Interactions between Glossina pallidipes salivary gland hypertrophy virus and tsetse endosymbionts in wild tsetse populations

Background Tsetse control is considered an effective and sustainable tactic for the control of cyclically transmitted trypanosomosis in the absence of effective vaccines and inexpensive, effective drugs. The sterile insect technique (SIT) is currently used to eliminate tsetse fly populations in an area-wide integrated pest management (AW-IPM) context in Senegal. For SIT, tsetse mass rearing is a major milestone that associated microbes can influence. Tsetse flies can be infected with microorganisms, including the primary and obligate Wigglesworthia glossinidia, the commensal Sodalis glossinidius, and Wolbachia pipientis. In addition, tsetse populations often carry a pathogenic DNA virus, the Glossina pallidipes salivary gland hypertrophy virus (GpSGHV) that hinders tsetse fertility and fecundity. Interactions between symbionts and pathogens might affect the performance of the insect host. Methods In the present study, we assessed associations of GpSGHV and tsetse endosymbionts under field conditions to decipher the possible bidirectional interactions in different Glossina species. We determined the co-infection pattern of GpSGHV and Wolbachia in natural tsetse populations. We further analyzed the interaction of both Wolbachia and GpSGHV infections with Sodalis and Wigglesworthia density using qPCR. Results The results indicated that the co-infection of GpSGHV and Wolbachia was most prevalent in Glossina austeni and Glossina morsitans morsitans, with an explicit significant negative correlation between GpSGHV and Wigglesworthia density. GpSGHV infection levels > 103.31 seem to be absent when Wolbachia infection is present at high density (> 107.36), suggesting a potential protective role of Wolbachia against GpSGHV. Conclusion The result indicates that Wolbachia infection might interact (with an undefined mechanism) antagonistically with SGHV infection protecting tsetse fly against GpSGHV, and the interactions between the tsetse host and its associated microbes are dynamic and likely species specific; significant differences may exist between laboratory and field conditions. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s13071-022-05536-9.


