The dynamics of viral RNA levels over time are virus-specific

Our laboratory has previously established persistent viral mono-infections in D. melanogaster flies of the w¹¹¹⁸ genetic background using different viruses: DAV, DCV, Bloomfield virus, and Nora virus [25]. All these positive-sense single-strand RNA viruses belong to different viral families and naturally infect D. melanogaster. They are transmitted primarily through orofecal contamination. In this work, we measured key biological traits over time in the infected and the uninfected control population (Fig 1A). To maintain continuous exposure to contaminated food while preventing excessive viral accumulation, flies from the different persistently infected lines were transferred to fresh food every other day. Persistent viral infections significantly reduce the survival of infected flies, as previously described (Fig 1B:). DAV infection has the most pronounced effects on survival, followed by DCV, Bloomfield virus, and Nora virus (Mantel–Cox test infected versus uninfected p < 0.001 for all the virus-infected populations). This is reflected by the time at which 50% of the population has perished, the median survival (Fig 1B): 25 days post-eclosion (dpe) for DAV, 28 dpe for DCV, 36 dpe for Bloomfield virus, and 52 dpe for Nora virus, compared to 63 dpe for uninfected flies. This information enabled us to establish different time points for further characterization of the infections: 1 dpe as an early time point, 12 dpe as an intermediate time point just before the onset of virus-induced mortality, and the median survival as a late time point specific for each condition. Additionally, we added an additional time point for DCV at 6 dpe, as mortality starts earlier for this virus. In the present study, these time points served as the basis for characterizing the impact of persistent viral infections on the physiology and behavior of infected hosts (Fig 1A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Experimental set-up, survival, offspring production, viral dynamics, and bacteriome analysis of persistently infected fly populations.

A) Experimental design. The large arrow indicates the days post-eclosion (dpe) during which specific experiments were conducted. The survival of persistently infected stocks has been monitored from eclosion until death across all stocks. The viability of offspring was assessed after a 24-hour mating period, until no more embryos were laid (17 dpe). Black dots indicate the specific days when the climbing assay and the sampling for the bacteriome studies were performed for all stocks. At the bottom of the graph, regarding the transcriptomics experiment, each colored dot signifies the date of sample collection for RNA sequencing corresponding to each different fly stock. B) Left panel shows the survival of w¹¹¹⁸ flies uninfected (green) or persistently infected with DAV (red), DCV (yellow), Bloomfield virus (dark blue), or Nora virus (purple). Shaded regions indicate 95% confidence intervals. The data (n = 180 flies/condition in 3 independent experiments with 3 biological replicates/experiment) were obtained from Nigg and colleagues. Statistical significance was assessed via the Mantel–Cox test. Right panel indicates the dpe at which 50% of the population is deceased. C) Viral infection dynamics. The upper panel shows viral RNA levels for DAV, DCV, Bloomfield virus, and Nora virus in infected populations. Viral RNA levels were measured in whole individuals at various times post-eclosion. Only data from infected individuals is presented. The Y-axis denotes the relative viral RNA level as log₁₀ 2^ΔCt between the viral RNA and the housekeeping gene, Rp49. The red bars represent the median values along with the interquartile range. For each virus, significance was calculated using a general linear model where the time point was the fixed factor; Tukey contrasts were used for post-hoc analyses. Letters over the bars indicate letter-based grouping for multiple comparisons. Bottom panel: viral prevalence showing the proportion of infected (black) or noninfected (gray) flies at the corresponding time point. Numbers below the pie charts represent the number of infected individuals over the number of individuals sampled. D) Cumulative viable offspring for each population, normalized over female number. Total viable offspring produced by each female was measured cumulatively across the different fly populations throughout the duration of the experiment, accounting for fly mortality. The Y-axis shows the total viable offspring per female and the X-axis the population studied. Significance was calculated using a general linear-mixed model where the virus status was the fixed factor and experiments a random effect; Tukey contrasts were used for post-hoc analyses. E) Relationship between the offspring and the survival of fly stocks. On the Y-axis the total offspring per female. On the X-axis, the median day of survival, in dpe, of each infected stock. Each dot on the graph represents a total offspring experiment. All infected samples from the same infected stock have the same median survival. For each virus-infected stock, the mean offspring is denoted as a black rhombus. Pearson’s correlation coefficient r = −0.633 and p value < 0.001 are depicted in the graph. F) Bacterial composition of pools of at least 20 females from the studied fly populations except for the one infected with Bloomfield virus. Females coming from the uninfected or persistently infected stocks were collected at 12 dpe, surface sterilized and their DNA was extracted. For each population, microbiome composition is expressed as the relative abundance in percentage of bacterial genera detected by sequencing of the full bacterial 16S rRNA gene. Each color corresponds to a specific bacterial genus sequenced in the microbiome. Bacteria belonging to different operational taxonomy units (OTUs) but belonging to the same bacterial genus are displayed with the same color. For all the figure, ***p < 0.001, **p < 0.01, *p < 0.05. G) Total bacterial load of pools of 10–20 females coming from the uninfected or persistently infected stocks except for the one infected with Bloomfield virus. Females coming from the uninfected or persistently infected stocks were collected at 12 dpe, surface sterilized and their DNA was extracted. The experiment was performed twice, with three pools per line analyzed in each experiment. Bacterial load was assessed via qPCR targeting the 16S bacterial ribosomal gene. The Y-axis denotes the relative bacterial DNA level as log₁₀ 2^ΔCt between the bacterial 16S gene and the housekeeping gene, Rp49 (mean ± SEM). Shape fill indicates the replicate. Statistical significance was determined with a generalized linear model using log₁₀ 2^ΔCt values and condition as fixed factors (F_(4,24) = 26.63, p < 0.001) and the replicate as a random factor. Tukey’s adjustment was applied to post-hoc comparisons. Underlying data for all panels in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003437.g001

