0.0005, 0.00025, and 0.0001 mg/ml. All dilutions were kept on ice until feeding to avoid chemical degradation. Fresh HPLA stock concentration was 2 mg/ml in sterile water. After 3 days, flies were offered a normal bloodmeal to test if HPLA impaired a second feed. To administer HPLA by injection, 1-week old male flies that had taken 3 previous bloodmeals (to ensure good health and adequate hydration) were separated into 10 cages. Each group received 2 μl of the selected HPLA concentration via thoracic injection. Stock HPLA was 2 mg/ml, and the concentrations injected were 1.0, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001, 0.0005, and 0.0001 mg/ml with sterile water injected as a control. Flies were allowed to recover for 24 hours and then offered a normal bloodmeal. Fly mortality was monitored throughout in both experiments for a week after HPLA administration, and no group exceeded 10% mortality.

In vivo oral administration of NTBC to rats

Rats received an oral dose of NTBC solubilised in sterile PBS by gavage. The doses administered were 0.1, 0.2, 0.5, 1.0, and 2.0 mg/kg, with controls receiving an equal volume of PBS. After 90 minutes, the rats were anaesthetized with an intraperitoneal injection of 80 μl of 3% ketamine and Xilazin 0.33% diluted in sterile PBS. Ketamine and Xilazin do not affect tsetse mortality. Groups of G. pallidipes flies were fed on either PBS- or NTBC-treated anaesthetized rats.

Topical application of mesotrione and NTBC to tsetse cuticle

Mesotrione and NTBC were solubilised in 100% acetone at the following concentrations (Mesotrione: 10, 5, 2.5, 1, 0.5, and 0.1 mg/ml. NTBC: 1, 0.5, 0.25, 0.125, 0.0625, 0.03175, 0.0156, 0.0078, and 0.0039 mg/ml). One microliter of solution was topically applied with a P2 micropipette to the ventral surface of the abdomen of a cold-anaesthetized male G. m. morsitans, either immediately before ingesting a blood meal, at different times before or after a bloodmeal. The control flies received 1 μl of 100% acetone. Only fully engorged flies were used.

Metabolomic analysis

Male G. m. morsitans were dissected at different times (0, 5, or 10 hours) after ingesting horse blood supplemented with NTBC 0.01 mg/ml or PBS. The flies were chilled for 10 minutes at 4°C and kept on ice. The legs of 20 flies were severed and collected into 1.5 ml plastic tubes containing 300 μl of 0.9% NaCl and 13.34 μM phenylthiourea (internal standard that prevents melanization) in ultrapure water. The samples were vortexed for 1 minute, centrifuged for 10 minutes at 14,000 rpm to collect the supernatant and then syringe-filtered (Millipore 0.22 μm syringe filter) to remove cells, such as haemocytes, present in the haemolymph. Filtered samples (150 μl total volume) were transferred to a 1.5 ml tube containing 50 μl of chloroform, 75 μl of methanol, and 75 μl of water containing B-methylamino-L-Alanine (53.4 μM) as an internal standard for normalisation. After centrifugation, the polar and apolar phases were collected and dried in a speed vac concentrator (ISS100, Thermo Scientific, Dreieich, Germany), and samples were stored at −80 until analysis. Five samples in total were collected from 2 independent experiments for each group and time point.

LC-MS and LC-MS/MS analysis

Polar extracts were reconstituted in Methanol:Water (1:1, V/V) containing 5 μM 13C5, 15N-Valine as an internal standard. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was performed as described previously [63]. LC analysis was performed using a Dionex UltiMate LC system (Thermo Fisher Scientific, Loughborough, UK) with a ZIC-pHILIC column (150 mm × 4.6 mm, 5 μm particle, Merck Sequant, Umea, Sweden). A 15-minute elution gradient of 80% Solvent A (20 mM ammonium carbonate in Optima HPLC grade water, Sigma Aldrich) to 20% Solvent B (acetonitrile Optima HPLC grade, Sigma Aldrich) was used, followed by a 5-minute wash of 95:5 Solvent A to Solvent B and 5-minute re-equilibration. Other parameters were as follows: flow rate 300 μl/min; column temperature 25°C; injection volume 10 μl; and autosampler temperature 4°C.

All metabolites were detected across a mass range of 70 to 1050 m/z using a Q Exactive Orbitrap instrument (Thermo Fisher Scientific, Loughborough, UK) with heated electrospray ionisation and polarity switching mode at a resolution of 70,000 (at 200 m/z). MS parameters were as follows: spray voltage 3.5 kV for positive mode and 3.2 kV for negative mode; probe temperature 320°C; sheath gas 30 arbitrary units; and auxiliary gas 5 arbitrary units. Pooled biological quality control (PBQC) samples were analysed throughout the run to provide a measurement of the stability and performance of the system. Parallel reaction monitoring (PRM) was used at a resolution of 17,500 to confirm the identification of metabolites; collision energies were set individually in high-energy collisional dissociation (HCD) mode. Data were acquired using Xcalibur 3.0.63 (Thermo Fisher Scientific, Loughborough, UK), and Progenesis (Nonlinear Dynamics, Newcastle, United Kingdom) was used for data alignment and peak detection. Data were normalised against internal standards. Metabolites were considered significantly altered using 2-way ANOVA (p < 0.05) and Bonferroni post hoc test. Annotations were assigned to accurate masses with a maximum error of 5 ppm using Metlin [64], LipidMaps [65], Kegg [66], and HMDB [67], which were searched simultaneously using the CEU Mass Mediator engine (http://ceumass.eps.uspceu.es/mediator/)). Features that did not present any hint in the database were filtered out. Metabolomics data (five replicates/time point and condition) was deposited into the Metabolights repository (https://www.ebi.ac.uk/metabolights/) under accession number MTBLS2166.