Introduction
Mutualistic bacteria are functionally essential to the physiological well-being of their animal hosts. They benefit their hosts by providing essential nutrients, aiding in digestion and maintaining intestinal equilibrium. Furthermore, mutualistic symbionts foster the development, differentiation, and proper function of their host's
Flies in the genus Glossina (tsetse flies) are unique to Africa and are of great medical and economic importance as they serve as a vector for the trypanosomes responsible for sleeping sickness in humans (human African trypanosomosis or HAT) and nagana in animals (African animal trypanosomosis or AAT) [31,32]. The presence of tsetse and trypanosomes is considered one of the major challenges to sustainable development in Africa [33,34]. The lack of adequate and affordable vaccines coupled with pathogen resistance to drug treatments severely limits AAT control, leaving vector control as the most feasible option for sustainable management of the disease [31,32]. In addition to various pesticide-and trapping-based methods for tsetse control, the sterile insect technique (SIT) is considered an efficient, sustainable and environmentally friendly method when implemented in the frame of area-wide integrated pest management (AW-IPM) [35,36]. However, the SIT requires the mass rearing of many males to be sterilized with ionizing radiation before release into the targeted area [33,37].
Tsetse fly biology is characterized by its viviparous reproduction rendering tsetse mass rearing a real challenge. Tsetse flies nourish their intrauterine larvae from glandular secretions and give birth to fully developed larvae (obligate adenotrophic viviparity) [38,39]. They also live considerably longer than other vector insects, which somewhat compensates for their slow reproduction rate [40]. The ability to nourish larvae on the milk gland secretion, although limiting the number of larvae produced per female lifetime (8)(9)(10)(11)(12), facilitates the transmition of endosymbitic bacteria and pathogens from females to larvae such as Wigglesworthia, Sodalis, Wolbachia, Spiroplasma, and GpSGHV [8,10,13,41]. Moreover, as strictly hematophagous, tsetse rely on the associated endosymbionts to obtain essential nutrients for female reproduction. Therefore, tsetse well-being in mass rearing for SIT is affected by the status of its endosymbionts as well as infection with pathogenic viruses and the interactions between them. Although Wigglesworthia is an obligate endosymbiont and found in all tsetse species, Sodalis, Wolbachia, and Spiroplasma infection varied from one species to another [8,10,[42][43][44][45]. In addition, infection with GpSGHV, although reported in different tsetse species, is mainly symptomatic in G. pallidipes [46][47][48]. As GpSGHV is horizontally transmitted via the feeding system under laboratory conditions, leading to high infection rates [49][50][51], and the virus has a negative effect on the reproductive system of the host causing reduced fecundity and fertility [52,53], control of the virus infection is important in tsetse mass rearing for efficient production of irradiated males for SIT program implementation.
The variable responses of different tsetse species to the GpSGHV infection might indicate a possibility of the tsetse microbiota modulating the molecular dialogue among the virus, symbiont, and host, shaping the response of each species to the virus infection. It was necessary, therefore, to investigate the infection status of the major tsetse endosymbionts (Wigglesworthia, Sodalis, and Wolbachia) in different tsetse species and their potential interactions. We have recently investigated the interaction between GpSGHV and tsetse symbionts in six tsetse species after virus injection under laboratory conditions [54]. The results indicated that the interaction between the GpSGHV and tsetse symbionts is a complicated process that varies from one tsetse species to another. It is worth noting that the study of Demirbas-Uzel et al. [54] was conducted in tsetse flies maintained under controlled laboratory conditions (sustainable food availability, constant environmental conditions (temperature and humidity), and high density of the flies), which favors the increase of tsetse symbionts [45,55,[55][56][57]. In addition, this study was done using adults artificially infected with GpSGHV by injection. Therefore, we investigated the associations of the GpSGHV and tsetse symbionts in field-collected samples by evaluating the prevalence of co-infection of GpSGHV and Wolbachia and their potential association with Wigglesworthia and Sodalis infection in natural tsetse populations. The results are also discussed in the context of developing an effective and robust mass production system of highquality sterile tsetse flies for implementing SIT programs.

Tsetse samples, extraction of total DNA, and PCR amplifications
The field collection of tsetse fly samples, DNA extraction, and the PCR-based prevalence of GpSGHV and Wolbachia infections were reported previously [7,47,58,59]. Based on these publications, and using G. m. morsitans, G. pallidipes, G. medicorum, G. brevipalpis, and G. austeni samples collected from Burkina Faso, South Africa, Tanzania, Zambia, and Zimbabwe, four infection patterns (i.e. presence) were determined: (i) flies PCR positive for both GpSGHV and Wolbachia (W + /V + ), (ii) flies PCR positive for Wolbachia alone (W + /V − ), (iii) flies PCR positive for GpSGHV alone (W − /V + ), and (iv) flies PCR negative for both GpSGHV and Wolbachia (W − /V − ). It has to be noted that the prevalence of the symbionts was assessed using a conventional PCR assay while their densities (see below) were determined using a qPCR assay. Since these two assays were different in several aspects including the size of the amplicons and visualization process, this resulted in some discrepancies regarding the infections status of some virus samples initially considered virus free by conventional PCR that were found to be positive during the qPCR analysis.