To characterize the dynamics of viral load within the infected populations, we examined viral RNA levels in individual flies at the aforementioned early, intermediate, and late stages of infection. Viral loads were quantified by RT-qPCR by measuring viral RNA levels relative to the Drosophila housekeeping gene Rp49. Flies were classified as infected if their Ct values for viral RNA were below 35, with melting temperature profiles indicating the presence of a single amplicon. We observed distinct patterns of viral accumulation for each virus (Fig 1C, upper panels). DAV exhibited a rapid increase in viral RNA levels early post-eclosion, with a median increase of ~4 and ~5 log₁₀ orders of magnitude at 12 and 25 dpe, respectively. In contrast, DCV displayed a relatively stable pattern in viral RNA levels over time, with no significant changes in viral loads. Interestingly, we noticed that DCV-infected individuals stably clustered into two distinct groups, displaying viral RNA levels several orders of magnitude apart. Bloomfield virus RNA levels rose steadily over the lifetime of infected individuals, increasing by 1 and 2 log₁₀ orders of magnitude at 12 and 36 dpe, respectively. In contrast, Nora virus RNA levels remained consistent at the earlier time point but rose significantly by ~3 log₁₀ orders of magnitude by the final time point (General Linear Model [GLM] letters over the conditions indicate grouping for multiple comparisons).

Our examination of viral RNA levels allowed us to estimate infection prevalence in the different populations over time (Fig 1C, lower panels). A viral infection is considered prevalent in a population if individuals continue to be infected over time. DAV and Nora virus were consistently detected in 100% of individuals, except for a slight reduction in Nora virus prevalence at the later stage (~91% of individuals infected at 52 dpe). Bloomfield virus was highly prevalent in the population at all time points, although not all individuals were infected at any given time, and infection prevalence varied from ~91% to ~72%. DCV was detected in all individuals at 1 dpe, and its prevalence in the population progressively decreased over time, with only ~36% of the individuals infected at the latest time point.

Persistent viral infections impact viable offspring production

Persistent viral infections can reduce individual survival rates, which may consequently affect the viability of the population over time. To understand the broader impact of these infections, we investigated whether flies from populations with a prevalent viral infection produced altered numbers of viable offspring compared to flies from the uninfected population. For this experiment, we measured viable offspring production as the number of progeny that successfully emerged as adults. This metric is a crucial aspect of reproductive success and population dynamics. Any impact on the production of viable offspring would indicate that persistent infections influence additional critical biological traits related to fitness beyond the previously observed reduction in survival.

To explore potential changes in viable offspring production in the infected populations, 20 males and 20 females from each infected population or from the uninfected population were collected on the day of eclosion and placed together in fresh vials for 24 hours to allow mating. After 24 hours, males were discarded, and the 20 females were transferred into new vials with fresh food each day. Initial female density across all populations was uniformly established at 20 females per tube, as these conditions provide optimal surface area and total tube space, thereby minimizing crowding effects and promoting normal development. The viable offspring were counted daily as they emerged and the viable offspring count was normalized based on the number of live females remaining in each vial. This normalization allowed us to account for changes in the density of females per tube due to mortality over the course of the experiment.

Flies collected from all four virus-infected populations exhibited distinct patterns of offspring production compared to uninfected controls (S1 Fig). Uninfected flies exhibited a peak in viable offspring production at 1 dpe (10.3 ± 2.02 adults emerging per female, mean ± SD), followed by a gradual decrease in offspring production until 6 dpe and a pronounced decline thereafter, reaching zero by 16 dpe. Similarly, populations infected with DAV and Bloomfield viruses also peaked in viable offspring production at 1 dpe (11.5 ± 2.07 and 11.6 ± 1.30 offspring per female, respectively) and experienced a sharp decline in offspring production after 6 dpe. However, during early time points, DAV, and Bloomfield virus-infected populations produced more viable offspring per female than the uninfected control population (GLMM, DAV P = 0.022, Bloomfield P = 0.031). DCV-infected flies displayed a distinct pattern, producing significantly more viable offspring than the uninfected ones only at later time points (GLMM, P < 0.001). In contrast, Nora virus-infected flies consistently produced fewer viable offspring (GLMM, P = 0.010), with their offspring production peaking at 4 dpe (5.28 ± 2.15 viable offspring per female).

