Identification of Sialidase-Producing Bacteria in the Tsetse Gut and Characterisation of a Paenibacillus Sialidase: A Potential Tool for Paratransgenic Control of African Trypanosomes

TL;DR

This study identifies bacteria in tsetse flies that produce sialidase enzymes, which could help control African trypanosomes by interfering with parasite survival.

Contribution

The discovery of a Paenibacillus sialidase with dual pH optima and potential for paratransgenic control of trypanosomes is novel.

Findings

Sialidase activity was detected only in tsetse samples from Cameroon.

A Paenibacillus sialidase gene was identified, cloned, and characterized with specific enzymatic properties.

The sialidase enzyme shows potential as a tool for paratransgenic control of trypanosome colonization.

Abstract

The gut microbiota of Tsetse influences several aspects of the host physiology and vector competence. Trypanosome survival in the tsetse midgut depends on surface sialylation. We hypothesised that bacterial sialidases in the tsetse gut could interfere with parasite transmission by counteracting trypanosomal trans-sialidase activity. This study aimed to detect sialidase activity within the tsetse gut, isolate sialidase-producing bacterial strains, and to characterise the enzymatic properties of bacterial sialidases. Tsetse collected from five African countries (Cameroon, Chad, Ethiopia, Ghana and Nigeria) were screened for sialidase activity using a newly developed field-compatible assay. Detectable activity was found only in samples from Cameroon. Bacterial isolates from these samples, representing several genera (Bacillus, Stenotrophomonas, Pseudomonas, Paenibacillus, Enterobacter), were screened for sialidase production. Whole-genome sequencing identified a sialidase gene in a Paenibacillus species. The gene was cloned, expressed in E. coli, and the recombinant enzyme was purified and biochemically characterised. The sialidase exhibited a molecular weight of approximately 120 kDa, showed dual pH optima at pH 5 and 7. Enzymatic activity was enhanced by Cu2+ ions, and kinetic analysis revealed a vmax of 2.514 nmol/mg/min and a KM of 332.7 μM for the synthetic substrate 4-methylumbelliferyl-N-acetylneuraminic acid. These findings indicate that sialidase-producing Paenibacillus and its enzyme may offer promising tools for paratransgenic approaches aimed at limiting trypanosome colonisation in tsetse flies.

Figures

B](#F1)) was used for rapid detection of sialidase harbouring candidates before inoculation. No clear indication for sialidase acitivity was observed in the field, resulting in inoculation of 10-fold dilution series of all 57 gut samples. Under laboratory conditions, all samples showed a high concentration of free SA, including the substrate control, giving evidence for a challenge with substrate stability ([Fig. 1C](#F1)). However, when looking at the MU-mediated fluorescence, thirteen samples with a fluorescence increase higher than 30-fold over autofluorescence control were considered to be

A](#F3), Fig. S3). Substrate dependent MU-NANA cleavage is shown in [Fig. 3B](#F3).

A](#F8)), the first 31 amino acids represented the signal peptide while the remaining amino acids (32–917) are extracellular. No transmembrane portion was detected. The [Fig. 8B](#F8) represents the predicted 3D model of the *Paenibacillus* sialidase.

Full text

Introduction

Tsetse flies are the primary vectors of African trypanosomes, the causative agents of human African trypanosomiasis (HAT, or sleeping sickness) and African animal trypanosomosis (AAT). The disease remain endemic in 36 sub-Saharan Africa countries, with approximately 56 million people estimated to be at varying levels of risk for HAT^1^. AAT imposes significant socioeconomic burdens on livestock-dependent communities across the African continent^2^. The emergence of resistance to trypanocidal drugs^3–5^ underscores the urgency for novel vector- and pathogen-targeted control strategies.

Vector control remains a cornerstone of disease control strategies and trypanosomiasis management^6^. Upon ingestion of an infected blood meal, trypanosomes encounter the tsetse gut microbiome, which includes both obligate and facultative symbionts such as Wigglesworthia, Sodalis, Wolbachia and Spiroplasma^7–9^. In addition to these endosymbionts, the tsetse gut also harbors environmentally acquired bacteria, whose composition varies with tsetse species and geographic origin of the flies^10,11^. These microbial communities influence vector competence, either directly through bioactive metabolites or indirectly by modulating host immune responses, that in turn can clear the pathogenic microbes^12^, as demonstrated for Kosakonia cowanii Zambiae^13^.