Analysis of the associations among SGHV and Wolbachia, Sodalis, and Wigglesworthia infection in wild tsetse populations
The associations among GpSGHV and Wolbachia, Sodalis, and Wigglesworthia were assessed by qPCR analysis.  Table 1). The qPCR analysis was performed as previously described [47,53,60]. In brief, for the standard curve, total DNA was diluted tenfold before being used for qPCR analysis on a CFX96 real-time PCR detection system (Bio-Rad, Hercules, CA) using the primers and conditions presented in Additional file 2: Table S1. The estimated copy number by qPCR for each sample compared with the standard curve was determined in diluted DNA (4 ng/μl) and corrected through the multiplication by the inverse dilution factor to reflect the GpSGHV, Wolbachia, Wigglesworthia, or Sodalis copy number (hereafter mention as density) per fly. Analysis of the Wolbachia, Wigglesworthia, Sodalis, and SGHV density levels (titers) was based only on qPCR data with the expected melting curves at 85.5-86 °C, 78.5-80 °C, 81.5-82 °C, and 76.5-77 °C, respectively. Data with a melting curve outside the indicated range were excluded from the analysis. The status of Sodalis and Wigglesworthia infection of the samples used for the qPCR analysis was not determined by traditional PCR. Based on the estimated copy number per fly for SGHV, Wolbachia, Wigglesworthia, and Sodalis, the average copy number was calculated for all tested flies. Flies with copy number values less than the median were considered infected with low density and flies with copy number value greater than the median were considered infected at a high level. The median copy numbers of the GpSGHV, Wolbachia, Sodalis, and Wigglesworthia in all tested samples were 10 3.31 , 10 7.36 , 10 6.07 , and 10 6.84 per fly, respectively.

Statistical analysis
The proportion of single and double infections (GpSGHV and Wolbachia) in wild flies was analyzed by location and species and for all samples together using the Chisquared test. The Chi-squared tests for independence, Spearman correlation coefficient, and Cochran-Mantel-Haenszel test for repeated tests of independence were performed using Excel 2010. P-values were calculated from the data with the significance threshold selected as 0.05.
The difference in Wigglesworthia, Sodalis, Wolbachia, and GpSGHV density between different locations and tsetse species and the correlation between densities as well as preparing figures were executed in R v 4.0.5 [61] using RStudio v 1.4.1106 [62,63] with packages ggplot2 v3.3.2.1 [64], lattice v0.20-41 [65], car (version 3.1-0) [66], ggthems (version 4.2.4) [67], and MASS v7.3-51.6 [68]. All regression analyses of symbionts and GpSGHV densities were conducted using the generalized linear model (glm) for different tsetse species and different countries with analysis of deviance table (type II tests). Pearson correlation coefficient between the density of Wolbachia and Wigglesworthia and the log transformed density of GpSGHV and Sodalis was conducted in R. The analysis details are presented in Additional file 1. Overall similarities in Wolbachia, Wigglesworthia, Sodalis, and GpSGHV density levels between tsetse species, countries, and infection pattern were shown using the matrix display and metric multidimensional scaling (mMDS) plot with bootstrap averages in PRIMER version 7 + and were displayed with a Bray and Curtis matrix based on the square root transformation [69]. The tests were based on the

Prevalence of co-infection with GpSGHV and Wolbachia in wild tsetse flies
Analysis of the Wolbachia and GpSGHV infection status for each individual tsetse adult in the previously reported data [7,47,58,59] indicated that the single infection rate was 10.21% (n = 459) and 15.12% (n = 680) for GpSGHV and Wolbachia, respectively, over all taxa and locations combined (Additional file 4: Fig. S1A). No Wolbachia infection was found in two taxa, G. f. fuscipes and G. p. palpalis, and these were excluded from further examination (Table 1). A Cochran-Mantel-Haenszel test for repeated tests of independence showed that infection with GpSGHV and Wolbachia did not deviate from independence across all taxa (χ 2 MH = 0.848, df = 1, n.s.), and individual Chi-squared tests for independence for each taxon did not show any significant deviation from independence at the Bonferroni corrected α = 0.00714 (Additional file 3: Table S2). The prevalence of co-infection of GpSGHV and Wolbachia (W + /V + ) in wild tsetse populations varied based on the taxon and the location (Table 1 and Additional file 4: Fig. S1B). No co-infection was found in G. brevipalpis, G. m. submorsitans, and G. p. gambiensis, and co-infection was absent in many locations in the remaining taxa. However, a low prevalence of co-infection was found in G. medicorum (2%), G. tachinoides (0.7%), and G. pallidipes (0.2%). A relatively high prevalence of co-infection was only observed in G. austeni (26%) and G. m. morsitans (13%) (Additional file 4: Fig. S1B).