Studying both daily and cumulative offspring production offers a comprehensive perspective on reproductive fitness. While daily offspring production captures temporal patterns and short-term reproductive dynamics, cumulative counts reveal the total reproductive output across a female’s life span.

Analysis of cumulative offspring per female reveals that DCV-infected flies produced more viable offspring (91.5 ± 32.1) than the uninfected group (67.7 ± 11.6) (Fig 1D, GLMM, P = 0.004). In contrast, Nora virus-infected flies produced significantly fewer offspring per female (41.5 ± 15.5, GLMM, P = 0.034). Although DAV and Bloomfield virus-infected flies showed higher early production rates, their lifetime reproductive output did not differ significantly from uninfected controls. These results demonstrate that persistent viral infections have virus-specific impacts on reproductive fitness, with some viruses enhancing and others diminishing total offspring production.

Analysis of the relationship between total offspring production and female survival reveals a negative correlation across virus-infected stocks (r = −0.633, P < 0.001; Fig 1E). DAV and DCV infections resulted in higher offspring production but shorter lifespans, while Bloomfield and Nora virus infections showed the opposite pattern: lower reproductive output but longer survival. This trade-off suggests a potential reallocation of resources from reproduction to defense mechanisms in response to viral infections.

Harboring persistent viral infections changes fly bacteriome

We next investigated whether persistent viral infections influence the fly bacteriome—the bacteria associated with the host. All fly populations in this study shared the same genetic background and were reared under identical conditions, except for the presence of different persistent viral infections. This controlled setup allowed us to attribute any observed differences in the associated bacteria to the presence of different viral infections.

Pools of at least 20 flies in three replicates were collected at 12 dpe, surface sterilized and their DNA was extracted. Microbiome composition was assessed by amplification and sequencing of the full 16S rRNA bacterial gene through Oxford Nanopore sequencing. Despite good DNA yields for all samples, bacterial 16S amplification was not successful for Bloomfield virus-infected flies, even when DNA was extracted from pools of 40 flies. We therefore could not determine the bacterial composition of Bloomfield virus-infected flies.

An average of 82,053 reads was obtained per sample (min: 25,673; max: 139,361). As previously described, we found a relatively low bacterial diversity in all the analyzed populations, with only 7–21 operational taxonomy units (OTUs) identified in our samples. α diversity analyses did not reveal any significant difference in the OTU diversity of uninfected or persistently infected flies (S3A Fig, pairwise Wilcoxon test on Chao1 and Shannon indexes, P > 0.46 for all comparisons). Similarly, beta diversity analysis did not identify specific differences in the bacterial community structure of the different fly lines (S3B Fig, PERMANOVA on infection, F = 1.57, P = 0.24). After taxonomical assignation of reads, we confirmed that the microbiome of all our Drosophila populations was dominated by Lactiplantibacillus bacteria characterized by a 16S sequence identity higher than 99% (Fig 1F). The failed amplification of the 16S gene of Bloomfield virus-infected flies suggested a significant difference in the total microbiome load of our fly populations. We thus determined the total bacterial load in whole flies by 16S qPCR. Nora virus-infected flies showed a significant increase in total bacterial loads compared to uninfected flies. Similarly, flies persistently infected with DAV or DCV were characterized by a nonsignificant increase in their total bacterial loads (Fig 1G). Although we successfully amplify bacterial 16S from Bloomfield virus-infected flies (with amplification cycles earlier than no-template controls), their Ct values closely matched those of extraction negative controls—blank samples processed alongside fly samples. This confirms that flies persistently infected with Bloomfield virus harbor a markedly reduced microbiota load.

Taken together, these data indicate that persistent viral infection does not influence the composition of the microbiome but significantly impacts the total bacterial load colonizing the fly.

Persistent viral infections alter locomotor function in a virus-dependent manner

Locomotor function is a key indicator of an organism’s overall health and physiological status. In flies, locomotion reflects both muscular and neurological capacity, making it a valuable marker for assessing the general impact of viral infections on host physiology.

We assessed locomotor function using the reverse geotaxis assay, which measures the flies’ upward climbing instinct over a specified period [26]. The assay was performed on 20 mated females from the uninfected and the different infected stocks at 1, 6, 12, and 20 dpe. This final time point was chosen as a common late date for all fly stocks, occurring prior to the day of median survival for all of the infected populations. Flies were placed in an empty tube and collected at the bottom by gently tapping the tubes. After 3 s, a picture was taken and the position of each fly was recorded manually. Our observations indicated that persistent viral infections had a mixed effect on locomotor function over time (S2 Fig). DCV- and Nora virus-infected flies consistently showed impaired locomotion abilities compared to uninfected flies at almost all the tested time points (S2 Fig). In contrast, DAV- and Bloomfield virus-infected flies showed mixed locomotion patterns. DAV significantly reduced mobility at 1 dpe but significantly increased it at 6 and 20 dpe (GLMM, P < 0.001, P < 0.001, and P = 0.011, respectively). Bloomfield virus significantly reduced mobility at 1 dpe and increased it at 12 dpe (GLMM, P < 0.001 at both time points). When examining the fly distribution patterns along the tubes we obtained similar results, with a higher proportion of DCV- and Nora virus-infected flies found in the bottom of the tube at almost all time points (S2 Fig). DAV infection resulted in a higher proportion of flies climbing to the top of the tube at 6 dpe, but a higher proportion of flies at the bottom at 12 dpe (cumulative link mixed models [CLMM], P < 0.001 and P = 0.001). Bloomfield virus-infected flies were found more often on the bottom of the tube at 1 and 20 dpe, while they climbed more to the top at 12 dpe (CLMM, P < 0.001, P = 0.003, and P = 0.032, respectively). Taken together, these results indicate that persistent viral infections affect the climbing abilities of infected flies, with a general trend of lower climbing height, particularly in flies harboring Nora virus infections. These findings underscore the systemic impact of persistent viral infections on host physiology and behavior.