To establish infection in the vector, trypanosomes must first colonise the tsetse midgut^14^, where, they differentiate into procyclic forms. These forms express trans-sialidases (TS)^15–19^, enzymes that catalyze the transfer of terminal linked sialic acids (SA) from host-derived glycoconjugates to the parasite’s surface, forming a new α2,3-linkage^16,18,20,21^. Trypanosomes utilise TS activity to cover themselves with SA that they scavenge from the blood meal^15,16,22^, generating a negatively charged coat believed to protect the parasite from against digestive enzymes and immune factors. TS-mediated surface sialylation is essential for midgut colonisation; loss of sialylation severely impaires parasite survival (Fig. S1)^22^. In contrast, sialidases (neuraminidases) hydrolyse terminal SA, releasing the monosaccharide, which cannot be utilised by TS^16,18^.

Sialidase activity has been detected in the midguts of various hematophagous insects^23^ and multiple bacteria taxa^24^. Previous studies have identified bacteria within the tsetse gut microbiome that are taxonomically related to known sialidase producers, including Arthrobacter sp., Corynebacterium sp., Pseudomonas sp. and Salmonella sp.^24–29^. In bacteria, sialidases serve diverse physiological and pathogenic roles, including nutrition acquisition, host colonisation, immunomodulation, and promotion of virulence^30^. In their nutritional function, they enable the bacterium to procure SA from the host for use as carbon and energy sources^31^, while as virulence factors during bacterial pathogenesis, they can participate to colonization, toxin binding, and immunomodulatory actions^30,32,33^. In AAT caused by Trypanosoma congolense, sialidases play a key role in infection and anaemia^34,35^.

Many bacterial sialidases are secreted enzymes containing signal peptides, conserved sialidase catalytic domains, and accessory regions for substrate binding or localization^24^. Despite low overall sequence homology (<30%), sialidase domains share characteristic motifs such as multiple aspartate boxes^24,36^.

Bacterial sialidases vary in their molecular mass (40–150 kDa), optimal pH (commonly pH 4–7)^24,33,37^, substrate specificity, and kinetic properties. Each enzyme displays distinct regioselectivity and affinity for sialoglycoconjugate, independent of its structural features or cellular localisation^24^.

We hypothesised that bacterial sialidases in the tsetse gut may counteract trypanosomal TS activity by competing for sialylated glycoconjugates from the blood meal, the sole source of SA in the fly gut. By removing terminal SA, sialidases not only reduce donor substrate availability for TS but also increase the pool of exposed β-galactose residues on blood glycoconjugates, which serve as competing acceptors for TS activity.

If sialidase activity exceeds that of TS, it may prevent effective trypanosome sialylation and thereby impair midgut colonisation (Fig. S1).

However, most characterised bacterial sialidases exhibit optimal activity under acidic conditions and are largely inactive at the alkaline pH of the tsetse midgut^38^, limiting their applicability in paratransgenic approaches. In this study, we developed a field-compatible sialidase assay to screen tsetse gut samples for enzymatic activity. We identified and isolated sialidase-positive bacterial strains, cloned and expressed a candidate sialidase gene, and characterised its biochemical properties to evaluate its potential as a paratransgenic tool to interfere with trypanosome development in the vector.

Results

Optimisation of the sialidase assay

The on-field sialidase assay was optimised during several field collections. During the first large scale application of the assay in various sampling sites in Nigeria, controls indicated challenges with substrate and SA stability (Supplementary Information). As a consequence, only lyophilised samples were used in all assays and the free sialic acid control was included to monitor stability of the released SA.

The adjusted assay was further applied in field surveys, repeatedly in Nigeria, Cameroon and Chad and in single collections in Ghana and Ethiopia.

Indications for sialidase activity in tsetse gut

During the application of the sialidase assay in Dodeo, Cameroon, in 2016, 41 tsetse were collected and screened for sialidase activity. Sialidase activity was detected qualitatively by a simple filter spot test (Fig. 1B). Fluorescence was clearly visible in the sialidase positive samples. Adjusting the pH to 10.0 increased the signal. Of these 41 flies, six (14.63%) showed a high increase in MU mediated fluorescence, 100-fold higher than observed during previous collections. In the same samples, high levels of free SA were detected (Fig. 1A) that corresponded to a substrate turnover of 27 to 34%. Free SA present in gut samples without addition of SA donors were below detection threshold (below 1 μM) in all cases except one, sample 1055. A recent blood meal had been recorded in 1055, and, interestingly, this specimen also showed sialidase activity upon substrate incubation. Initial free SA corresponded to 2.4 μM and indicated sialidase activity also prior addition of substrate.

Several samples had a reduced free SA recovery rate. Nineteen of the 41 samples had a recovery rate below 75%, among these, 10 samples with a recovery rate of around 50 %. Only one of the samples indicating sialidase activity had a recovery rate below 75% (Fig. 1A, red). The free SA control had a recovery rate of 96%, confirming the stability of MU-NANA under the experimental conditions in the field. Overall, these results strongly indicate that sialidase activity was present in a subset of tsetse within the population collected in Dodeo.