Impact of co-infection (W + /V + ) on GpSGHV, Wolbachia, Sodalis, and Wigglesworthia density GpSGHV density
The GpSGHV qPCR data showed overall no statistically significant difference between flies with different infection patterns (W + /V + ), (W − /V + ), and (W + /V − ) (X 2 = 1.4625, df = 2, P = 0.481) regardless of tsetse taxon (Additional file 5: Fig. S2A). Moreover, no significant difference in GpSGHV copy number was observed between tsetse taxa (X 2 = 0.752, df = 3, P = 0.861) (Additional file 4: Fig. S1A). However, a significant difference in the virus copy number was observed between different countries (X 2 = 16.234, df = 4, P = 0.0027) where the virus copy number in the flies collected from Zambia a In these samples, the presence of Wolbachia was tested in Doudoumis et al. [7] b The individuals of G. f. fuscipes were considered negative for Wolbachia based on the results of the initial PCR amplification. The results from the reamplification method were not considered so that the conditions were consistent for all species c Samples used for qPCR analysis to determine the density of Wigglesworthia, Sodalis, Wolbachia, and GpSGHV was significantly lower than those collected from South Africa, Tanzania, and Zimbabwe (Additional file 1 and 6: Fig. S3A).
Overall, a significant difference in Wolbachia density was observed in the flies with different infection patterns previously determined by conventional PCR, where flies with a (W + /V − ) infection pattern showed significantly higher Wolbachia density than flies with a (W + /V + ) infection pattern regardless of the tsetse species (X 2 = 23.723, df = 2, P < < 0.001). This trend was observed in G. m. morsitans (t = 3.184, P = 0.0022) (Additional file 1). The Wolbachia density was highest in the flies collected from Zambia (Additional file 6: Fig. S3B). Analyzing only the flies with co-infection (W + /V + ) indicated that the Wolbachia density was statistically significantly higher in G. m. morsitans than in G.austeni (t = − 2.353, df = 1, P = 0.024) (Additional file 1 and 7: Fig. S4B).
The qPCR results showed that Wolbachia-infected flies had relatively high Wolbachia density (median 10 7.3 copies/fly) compared to the GpSGHV and other tsetse symbionts (Wigglesworthia and Sodalis) regardless of the species, country, or infection pattern (Fig. 3). The heat map analysis of the qPCR data of G. austeni and G. m. morsitans clearly indicates the contrast between Wolbachia copy number and Wigglesworthia copy number considering the infection pattern, tsetse taxa, or countries. In addition, it clearly shows the low copy number of GpSGHV in the samples showing a high Wolbachia copy number (Fig. 3, Additional file 8: Fig. S5). The bootstrap averages of the metric multidimensional scaling (mMDS) produced clusters based on the species, country, and infection pattern (Fig. 4). The PERMANOVA analysis of the density of GpSGH, Wolbachia, Wigglesworthia, and Sodalis based on the country, tsetse species, and infection pattern indicated that the clusters observed between infection pattern (P = 0.026) and country (P = 0.001) were statistically significant. The interaction between country and infection pattern was not statistically significant (P = 0.123) ( Table 2).