Persistent viral infections alter the fly activity and sleep duration while preserving circadian rhythms

We continuously monitored fly movement over a 72-hour period under 12:12-hour light:dark conditions to assess the impact of persistent viral infections on activity patterns and sleep behavior. The movement data revealed that all experimental groups maintained clear circadian rhythmicity, with peaks in activity occurring at light-dark transitions and similar overall temporal patterns of activity (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Activity patterns and sleep behavior of persistently infected flies.

A) Mean locomotor activity over a 72-hour period under 12:12 light:dark conditions (T0–T72). Activity was quantified as infrared beam crossings per minute using Drosophila Activity Monitors. Colored lines represent different infection conditions (uninfected: green, Bloomfield virus: blue, DAV: red, DCV: yellow, and Nora virus: purple), with shaded areas indicating 95% confidence intervals. Gray shaded vertical bands indicate dark periods. Time is shown in hours, with T0 representing the beginning of the first light cycle after one night cycle of acclimation. B) Sleep time comparison during night and day phases. Sleep was defined as periods with no activity for at least 5 consecutive minutes. Relative sleep time is shown as proportion of total available time in each phase. Box plots indicate the median and interquartile range; each point represents an individual fly. Statistical significance compared to uninfected controls was determined using a linear mixed-effects model with experimental replicate as a random effect (***P < 0.001, **P < 0.01, *P < 0.05). C) Activity intensity during wake periods. Box plots show the average number of beam crosses per minute during nonsleep bouts. Each point represents an individual fly. Statistical significance compared to uninfected controls is indicated (*P < 0.05, ns = not significant) and was determined using a linear mixed-effects model with experimental replicate as a random effect. Data represent three independent experiments, each including 8–12 flies per condition. All experiments were conducted with 12-day-old mated female flies at 25°C under a 12:12-hour light:dark cycle. Underlying data for all panels and the code used to generated are deposited in Zenodo (https://zenodo.org/records/17175344).

https://doi.org/10.1371/journal.pbio.3003437.g002

Analysis of sleep periods, defined as immobility bouts lasting at least 5 min ([27] Cirelli and Bushey, 2009), revealed significant differences between infected and uninfected flies (Fig 2B). During the night phase, flies infected with Bloomfield, DAV, or Nora virus spent significantly more time sleeping compared to uninfected controls (GLMM, all P < 0.001). DCV-infected flies showed no significant difference in night sleep compared to uninfected controls. Day-phase sleep patterns displayed a similar profile, with all infected flies exhibiting significantly increased sleep duration compared to uninfected controls (GLMM, all P < 0.001).

We further analyzed the intensity of movement during active periods (when flies were not asleep) to determine whether viral infections affected the quality of movement (Fig 2C). Bloomfield-infected flies showed significantly reduced movement intensity during their active periods compared to uninfected controls (GLMM, P < 0.05), suggesting that even when Bloomfield virus-infected flies were active, they moved less vigorously than uninfected flies. In contrast, flies infected with DAV, DCV, or Nora virus showed movement intensities during active periods that were comparable to uninfected controls, with no significant differences detected. Together, these results demonstrate that while persistent viral infections do not disrupt fundamental circadian rhythmicity in Drosophila, they significantly modify both sleep duration and movement intensity in a virus-specific manner.

Host transcriptional response is virus-specific

To better understand the alterations induced by persistent viral infections, we examined the transcriptional response of flies at early (1 dpe), intermediate (12 dpe), and late stages of infection (at median survival times). For each time point and condition, RNA was extracted from individual flies for individual measurements of viral load via RT-qPCR. Subsequently, aliquots of RNA from 3 to 5 infected individuals harboring similar viral load levels (Fig 1C) were pooled together for the generation of bulk RNA-seq libraries.

Principal component analysis revealed that each enteric virus elicited a distinct transcriptional response in the host (S4 Fig). Nora virus infection was characterized by the most unique transcriptional response at 1 dpe, with the differential regulation of several genes (S5 and S6 Figs). In contrast, at the same early time point, other viruses induced little changes in transcriptional profiles of infected flies and samples clustered together with uninfected flies (S4–S6 Figs). At a more advanced time point (12 dpe) the transcriptional profiles of DCV and Bloomfield virus-infected flies clustered together with uninfected flies, while Nora virus and DAV-infected flies were characterized by unique transcriptional responses (S4 Fig). By the days of median survival, all viruses had induced significant changes in transcriptional profiles (S4 Fig).