Field sialidase assay and bacteria isolation

Dodeo sampling site was revisited 12 months later to isolate gut bacteria from tsetse populations. 57 live flies were caught and dissected. The filter spot test (Fig. 1B) was used for rapid detection of sialidase harbouring candidates before inoculation. No clear indication for sialidase acitivity was observed in the field, resulting in inoculation of 10-fold dilution series of all 57 gut samples. Under laboratory conditions, all samples showed a high concentration of free SA, including the substrate control, giving evidence for a challenge with substrate stability (Fig. 1C). However, when looking at the MU-mediated fluorescence, thirteen samples with a fluorescence increase higher than 30-fold over autofluorescence control were considered to be potential candidates for sialidase harbouring bacteria. These were streaked out to obtain single clones for further analysis.

All isolated bacteria were genotyped by 16S rRNA amplification. We identified members of the phylum Firmicutes and Gammaproteobacteria (Fig. 2). Most prevalent representative of the Firmicutes was Paenibacillus that was closely related to Paenibacillus lautus. Obtained 16S sequences were not 100% identical, but between 99 and 99.8%. Bacterial isolates other than those related to Paenibacillus lautus were isolated from 7 out of the 13 gut inoculates investigated (Table 1).

In some selected samples we identified different Paenibacillus, related to Paenibacillus lactis. This bacterial cluster well separated from the Paenibacillus lautus isolates, indicate that they are indeed fly sample-specific bacterial isolates (Fig. 2).

Table 1 gives an overview about the identified trypanosome species and isolated bacteria from respective samples. The genera identified were Bacillus, Enterobacter, Paenibacillus, Pseudomonas, and Stenotrophomonas.

Liquid sialidase assay

Whole bacteria cultures, culture supernatants and bacteria lysates were assayed for sialidase activity using MU-NANA as substrate. The overall turnover detected was low, between 0.5 and 6 % of the total available substrate. In several samples, addition of CaCl_2_ seemed to enhance sialidase activity in the culture supernatant samples (Fig. 3A, Fig. S3). Substrate dependent MU-NANA cleavage is shown in Fig. 3B.

Isolates 1 and 16 were both closely related to Paenibacillus lautus, while isolate 19 was more closely related to Paenibacillus lactis. Interestingly, isolate 16 indicated the presence of sialidase activity in all screening approaches, showing maximum substrate turnover of 6.8 % and enhanced cleavage when supplementing CaCl_2_. This isolate was obtained from the 10^−5^ dilution of fly 1152. From the same fly, two other Paenibacillus lautus relatives were cultured (isolates 37 and 43), that did not show robust indication for sialidase activity. These were isolated from a different dilution (10^−2^). Comparison of the 16S rRNA sequences revealed one base pair difference between the three isolates.

These findings indicate presence of sialidase producing bacteria among the isolated cultures. Isolate 16 (Paenibacillus sp.) was selected as a promising candidate to pursue for further genetic and enzymatic characterisations.

Paenibacillus sp. (isolate 16) whole genome sequencing

The whole genome sequencing results shown the isolate 16 (Paenibacillus sp.) to have a genome size of 8,067,571 bp (DDBJ/ENA/GenBank accession number JBNDYG000000000). From the annotated genome, the putative sialidase gene was found having a size of 2754 bp, making up to 917 amino-acids residues. The sialidase complete gene sequence have been deposited in the NCBI database (GenBank) under accession number PV695982.

Cloning & expression of the Paenibacillus sp. sialidase gene

The sialidase gene was successfully cloned (Fig. S4) and expressed. Western blot confirmed the purity and the integrity of the protein with the presence of the 2 tags: His and Strep (Fig. 4). A molecular weight of about 120 kDa was noted on native SDS-PAGE.

Biochemical characterization of sialidase

The assay with 5 mM of different divalent ions and salts (NaCl, KCl) for their effect on the recombinant sialidase showed that CaCl_2_, MnCl_2_, MgSO_4_, NaCl and KCl slightly enhanced the activity of the enzyme; while the ZnCl_2_ had no effect. Surprisingly, Cu_2_SO_4_ significantly increased the activity of the enzyme (Fig. 5).

The enzyme displayed an acidic to neutral property, with highest activity observed at pH 5 and at pH 7 (Fig. 6). The activity declined in the alkaline side and was zero at pH 9. This is surprising, as the tsetse gut from which this isolate was obtained showed sialidase activity at pH9. The recombinant sialidase efficiently worked at temperatures ranging from 30 °C to 37 °C.