Discussion
The prevalence of GpSGHV and Wolbachia in natural tsetse populations clearly indicated that the two infections were independent (not correlated) in most of the tested tsetse species with only G. m. morsitans and G. austeni presenting a high proportion of coinfections. However, the number of co-infections originally determined by conventional PCR may have been underestimated with conventional PCR as the qPCR analysis carried out in the frame of the present study clearly indicated that a number of initially considered virus-free samples were found to be positive, albeit at low density. It should also be noted that the Wolbachia strains infecting G. m. morsitans and G. austeni are closely related but different, as has been shown by both MLST analysis and, more recently, genome sequencing [7,40,71].
Analysis of G. morsitans and G. austeni co-infected samples suggested that low density of GpSGHV is associated with high density of Wolbachia. Due to the low number of individuals showing this correlation, further analysis is required. Moreover, the screen of wild tsetse populations for GpSGHV and Wolbachia infection indicated that not all Glossina species harbor Wolbachia or GpSGHV. Furthermore, Wolbachia and GpSGHV prevalence was found to differ not only between different  tsetse host species but also between different populations within the same tsetse species [7,11,47,58,59,72].
The potential negative impact (antagonistic effect) of Wolbachia density on the GpSGHV density in natural tsetse populations is in agreement with the recent report on the interaction of Wolbachia and GpSGHV infection in colonized tsetse populations [54]. However, the number of tested flies was not equally distributed between the tsetse taxa and locations, which might explain the lack of detected co-infections in some taxa and, therefore, the low number of taxa (G. austeni and G. m. morsitans) used for investigating the interactions between the GpSGHV and tsetse symbionts. The negative correlation between Wolbachia and GpSGHV infections was also reported in wild-caught G. f. fuscipes collected from Uganda [72]. This conflicts with our findings as no G. f. fuscipes flies with GpSGHV were reported, which might be due to the low number of tested flies used in our study (n = 53).
Several reports have discussed and well documented the negative effect of Wolbachia on RNA viruses in different insect models such as mosquitoes and Drosophila [73][74][75], although there have also been reports about Wolbachia enhancement of both RNA and DNA viruses [76][77][78]. It is worth mentioning that the negative correlation of Wolbachia with GpSGHV was observed only when Wolbachia density was high as the results show the absence of high density (> 10 3.7 ) GpSGHV infection with high-density Wolbachia infection (> 10 7.5 ). However, at low Wolbachia density coinfection occurs with a prevalence of > 10%. Although our study indicated a correlation between high-density Wolbachia and low-density GpSGHV, previous reports suggested that the negative impact of Wolbachia on insect viruses is density dependent [76,79].
The assessment of the infection density (copy number per fly) of all four microbes (GpSGHV, Wolbachia, Wigglesworthia, and Sodalis) in the same tsetse flies indicated that Wolbachia infection at high density has a significant negative correlation with Wigglesworthia infection in G. m. morsitans but not in G. austeni. However, the latter might be due to the low number of analyzed G. austeni flies (n = 21) compared to G. m. morsitans (n = 91). On the other hand, Wolbachia density levels do not correlate with Sodalis. The nature of the negative interaction between Wolbachia and Wigglesworthia is unclear. Whether this negative correlation between Wolbachia and Wigglesworthia is present in other tsetse species beyond G. m. morsitans remains to be seen.
The positive correlation between GpSGHV infection and Wigglesworthia infection observed in G. m. morsitans conflicts with the negative correlation observed in the same species of colonized flies [54]. This result might reflect a specific adaptation between a specific strain of Wigglesworthia, which reacts in a specific way to increase its density in the presence of GpSGHV as a manner to restore and enhance the host immune system against the virus infection [80]. The difference in the interaction between the GpSGHV and Wigglesworthia between the results of this study and the results of Demirbas-Uzel et al. [54] might be due to: (i) difference in the host strain/genotype as the G. m. morsitans individuals were collected from several countries in east Africa (Tanzania, Zambia, and Zimbabwe) while the colonized flies originated from Zimbabwe and have been maintained in the colony since 1997; (ii) different strain(s) of Wigglesworthia circulating in the field samples compared to the ones present in colonized flies [60]; (iii) different strain(s) of the GpSGHV in the field samples [58]; (iv) difference between field and laboratory conditions where the stress from handling the large number of flies in high density in the laboratory might negatively affect Wigglesworthia density levels and/or performance. The same reasons may also explain the difference observed between field and

Conclusions
The present study, despite its limitations regarding the size of samples and the lack of knowledge about the age, nutritional and trypanosome infection status, and environmental conditions at the time of collection of field specimens, shows a snapshot image of the density levels of tsetse symbionts and SGHV under field conditions and clearly indicates that the interactions/association between the tsetse host and its associated microbes are dynamic and likely species specific, and significant differences may exist between laboratory and field conditions. Further studies are needed to clarify the interaction between tsetse symbionts and GpSGHV under field conditions.