No common differentially expressed gene (DEG) was identified across all viral infections at any time point (S7 Fig). Furthermore, we could not identify a consistent set of host genes responding to the infection over time, thereby illustrating the differential impacts of viral infection on the host as the infection progresses (S7 Fig). Even at median survival, no common transcriptional profile was shared among all infections (S7 Fig). We also observed an increase in the number of DEGs as infection progressed for the flies persistently infected with DAV, DCV, and Bloomfield virus (S6 and S7 Figs). Many of the top DEGs identified at early and intermediate time points are uncharacterized (S6 Fig).

Focusing on the genes known to participate in antiviral immune responses, Nora virus elicited the highest levels of immune regulation at 1 dpe (Fig 3A). At 12 dpe, DAV had the highest upregulation of genes associated with immune defense and stress response pathways, indicating a potent antiviral response (Fig 3A). Interestingly, specific immune genes upregulated during DAV infection were downregulated during Nora virus infection, suggesting fundamentally different immune responses to these viruses (Fig 3A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Transcriptomic profiles of uninfected or persistently infected flies.

A) Heatmaps expressing the Z-scores of gene expression fold change of infected populations in comparison with uninfected ones. Genes are represented in the Y-axis and the infecting virus on the X-axis. For each virus, the results are calculated on the average of 4 samples. Top left: heatmap showing all the genes reported to be involved in antiviral immunity (subset of 191 immune antiviral genes, S1 File) and their differential expression in our samples at 1 dpe. Top right: heatmap showing the top 20 differentially expressed immune genes at 1 dpe. Bottom left: heatmap showing all the genes reported to be involved in the insect’s immune response and their differential expression in our samples at 12 dpe. Bottom right: heatmap showing the top 20 differentially expressed immune genes at 12 dpe. On the right side of each heatmap, a colored bar indicates the pathway in which the gene is reported to participate. B) Heatmaps showing all the genes reported to be involved in the insect’s locomotion (a subset of 264 locomotion genes, S1 File) and their differential expression in our samples at 1 dpe1. Top left: heatmap showing the top 20 differentially expressed locomotion genes at 1 dpe. Top right: heatmap showing all the genes reported to be involved in the insect’s locomotion and their differential expression in our samples at 12 dpe. Bottom left: heatmap showing the top 20 differentially expressed immune genes at 12 dpe. Bottom right: heatmap showing the top 20 DEGs at 12 dpe. Underlying data for all panels in this figure can be found in S2 Data.

https://doi.org/10.1371/journal.pbio.3003437.g003

Analysis of genes involved in locomotion revealed significant differential expression in Nora virus-infected flies at 1 dpe. Notably, Lsp2, a gene reported to be involved in motor neuron axon and synaptic target inhibition [28], was upregulated in Nora virus-infected samples (Fig 3B). As infection progresses, we observed an increasing number of DEGs related to locomotion, with Nora virus infection showing an opposite transcriptional profile compared to DAV infection (Fig 3B). At 12 dpe, Nora virus infection upregulated genes involved in gravitaxis regulation (CG3,857), negative phototaxis (Ptth), light reception (Rh5), and neuronal function (nompC, Pdf) [29–33]. Furthermore, Nora virus infection induced the upregulation of M6 (involved in several neurological processes, including positive phototaxis) and the downregulation of fz (neuronal function) and Neto (neuromuscular roles and hatching behavior) [34]. Conversely, DAV-infected samples showed upregulation of Neto, esg (related to stem cell maintenance and nervous system development), and Mlc2 (flight musculature) [35–38]. DCV and Bloomfield virus infections induced more modest differential expression of locomotion-related genes. Taken together, these results indicate that the host response to persistent infection is virus-specific.

Viral infection induced transcriptional changes even after viral clearance

While persistent viral infections remain prevalent at the population level, individual flies can clear the infection over time (Fig 1C; [25,39]). This was particularly evident for DCV, where all individuals were infected at 1 dpe, but more than half had cleared the virus by 28 dpe (Fig 1C). This provided an opportunity to examine whether viral clearance restores normal gene expression or leaves a lasting transcriptional signature.

Therefore, we analyzed the transcriptional response of flies that had cleared DCV infection by 28 dpe and found that it remained significantly altered compared to the never-infected control population (Fig 4A). Comparison of the transcriptional response of currently infected and cleared flies showed minimal differences between the two (Fig 4B), with most DEGs related to mitochondrial function. These included mitochondrial transfer RNAs and mitochondrial proteins such as mt:ATPase8 (mitochondrial ATPase subunit), l(3)neo43 (cytochrome c oxidase assembly), Mcad (mitochondrial fatty acid beta-oxidation), mtSSB (mitochondrial single-stranded DNA-binding protein), ND-MWFE (respiratory chain complex component), and sea (inner mitochondrial membrane carrier protein). There is a marked downregulation of mitochondrial genes in the cleared samples, which indicates that energy metabolism is upregulated during DCV infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Transcriptomic profile of flies at 28 dpe that were infected with DCV but cleared the viral infection.