Initial-velocity determinations of the recombinant Sialidase were carried out over wide concentration ranges (0 to 500 μM) and 25 ng/μL of enzyme. According to nonlinear regression on GraphPad Prism 9 (Fig. 7) and transformed into the Lineweaver-Burk plot (Fig. 7 inset) for calculation of the kinetic parameters Vmax and KM. The Vmax was 2.5 nmol/min/mg and the KM, 332.7 μM.

Gene sequence analysis

From the prediction of the sialidase topology (Fig. 8A), the first 31 amino acids represented the signal peptide while the remaining amino acids (32–917) are extracellular. No transmembrane portion was detected. The Fig. 8B represents the predicted 3D model of the Paenibacillus sialidase.

The evolutionary history of the isolated sialidase amino-acid sequence with other Paenibacillus sialidase presented on the phylogenetic tree (Fig. 9) showed that its amino-acid sequence is similar to the one from Paenibacillus sp. GM2FR (WP100540498.1)

Discussion

Trypanosomes, causative agents of human and animal Trypanosomiasis, have been shown to rely on a surface coat of SA when establishing an infection in the tsetse gut^22^. They express trans-sialidases, enzymes that transfer SA from one sialoglycoconjugate to their own surface glycans. Sialidases can counteract this activity, releasing bound SA and thus rendering it useless for trans-sialidases, unmasking the trypanosome surface.

Sialidase activity had so far not been detected in tsetse and its known endosymbionts. Therefore, any potential sialidase activity detected in our study would be due to environmentally acquired bacteria.

Field assays as developed during this study are prone to disturbances, they encounter uncontrollable conditions that could not all be anticipated and modelled in the laboratory. In the course of this study, we therefore altered the protocol by using lyophilised substrates and included additional controls. The adjusted assay protocol proved to be robust in several surveys, although some problems with substrate stability were encountered again during the last application. The optimised protocol gave indications for sialidase activity in several specimens from a fly population in Dodeo, Cameroon. Only one of these flies contained trypanosomal DNA, originating from T. grayi. Although TS-like genes can be found in the genome of T. grayi^42,43^, it is not clear whether these enzymes are active and expressed in the different life stages of the parasite. However, other samples carrying the same trypanosome did not show enhanced free SA concentrations or enhanced fluorescence, therefore, it is unlikely that Trypanosoma TS is the origin for the potential sialidase activity detected.

Bacteria colonising the tsetse gut, apart from the common endosymbionts, have to be acquired from the environment. Blood, the main food source for tsetse, is sterile. Bacteria diversity in the tsetse gut might therefore result from feeding on a septic host or by taking up bacteria from the hosts’ skin during feeding, which has been shown to occur in experimental setting with bacteria smears on rabbit ears^44^. Tsetse have also been shown to occasionally feed on water or plant nectar^45^, which can harbour a high bacterial diversity.

With our culture-dependent approach to investigate tsetse from Dodeo for sialidase producing bacteria, we isolated members of the Firmicutes and Proteobacteria, including Bacillus and Paenibacillus, and Enterobacter, Stenotrophomonas and Pseudomonas. A Bacillus sp. was identified in one fly sample. Though closely related to Bacillus circulans, it was not possible to distinguish between B. circulans, B. thuringiensis and B. anthracis based on 16S rRNA sequences. However, due to the pre-screens for B. anthracis that turned out negative for all samples (data not shown), this very pathogenic species can be excluded.

The bacterial diversity identified in our study is lower than in previous studies based on culture-dependent identification of gut bacteria^25,27,28^. Cultivating under anaerobic conditions was found to enhance the number of cultivable genera from Glossina pallidipes (15 versus 5 under aerobic conditions)^28^. However, aerobic cultures revealed additional genera that were not isolated under anaerobic conditions, indicating that both culturing approaches should be applied in parallel to obtain a more representative picture of cultivable bacteria diversity. Other surveys in Southern Cameroon revealed a higher bacterial diversity, both with culture dependent and deep-sequencing methods. The diversity was found to vary from specimen to specimen regardless of geographic origin^25,29,46^. Therefore, the combination of culture-dependent and independent approaches might be useful to assess the true bacteria diversity within the samples.

Interestingly, the flies indicating sialidase activity in our study were collected geographically very close to each other. Four flies were from the same trap, another from a trap 82 m apart. The remaining fly with sialidase activity was collected roughly 520 m from the latter. This might indicate that the potential sialidase producing bacteria are restricted to these locations, and might not be encountered by flies in other sites. Sialidase activity was not recorded during surveys in the other sites.