Metabolomic pathway analysis and heat map

The mean relative abundances of all identifiable metabolites for each time point for both treated and untreated flies were divided by the mean abundance of the metabolite at the corresponding time point from untreated flies. Results were expressed in fold change values for each individual metabolite, with the untreated group serving as the baseline. Thus, metabolites with abundance fold changes greater than 1 were said to be up-regulated in flies treated with NTBC as compared to control, while metabolites with abundance fold change less than 1 were said to be down-regulated.

The abundance fold-change values were then plotted onto a heatmap using gplot's heatmap.2 program (version 3.0.1.1, http://cran.r-project.org). The heatmap colour scale was limited from 0 to 4, with additional non-cosmetic options as follows: distfun = function(x) dist(x, "manhattan"), hclustfun = function(x) hclust(x, "centroid"), Colv = FALSE.

Subsequent to this, metabolic pathways were assembled using Pathway Collages from BioCyc [68,69], focusing on pathways relevant to tyrosine degradation, carbon metabolism, and amino acid biosynthesis. Fold changes from all identified metabolites were overlaid onto this “combined” metabolic pathway, and statistically significant metabolites (as determined by the 2-way ANOVA test as described previously) were highlighted. Metabolites that were identified from LC-MS/MS but were not identifiable within the “combined” metabolic pathway were not mapped. Pathways that did not have any LC-MS/MS-identified metabolites mapped to it were removed from the “combined” pathway.

In vitro comparative metabolism of NTBC against insect microsomal preparations

Tsetse microsomes

Virgin female G. m. morsitans (n = 174) were maintained on three bloodmeals over the week and then starved for six days to reduce bloodmeal contaminants during tissue homogenisation. The intact thorax and abdomen were removed by a pair of forceps, snap frozen in liquid nitrogen, and homogenised as described with mosquito microsomes. Because of the low P450 content in insecticide-susceptible tsetse flies and mosquitoes, it was not possible to determine P450 content by the traditional carbon monoxide-reduced spectral assay [74], thus protein concentrations were adjusted to 4 mg/ml using 0.1 M Potassium Phosphate buffer (pH 7.4) to normalise for protein content. Recombinant An. gambiae CYP6P3 enzyme was used as a positive control for P450s microsomal activities. All microsome preparations, in comparison to CYP6P3, were tested for O-dealkylation activities against the P450’s generic substrate “Diethoxyfluorescein” before the setup of the NTCB metabolism assay that identifies P450s enzyme activities in microsomal extracts. Briefly, for activity assays, 200 μl of reaction media were set up in triplicate containing 0.5 mM NADPH, 5 μM diethoxyfluorescein, 4 mg/ml microsomes or with mixture of 0.1 μM CYP6P3, 0.8 μM cytochrome b5 and incubated at 30°C for 15 minutes. The enzyme activity was recorded versus negative control (no NADPH) as 147.93 ± 32, 408.64 ± 45.96, 51.5 ± 2.63, and 372 ± 9.8 relative fluorescence units (RFU) for tsetse, Ae. aegypti, An. gambiae, and CYP6P3, respectively. This showed the suitability of microsomes, CYP6P3, and NADPH for running the NTBC comparative metabolism.

NTBC stability under environmental stressors

NTBC was diluted in 10% (w/v) sucrose to a final concentration of 0.01 mg/ml and stored at either 4°C or 25°C in dark or transparent 1.5 ml polypropylene tubes. One volume of NTBC was then mixed with 9 volumes of sterile defibrinated horse blood (final NTBC concentration: 0.001 mg/ml), and these bloodmeals were used to weekly screen for tsetse lethality throughout a 5-week period.

Additionally, NTBC was diluted in 10% (w/v) sucrose to a final concentration of 0.01 mg/ml and subjected to 10 freeze–thaw cycles. Following temperature stress, one NTBC volume was mixed with nine volumes of sterile defibrinated horse blood (final concentration: 0.001 mg/ml), and these bloodmeals were fed to male tsetse. Mortality was monitored over 24 hours.

Mathematical modelling

A discrete time (one-day time step) compartmental model was constructed to describe the key processes underlying the transmission of African trypanosomiasis. These parasites have a broad reservoir of host animals, so these were categorised according to whether they were livestock or wildlife. The transmission between three host types in total (humans, livestock, and wildlife), and the tsetse fly vector was simulated for Trypanosoma brucei rhodesiense (S1 Methods). For T. b. gambiense, animal reservoirs likely play a reduced role, so simulations focused on transmission between tsetse flies and humans only. In both cases, the pre-intervention R₀ was assumed to have a value of 1.1 [75,76]. The impact of vector biting behaviour was examined by apportioning the split of bites on humans, livestock, and wildlife randomly across the full possible range (500 iterations) and, through simulation, determining the frequency of NTBC application required to drive the effective reproduction number (Re) below unity. See S1 Methods for further details about the mathematical modelling.

Statistical analysis

Tsetse survival was scored daily post-treatment. Statistical analysis and graphics were performed using Prism 6.0 software (GraphPad Software, San Diego, United States). The data from multiple experiments were combined into a single graph. The log-rank (Kaplan–Meier) test was used to evaluate significant differences in survival between the experimental and control groups. Dose-response curves were completed using a plot of nonlinear fit for log of inhibitor versus normalised response (variable slope). LD₅₀ and LC₅₀ were calculated using Probit analysis (POLO Plus version 2.0). Numerical source data underlying all figures can be found in S1 Data.