A) Principal component analysis of uninfected samples, DCV Infected, and DCV Cleared. On the X-axis, we indicate the first principal component (PC1) and the percentage of variance explained by this component in brackets. On the Y-axis, we indicate the second principal component (PC2) and the percentage of variance explained by this component in brackets. B) Volcano plot indicating the differentially expressed genes (DEGs) between DCV Cleared samples and DCV Infected ones. Genes significantly up- (in red) or down- (in blue) regulated. In gray, we represent those genes with differential expression values lower than 1.5 log₂ fold change and p-adjusted values lower than 0.05. These thresholds are indicated by dotted lines in the graphs. The names of the DEGs are shown. C) Heatmaps representing the differential expression patterns between DCV Infected and DCV Cleared samples when compared to the Uninfected of the same age. Heatmap on the right shows the top 200 DEGs in our samples in terms of absolute log₂ fold change value at dpe 28. Heatmap on the left shows the top 20 DEGs at dpe 28; gene names are indicated on the left. D) Heatmaps representing the differential expression patterns for reported immune genes between DCV-infected and DCV Cleared samples when compared to the Uninfected of the same age. Heatmap on the left shows all the genes reported to be involved in the insect’s immune response and their differential expression in our samples at dpe 28. Heatmap on the right shows the top 20 differentially expressed immune genes at dpe 28; gene names are indicated on the left. On the right side of each heatmap, a colored bar indicates the pathway in which the gene is reported to involved. Underlying data for all panels in this figure can be found in S3 Data.

https://doi.org/10.1371/journal.pbio.3003437.g004

Both infected and virus-cleared flies maintained significant transcriptional differences compared to never-infected controls, including sustained upregulation of immune response genes (Fig 4D). These results indicate that viral infections induce long-lasting or potentially permanent alterations in host gene expression, particularly affecting mitochondrial and immune functions, that persist even after viral clearance.

Fly stocks

For all experiments, we used Wolbachia-free w¹¹¹⁸ flies maintained on a standard cornmeal diet. This diet was prepared by combining 440 g inactive dry yeast, 440 g corn meal, and 60 g agar in 6 L of osmotic water. The mixture was autoclaved and after cooling, 150 ml moldex solution (20% methylhydroxybenzoate) and 29 ml propionic acid were added as preservatives. All experiments were done at 25°C under a 12:12 light:dark cycle. All fly lines were cleaned of possible persistent infections (viruses and Wolbachia) as described previously (Merkling and van Rij, 2015).

Viral infections

Persistently infected w¹¹¹⁸ flies were established following the method described by Nigg and colleagues [25] and are the same stocks used in the work of Nigg and colleagues. Persistent infections of four RNA viruses were established in w¹¹¹⁸ D. melanogaster using tailored methodologies. For DCV, a persistently infected bw1; st1 Ago3t3/TM6B; Tb+ line (Bloomington #28270) was used to environmentally transfer DCV to naïve w¹¹¹⁸ flies. Infected donors were housed in a fresh vial for 3 days; after their removal, w¹¹¹⁸ flies were exposed to the contaminated environment for 3 days. The F0 generation was transferred to a new vial, and their F1 progeny were confirmed as persistently infected. For DAV, the Australian isolate (DAV_HD from the van Rij lab) was injected into w¹¹¹⁸ adults (50 nl/fly). Identical environmental transfer followed: injected flies contaminated a vial for 3 days before naïve w¹¹¹⁸ exposure. Persistently infected F1 progeny were isolated and maintained. Bloomfield virus was identified in a contaminated Dipt-GFP stock (BDSC Cat#55709). Homogenate from these flies was filtered (0.22 µm) and injected into w¹¹¹⁸ adults. Injected F0 flies were removed after 3 days; F1 and F2 progeny underwent sequential generational passages (5–9 days/vial) to establish persistence in F3 adults confirmed via RNA sequencing. The Nora virus-persistent line originated from a w¹¹¹⁸ line persistently infected with Nora virus that was gifted by Dr. Stefan Ameres. Upon receipt, we repeatedly backcrossed the infected w¹¹¹⁸ flies to the w¹¹¹⁸ line maintained in our laboratory. Subsequently, we confirmed that the backcrossed flies retained persistent infection with Nora virus by RNA sequencing. The stocks were maintained at 25°C under a 12:12 light:dark cycle.

The presence or absence of persistent infections was determined by RT-PCR with specific primers for Nora virus, DAV, and DCV (S2 Table). In addition, total RNA from persistently infected adults was extracted and sequenced. The resulting reads were mapped to a database of all known Drosophila viruses (https://obbard.bio.ed.ac.uk/data.html) confirming that only the desired virus was present [25].