Sialidase activity has been identified in several bacteria genera^24^. The isolated Paenibacillus (Isolate 16) showed sialidase activity under various conditions, and whole genome sequencing indeed predicted a sialidase gene in this isolate. The nucleotide sequence was not closely related to any other reported sequence in the NCBI database, only two similar nucleotide sequences were found: 89.58% identity with Paenibacillus lautus strain E7593–69 (CP032412.1) and 81.7% with Paenibacillus ihbetae strain IHBB 9852 (CP016809.1). The BLAST search with the amino-acid sequence found one closely related (100 % identity) Paenibacillus sp. GMR2, isolated from the plant Festuca ruba (WP100540498.1). The size of the protein is with ~100 kDa (917 amino acids) particularly larger than that of many other bacterial sialidases which average size is 60 kDa^24^. Among Paenibacillus, sialidase activity was described for one isolate from soil in Japan. Other Paenibacillus isolates tested did not show sialidase activity^47^. Interestingly, Paenibacillus was found infecting humans globally in an opportunistic manner^48^. Paenibacillus lautus and P. lactis have previously been isolated from human blood in Spain^49^ and P. lautus was isolated from ticks feeding on rats in Malaysia^50^. This indicates that these bacteria species are widely distributed and can circulate in the bloodstream of mammals and thus could be encountered by tsetse during feeding. Additionally, several Paenibacillus species have been shown to produce antimicrobial peptides^48,51^, which is an additional interesting feature to monitor with respect to the effect of these compounds on trypanosome infections.

To investigate the enzymatic activity of the encoded protein, the entire sialidase gene of our isolated Paenibacillus sp. was cloned an expressed in Escherichia coli strain Rosetta. A signal peptide for protein excretion was predicted, therefore it was not surprising to detect large part of the expressed protein in the pellet where denaturation/renaturation steps are required before obtaining a functional protein. Surprisingly, an important fraction was also directly available in the soluble fraction. This implies that the protein could also be a peripheral membrane protein. Although the prediction showed a signal peptide, there was no transmembrane portion. Therefore, the protein could easily get detached from the membrane and get available in the soluble fraction. The sialidase activity on MU-NANA in the presence of metal ions was found to vary in different species^37,52^. It was observed that the Paenibacillus sp. clone 16 sialidase activity was slightly enhanced in the presence of some tested metal ions, but the highest was surprisingly with CuSO_4_ which is rarely found to be cofactor of sialidase. It can be considered as the best co-factor for the recombinant sialidase.

The protein was found to be active from pH 4.0 to pH 8.5. Surprisingly, the pH curve showed two peaks (at pH 5.0 and pH 7.0), thus the existence of two optimum pH of the enzyme. Bacterial sialidases are known to having their maximum activity at acidic pH usually from 4.5 to 5.5^37,52,53^. A sialidase from Sphingobacterium sp. strain HMA12 was found to have its optimal pH at neutral ranges, pH 6.5–7.0^54^. The double peaks may be due to the presence of two catalytic domains in the protein: one working at acidic pH while another is working at neutral pH. This situation has been previously described with other proteins^55^. The working temperature of 30°C to 37 °C is similar to that of many other^52^. In previous studies, sialidases have been shown to be functional in a large temperature range, retaining activity up to 50 °C^56^.

The maximum velocity of 2.5 nmol/min/mg obtained in this study was lower compared to that of other sialidases tested with MU-NANA: 706 μmol/h/mg, 1.58 μmol/h/mg and 2.39 μmol/h/mg respectively for T. congolense sialidase, zebrafish Neu3.2 and Neu3.3^57,58^. However, this velocity was higher than 25 nmol/h/mg and 5 nmol/h/mg obtained from tilapia sialidase Neu1a and Neu1b respectively^53^. The estimated KM of 332.7 μM was higher compared to 25 μM of T. congolense sialidase^58^. These results suggest that the MU-NANA may not be the ideal substrate for our enzyme. Other substrates should be tested.

Conclusion

We developed a protocol for a sialidase assay that allows detection of sialidase activity in the fly gut and preserves the DNA integrity on the level of the individual fly. Using this assay, we have successfully identified several potential candidates, isolated a sialidase-expressing Paenibacillus sp, and characterised the sialidase activity of the recombinant enzyme. The enzyme showed particular features that are not common in other bacterial sialidases.

Little is known about the multifactorial interplay between tsetse, the obligate and secondary endosymbionts, microbiome and trypanosomes. The sialidase-producing Paenibacillus sp. and the sialidase enzyme isolated in this study are promising candidates of investigation towards biological control of trypanosomes transmission by tsetse flies. This includes investigating their influence on trypanosome infections, tsetse fly longevity and suitability for paratransgenesis approaches.

Material and methods

Field collection

Tsetse flies were collected during five trapping surveys conducted in Cameroon, Chad, Ethiopia, Ghana and Nigeria, using biconical traps. Detailed descriptions of sample collection site and protocols have been described previously^39,59–61^.