Survival analysis

Survival analyses of the flies coming from persistently infected stocks were conducted using 20 female flies/replicate sorted at 1 dpe. Male and female flies were collected on the day of eclosion and left to mate for 24 hours. Subsequently, only female flies were sorted in new tubes containing 20 females each. Each tube was considered a biological replicate, and a total of nine biological replicates (from three independent experiments) were analyzed. Survival was monitored daily by counting the number of dead flies in each vial, and flies were transferred to new vials every 2 days. Fly survival was assessed from the day of adult eclosion until death.

Viral quantification

Individual female flies were collected in 300 μl of TRIzol reagent (Invitrogen, 15596026). The samples were homogenized using a pestle, and RNA was extracted following manufacturer’s instructions, supplemented with GlycoBlue (Invitrogen, AM9516) during the precipitation process. RNA concentrations were measured using the Qubit RNA BR Assay Kit (Invitrogen, Q10211). Following extraction, 4 μl of each RNA sample were treated with DNAse I (Roche, 04716728001), and 2 μl of the final volume were used in the RT reaction. RT was performed using random primers and Maxima H Minus Reverse Transcriptase (Thermo scientific, EP0751) following manufacturer’s instructions. The resulting cDNA was diluted 1:10 with MilliQ water. The qPCR was performed using Luminaris Color HiGreen qPCR Master Mix (Thermo Scientific, K0374) in a QuantStudio 7 flex machine (Thermo Scientific). The reactions were done in triplicates of 10 μl each and using the corresponding set of primers targeting a replication protein for each virus (S1 Table). The cycling conditions were as follows: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 60 s at 60°C. A standard melt curve analysis was performed after the cycling protocol. For the analysis of the results, the threshold for viral RNA detection was set to 35 Ct (i.e., if a sample had Ct higher than 35 it was considered noninfected). The Ct values of the virus for a sample were normalized against the Ct value of the housekeeping gene Rp49 for that sample (ΔCt). The 2^−ΔCt values were calculated per sample and their log₁₀ plotted.

Viable offspring production

To assess viable offspring production, 20 female and 20 male flies from infected or uninfected stocks were left to mate in clean fly vials for 24 hours after eclosion. After this period, males were discarded, and the mated females were transferred to new tubes daily until no more progeny was produced (17 days). All tubes were kept at 25°C under a 12:12 light:dark cycle, and the number of adult flies (i.e., the viable progeny) in each tube was counted after 14 days. The number of offspring was normalized to the number of adult females placed in each tube daily to account for those that did not survive throughout the test duration.

Climbing assays

Climbing assays were performed following the method described by Barone and Bohmann [68]. Briefly, a 9 cm empty tube was used, and 20 female flies were placed inside. With the cotton cap on, tubes had a maximum climbable height of 7.2 cm. The tubes were tapped three times over the span of 3 s, and the climbing height was assessed by taking a picture 3 s after the last tap. Flies’ height was assessed manually on the pictures using a metric guide on the back of the tubes. For categorical classification, flies were considered “Top” if they surpassed 6.2 cm, “Bottom” if they were between 0.0 and 1.0 cm, and “Middle” if they were between 1.0 and 6.2 cm.

Behavioral analysis

We conducted behavioral analyses using DAM5H Drosophila Activity Monitors (Trikinetics, Princeton, MA, USA). Using the DAM5H we tracked the locomotor activity of individual flies housed in a 5 mm-diameter, 80 mm-length tube. The system records activity when a fly crosses one of 15 infrared beams of each tube. The monitor quantifies movement by registering infrared beam disruptions as activity “counts.” A fly that moves continuously through a beam zone or returns to the same beam multiple times generates several counts, whereas a fly that simply enters a beam area and stops moving is recorded as only a single count. The count data were registered per minute and were used to evaluate the locomotion activity and to identify sleep periods, which were defined as immobility bouts lasting at least 5 min. All behavioral experiments utilized 12-day-old mated female flies. Flies were individually housed in DAM tubes containing standard Bloomington cornmeal diet. All experiments were conducted at 25°C under a 12:12 light:dark cycle. To allow for acclimation, flies were placed in the monitoring tubes 1 hour before the start of a dark cycle. Data collection began at the onset of the following light cycle (designated as T0) and continued for three complete cycles (until T2). We performed three independent experiments, each including 8–12 flies per experimental condition.