Dissection of tsetse flies and assay preparation

Live tsetse flies were dissected under semi-sterile conditions. The gut was excised and homogenised in 200 μl sterile 50 mM Tris-HCl buffer (pH 9.0). For nucleic acid analysis (Fig. 10.1), 50 μL of the homogenate was transferred to 450 μL of nucleic acid preservation agent (NAPA: 25 mM sodium citrate, 10 mM EDTA, 70 g ammonium sulfate/100 ml, pH 7.5). The remaining homogenate was used for sialidase assays (2–4), and bacterial inoculation (5). Except for bacteria inoculations, all samples were kept cool (on ice) or stored at −20°C during field work and at −80°C for long-term storage.

Field sialidase assay

To detect sialidase activity in single tsetse gut samples under field conditions (with limited infrastructure and cooling), assay components were lyopholised in 50 mM Tris/Cl pH 9 for transportation.

To determine background free SA from recent blood meals, 40 μL homogenate was precipitated with 4 volumes of acetone (Fig. 10.3), no substrate control). To assess SA stability and potential enzymatic degradation by sialic acid lyases, 40 μL homogenate was incubated with 100 μM free SA (Fig. 10.4). To detect sialidase activity, 40 μL homogenate was incubated with 80 μg fetuin (0.8 – 0.9 mM fetuin-bound SA) and 1 mM 4-Methylumbelliferyl-N-acetylneuraminic acid (MU-NANA) as substrate (Fig. 10.2). All reactions were incubated for one hour at ambient field temperatures (25–38 °C) and terminated by acetone precipitation.

Qualitative assessment of sialidase activity in field samples was performed using a simple filter spot approach. 1 μL of the acetone supernatant (10x diluted original sialidase assay) was spotted onto a Whatman filter. MU-fluorescence was visualised at pH 9.0/10.0^62^ by adding 2 μl of 1 M glycine stock solution (pH 10) and exposing to blue light (395–400 nm). Positive samples were selected for bacterial isolation.

Bacteria inoculation under field conditions

For isolation of tsetse gut bacteria (Fig. 10.5), dissections were performed using a semi-sterile cabinet (detailed in supplementary information). Tsetse flies were surface-sterilised with 70% ethanol.

As gut bacteria are exposed to microaerobic conditions, we aimed to isolate facultative anaerobic bacteria. Exactly 30 μL of homogenate were added to 150 μL sterile MMI (Maramorosch and Mitsuhashi Insect medium) buffered with 10 mM sodium phosphate (pH 7.0). Samples showing fluorescence in the sialidase filter spot assay were inoculated into microaerobic cultures.

Serial dilutions (10^0^-10^−10^) were prepared in sterile MMI, and 20 μL of each dilution was injected with a sterile syringe into 2 mL tubes containing 1.9 mL 0.3 % Bactoagar (Sigma Aldrich, Munich, Germany)/ MMI supplemented with 20 % foetal calf serum (FCS). Inoculation tubes were prepared by solidifying the MMI-agar upside down at room temperature, to allow low-oxygen culture injection into the medium with low oxygen concentration.

Sterility controls included uninoculated MMI and MMI in which a surface-sterilised fly was immersed. All cultures were incubated upside down under field conditions and stored at 4 °C prior to laboratory analysis.

Fluorescence-Based Detection of MU Release (sialidase activity)

Acetone-precipitated samples were centrifuged at 14,000 rpm for 20 minutes at 4°C to remove precipitated proteins. The resulting supernatants were dried and subsequently reconstituted in 400 μL of double-distilled water (ddH_2_O), followed by sonication (three cycles of 10 seconds with 20-second intervals) to ensure complete dissolution of released methylumbelliferone (MU). Reconstituted samples were stored at −80°C until further analysis. For fluorescence-based detection of sialidase activity, 40 μL of the dissolved sample was transferred into a black 96-well microtiter plate (Thermo Fisher Scientific) and mixed with 100 μL of 1 M glycine buffer (pH 10.0). Fluorescence intensity, indicative of MU release, was measured using a Tecan Infinite 200 Pro microplate reader (Tecan Life Sciences, Switzerland) at an excitation wavelength of 360 nm and an emission wavelength of 465 nm. To account for potential interference from residual blood components, control reactions without added sialidase substrates were included for each sample.

Detection of free sialic acid by RP-HPLC

α-Keto acids react specifically with phenyldiamines to form quinoxalines. This derivatization can be used for the selective separation and detection of SA via reverse phase high performance liquid chromatography (RP-HPLC) coupled to fluorescence detection. Free SA was derivatised using 1,2-diamino-4,5-methylendioxybenzene (DMB) or 4,5-dimethylbenzene-1,2-diamine (DMBA) and analysed via RP-HPLC. Concentrations of free SA in fly samples were determined following adaptations of Hara et. al.^63^ and Wang et al.^64^. Detailed protocols are provided in the Supplementary Information.

Detection of Trypanosoma sp.

As Trypanosoma TS possess hydrolytic activity in the absence of suitable SA acceptors, all gut samples were screened for trypanosomal DNA to exclude TS-derived activity. DNA extracted from NAPA-preserved gut tissue was used in internal transcribed spacer 1 (ITS1)-based diagnostic PCR for Trypanosoma spp.^39,65^. Results were previously published^39,66–68^. (Detailed in supplementary information).

Bacteria idendification and 16S rRNA sequencing

The presence of Bacillus anthracis in all samples was first excluded by a pre-screens in selective media: Cotrimoxazol- blood agar (selective for Bacillus anthracis) and MYP-agar or BACARA^®^ (selective for Bacillus cereus)^69^.

Single cultures were isolated by repeated streaking on MMI agar plates supplemented with FCS. Initial streaks were done in presence of 20% FCS while sub-streaks were done in 10% FCS. Plates were cultured at 30 °C under anaerobic conditions generated by Oxoid AnaeroGen sachets (Thermo Scientific, Dreieich, Germany) until colony growth was visible.

DNA was crudely extracted from bacterial isolates by bead-beating with 0.75–1 mm glass beads (detailed in Supplementary information).

The bacterial 16S rRNA gene was amplified with generic primers according to Grönemeyer et al.^70^. Amplicons were sequenced by a commercial company (Microsynth SeqLab, Göttingen, Germany). Sequences were compared using BLAST search in the NCBI database to identify closest relatives. Phylogenetic relationships with other reference sequences were inferred (detailed in Supplementary information).

Sialidase assay of liquid bacterial cultures

To detect sialidase activity, bacterial isolates were cultured in 15 mL tubes containing 8 mL liquid MMI medium supplemented with 20% FCS for 5 days at room temperature (20°C). Whole bacteria cultures, culture supernatants and bacteria lysates were screened for sialidase activity. The lysate was obtained by lysozyme digestion. Samples (100 μL) were investigated for sialidase activity by adding 1 mM MU-NANA with and without additional 4 mM CaCl_2_. A control sample without any supplements was incubated in parallel. Samples were incubated overnight at 30°C in the dark. Standard curves of MU in MMI and lysis buffer were loaded on the same plate for reading. Fluorescence was measured after addition of 100 μL of 1 M glycine pH 10.0 in a black 96-well plates (Detailed description in Supplementary Information).

Purification of bacterial genomic DNA

Genomic DNA was isolated from pure bacterial culture using standard phenol-chloroform method^71^ with few modifications (detailed in Supplementary information). The gDNA concentration and the purity were assessed on NanoDrop spectrophotometer (ND1000, Thermo Scientific, Dreieich, Germany).

Whole genome Sequencing

The genome sequencing was performed at AG Reinhold, Molecular Plant-Bacteria Interactions, University of Bremen, Germany (Data not shown). Gene prediction and annotation was performed using the Rapid Annotation System Technology (RAST) server (https://rast.nmpdr.org).

Cloning

The hypothetical sialidase gene was isolated by PCR amplification using Phusion High-Fidelity DNA Polymerase (ThermoFisher) and the primers designed to include BsphI and EcoRI restriction enzyme recognition sites: BsphI_His_Sia: 5’-tatcatgagccatcaccatcaccatCACATTCCATTTCG-3’ and EcoRI_Sia: 5’-tcgaattcATCGCCATCATGTTCATCCTC-3’. Nucleotides in lower case are extensions containing the recognition sites for restriction enzymes (underlined) and the 6X His-Tag (double underlined). The PCR conditions are detailed in supplementary file. The amplicons were purified using the GeneJET Gel Extraction Kit (ThermoScientific) according to the manufacturer instructions.

The sialidase was cloned into a modified pET28a+ expression vector containing a N-terminal His and a C-terminal Strep Tags^72^. The plasmid and insert were digested with FastDigest BspHI, EcoRI and NcoI (ThermoScientific). Ligation was performed using T4 DNA Ligase (ThermoScientific) and the mixture was then transformed into E. coli Rosetta and plated on LB-agar containing kanamycin. The cloning procedure is detailed in supplementary file.

Recombinant Sialiadase expression

Recombinant protein expression of bacteria inoculates was induced by 50 μM IPTG for 8 hours at 18 – 20 °C. The recombinant sialidase was purified and detected using the Strep and His tags as previously described^21,72^. The final protein concentration was determined by Bicinchoninic acid assay using Pierce^™^ BCA Protein Assay Kit according to manufacturer instructions.