Microbiome composition

Female flies from uninfected or persistently infected stocks were collected at 12 dpe. External microbes were removed by surface sterilization using subsequent washes in 10% bleach for 5 min, 70% ethanol for 5 min and three rinsing steps in sterile water. Since the fly microbiota is fairly stable in populations reared together, microbiota composition was analyzed in three pools of 20 flies per condition. Flies were placed in a lysis buffer (10 mM Tris-HCl (pH 8), 26 mM EDTA, 0.5% SDS, and 5 mg/mL lysozyme) and homogenized with sterile pestles. Blank samples consisting of lysis buffer were processed alongside the flies and served as extraction negative controls. After incubating the lysates for 1 hour at 37°C, total DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer’s instruction. For each sample, 5 µL of undiluted genomic DNA was used as template in a 20 µL final reaction volume for amplification of the 16S rRNA gene. The primers containing locus-specific sequence 8/27_Forward and 1492_Reverse were employed (S1 Table). PCR amplification was carried out for a total of 35 cycles. No visible amplicons were obtained from extraction negative controls. The resulting amplicons were used for library preparation with the Ligation Sequencing Kit V14 (SQK-LSK114, Oxford Nanopore Technologies) following the manufacturer’s instructions. Subsequently, the ONT DNA libraries were sequenced on a PromethION 2 Integrated apparatus using a FLO-PRO114M flowcell. The raw ONT sequencing data were base-called (super-accurate algorithm), demultiplexed and sequencing adaptor trimmed using the software dorado (version 7.3.11). The produced long-reads in fastq format were quality checked using the software NanoPlot (version 1.42.0) and then quality filtered using the software filtlong (version 0.2.1), discarding any long-reads with an average quality score below Q17 and read lengths not in the range of 1,200–2,300 bp. The surviving long-reads were in-silico PCR amplified to verify the locus using the usearch software (version v11.0.667), then clustered using the software isONclust (version 0.0.6.1). For each cluster, a consensus sequence was generated to represent it as an OTU. OTU sequences were compared to the reference sequences of the NCBI targeted loci DB (https://www.ncbi.nlm.nih.gov/refseq/targetedloci/ - downloaded 15 Feb. 2024). Taxa were predicted and confidences were calculated using the SINTAX algorithm implemented in the usearch software (version v11.0.667). DNA library construction, sequencing, read processing, and taxonomy identification described in these sections were performed by Microsynth AG (Balgach, Switzerland). Alpha and beta diversity measures were estimated using the Chao1 and Shannon indices (alpha diversity) and the

Bray-Curtis dissimilarity (beta diversity) in R using the phyloseq package (version 1.42.0). Both alpha and beta diversities were calculated on microbiota composition at the OTU level. Pairwise Wilcoxon tests were used to determine differences in alpha diversity measures among the different fly lines, while PERMANOVA was used to determine the effect of the infection status on Bray–Curtis dissimilarities. The barplot was produced in Prism (version 10.4.2). As Lactiplantibacillus OTU sequences were characterized by a sequence identity >99%, we decided to display the bacterial genus without indicating the bacterial species. Raw sequencing reads will be made available upon publication.

Bacteriome load

Female flies from uninfected or persistently infected stocks were collected at 12 dpe. External microbes were removed as described above. Ten to twenty flies per condition were collected in two independent experiments. Total DNA was extracted as described above. Bacterial DNA was quantified via amplification of the 16S rRNA bacterial gene and the fly ribosomal gene Rp49 via qPCR using the Luminaris Color HiGreen qPCR Master Mix (Thermo Scientific, K0374) in a QuantStudio 7 flex machine (Thermo Scientific). Primer sequences are listed in S1 Table. Ct values were retained only if they were lower than those of both no-template and extraction blank controls. Total bacterial load was calculated via the ΔCt method.

Transcriptome profiling by high-throughput sequencing

Raw sequencing reads can be found in the Sequence Read Archive under the BioProject PRJNA1235228. For each condition, four samples were sequenced. Each sample was made by pooling RNA from five individuals (3 ng of total RNA per individual), except for DCV at 28 dpe, where the four pools were made from three individuals each. RNA-seq libraries were prepared from 15 ng of pooled RNA using an NEBNext Ultra II RNA Library Prep Kit for Illumina (New England Biolabs, E7770L) with NEBNext Multiplex Oligos for Illumina (Dual Index Primers Set 1) (New England Biolabs, E7600S). All sequencing was performed on an Illumina NextSeq 500 instrument using a NextSeq 500/550 High Output Kit v2.5 (75 cycles) (Illumina, 20024906). Raw sequencing reads were mapped to the D. melanogaster genome release dmel-r6.56 with STAR version 2.7.11b [69]. Feature counting was performed with HTSeq version 0.11.2. [70], and differential gene expression analysis was performed with DESeq2 in R Studio (version 2023.03.1). Differential gene expression analysis results can be found in S2 Data. Log₂ fold changes were transformed into Z-scores using the “ashr” package, and only adjusted p-values were considered for analysis.

Statistical analysis

All statistical analyses were conducted using R version 3.6.1 within the Rstudio development environment version 2024.04.2 + 764. The level of significance was set at P < 0.05 and symbols indicate the significance of differences between conditions: ***P < 0.001; **P < 0.01; *0.01 < P < 0.05.

To account for the block effect, we employed linear mixed-effects or general linear mixed-effects models. Generalized linear models (GLM) were fitted to the data using the glm function from the “stats” package. Generalized linear mixed models (GLMM) were fitted to the data using the lmer function from the “lme4” package version 1.1-35.5. CLMM using the clmm function from the “ordinal” package version 2023.12-4.1. Post-hoc pairwise comparisons were performed using the emmeans function from the “emmeans” package version 1.10.5. Tukey’s or FDR correction was used to adjust for multiple comparisons. The letter-based grouping for pair-wise comparisons was created using the cld function from the “multcomp” package version 1.4-26. The specific statistical methods applied to each figure are detailed in the legend of the corresponding figure.