SDS-PAGE and Western Blot of the recombinant sialidase

Bacterial lysate, and eluted fractions of Ni-NTA, Strep-Tactin and Sephadex G-25 columns were subjected to SDS-PAGE. The stacking and separating gels were 4% and 8% (w/v) acrylamide, respectively, 10 μL sample were loaded per lane. The gel was run at 15 A for 1 hour and stained with PAGE-Blue stain (Thermo Scientific, Germany).

Western Blot was performed to ensure expression of the whole recombinant protein. Blotting was done using 100 V for 1 h. The membrane was blocked for 1 h with 3% BSA in TBST (Tris-buffered saline + 0.15 % Tween 20). The primary antibodies used were sheep anti-Strep-tag or mouse anti-His-tag antibody diluted 1:1000 in TBST buffer. The secondary antibodies were Peroxidase conjugated Donkey anti-sheep IgG and Donkey anti-mouse (Jackson Immuno Research Europe Ltd, Hamburg, Germany) diluted 1:40000 in TBST + 3% BSA. Visualisation was done using the chemiluminescent reagent and 5 min exposure in an autoradiography film (Kodak).

Recombinant sialidase characterisation

The sialidase activity assay was carried out in triplicate using N-acetyl neuraminic acid-methylumbelliferone (MU-NANA) as substrate, the Arthrobacter ureafaciens neuraminidase as positive control and 10 mM Tris-HCl, pH 7 as negative control.

Exactly 20 μL reaction mixture consisting of 100 μM of MU-NANA, 5 mM of divalent cation, 37.5 mM of a buffer, and 1.7 μg of purified enzyme (in 5 μL of 10 mM Tris-HCl, pH 7) was incubated at 37°C for 1 hour in the dark. After incubation, the reaction was terminated by the addition of 40 μL glycine/NaOH buffer (pH 9). Exactly 20 μL of the mixture were transferred to a black 384 wells microtiter plate. Fluorescence was measured at excitation wavelength of 360 nm and emission of 465 nm on a Tecan infinite 200pro plate reader (Tecan Life Sciences, Switzerland). The approximate amount of released 4-MU was calculated using the calibration curve prepared from different concentration of commercial 4-MU.

Effect of divalent metal ions and salts

The effect of divalent metal ions was measured by using 5 mM final concentration of CaCl_2_, CuSO_4_, MgCl_2_, MnCl_2_, MgSO_4_, ZnCl_2_. The salts were 5 mM final concentration of NaCl and KCl. The enzyme activity in the absence of divalent ions was tested by adding a chelating agent EDTA (5 mM final concentration) and assayed under above-mentioned conditions.

Optimization of pH and Temperature of the recombinant sialidase

To determine the optimal pH, the reaction was carried out using 37.5 mM of different buffer systems: Sodium Acetate buffer (pH 4.0 – 5.0 – 6.0 – 6.4), Tris-HCl (pH 6.5 – 7.0 – 7.7 – 8.0 – 8.5), and Glycine/NaOH (pH 9.0) under above-mentioned conditions.

The activity of the enzyme was tested at 3 different temperatures: Room temperature (20–22 °C), 30 °C, and 37 °C under optimum pH and co-factor condition.

Enzyme kinetics

Kinetic constants of the recombinant sialidase were obtained by measuring initial velocity over wide range concentration of MU-NANA (0, 10, 25, 50, 75, 150, 250, 500 μM). The reaction was performed in 20 μL containing 500 ng of purified enzyme (25 ng/μL) in 37.5 mM Tris-HCl buffer pH 7.0 plus 5 mM CuSO_4_ at 37 °C. A standard curve obtained from the fluorescence signal of 4-MU at defined concentrations was used to quantify the released 4-MU. The Michaelis–Menten kinetics parameters (KM and Vmax) for hydrolysis of MU-NANA by the purified enzyme were determined by GraphPad Prism 9 software using nonlinear regression with different substrate concentrations. The sialidase activity was expressed as nmol 4-MU released min^−1^ mg^−1^ of the sialidase.

Gene sequence analysis

The topology of sialidase was predicted using the online transmembrane predictor, DeepTMHMM server available at: https://services.healthtech.dtu.dk/service.php?DeepTMHMM. The predicted 3D structure was built as previously described in Jeelani et al.^40^ using Phyre 2.2^41^.

The sialidase sequence was subjected to BLAST searches at NCBI data base to identify closest relatives. The MEGA X software^73^ was used to construct the phylogenetic tree. The evolutionary history was inferred using the Neighbor-Joining method, based on the Jones-Taylor-Thornton model with 100 bootstraps replicates.

Supplementary Files

This is a list of supplementary files associated with this preprint. Click to download.

MfopitSupplementaryinfo.pdf