The heat shock response plays a role in the immune defense against viruses across various organisms. Studies on model organisms show that inducing this response prior to viral exposure enhances host resistance to infections, while deficient responses make individuals more vulnerable. Moreover, viruses rely on components of the heat shock response for their own stability and viral infections improve thermal tolerance in plants, giving infected individuals an advantage in extreme conditions, which aids the virus in replication and transmission. Here, we examine the interaction between the nematode Caenorhabditis elegans and its natural pathogen the Orsay virus (OrV) under heat stress. We found that OrV infection leads to differential expression of heat-stress related genes, and infected populations showed increased resistance to heat shock. This resistance correlated with increased expression of argonautes alg-1 and alg-2, which are crucial for survival after heat shock and for OrV replication. Overall, our study suggests an environmental-dependent mutualistic relationship between the nematode and OrV, potentially expanding the animal’s ecological niche and providing the virus with extra opportunities for replication and adaptation to extreme conditions.

(a) Overlap of heat-stress related GO terms and OrV DEGs

The differentially expressed genes upon OrV infection in wild-type animals were obtained from. The genes belonging to the GO term categories (i) cellular response to unfolded proteins, (ii) heat-shock binding proteins and (iii) response to heat were extracted from WormBase Ontology Browser. Only genes annotated as directly involved in each category were used. Genes that were differentially expressed upon OrV infection and belonged to any of these three GO term categories were selected.

(b) Strain maintenance

Nematodes were maintained at 20 ºC on Nematode Growth Medium (NGM) plates seeded with Escherichia coli OP50 under standard conditions. ERT54 (jyIs8 [pals-5p::GFP; myo-2p::mCherry]X in an N2 Bristol background) served as wild-type. JU2624, (mjIs228[myo-2p::mCherry::unc54; lys-3p::eGFP::tbb-2] in a JU1580 background) was the natural isolate from which OrV was first identified. Both strains were generously provided by Prof. M.A. Félix.

(c) OrV stock preparation and quantification

JU2624 animals were inoculated with OrV isolate JUv1580 (a gift from Prof. M.A. Félix), allowed to grow for 5 days and then resuspended in M9 buffer (0.22 M KH₂PO₄, 0.42 M Na₂HPO₄, 0.85 M NaCl, 0.001 M MgSO₄), let stand for 15 min at room temperature, vortexed, and centrifuged for 2 min at 400 g. The supernatant was centrifuged twice at 21,000 g for 5 min and then passed through a 0.2 mm filter. RNA of the resulting viral stock was extracted using the Viral RNA Isolation kit (NZYTech). The concentration of viral RNA was determined by RT-qPCR and normalized using a standard curve. Primers used can be found in table 1.

For the standard curve cDNA of OrV was obtained using Accuscript High Fidelity Reverse Transcriptase (Agilent) and reverse primers at the 3’ end of the virus (see table 1 for primers). Approximately 1000 bp of the 3’ end of RNA1 and RNA2 were amplified using forward primers containing 20 bp coding the T7 promoter and DreamTaq DNA Polymerase (ThermoFisher). The PCR products were gel-purified using MSB Spin PCRapace (Invitek Molecular) and an in vitro transcription was performed using T7 Polymerase (Merck). The remaining DNA was then degraded using DNAse I (Life Technologies). RNA concentration was determined by NanoDrop (ThermoFisher) and the number of molecules per µL was determined using the online tool EndMemo RNA Copy Number Calculator (https://endmemo.com/bio/dnacopynum.php).

(d) Nematode synchronization and OrV inoculation

In order to obtain synchronized populations of wild-type and JU2624 animals, plates with eggs were carefully washed with M9 buffer to remove larvae and adults but leaving the eggs behind. Plates were washed again using M9 buffer after 1 h to collect larvae hatched within that time span. Synchronized nematode populations were inoculated with 4.92´10⁸ copies of OrV –a concentration that leads to activation of the pals-5p::GFP reporter in ~50% of the animals– by pipetting the viral stock on top of the bacterial lawn containing the animals. Twenty-four hours post-inoculation (hpi) pals-5p::GFP negative animals were manually removed from infected plates and negative control plates were checked to confirm absence of pals-5p::GFP activation.

(e) Heat-shock

Synchronized population were shifted from a 20 ºC incubator to a 37 ºC water bath at 48 hpi for 2 h. This heat-shock treatment results in ~50% mortality of control animals, and hereafter we will refer to it as semi-lethal. Plates were sealed with parafilm and placed bottom down in order to ensure a quick temperature shift. After the heat shock they were returned to 20 ºC (figure 1b).

(f) Mortality assay

Twenty-four hours post-heat-shock (hphs) mortality within the heat-shocked population was quantified. Animals were considered dead when they did not react to touch and had stopped pharyngeal pumping.

Mortality was evaluated in four and eight independent full blocks for JU2624 and wild-type populations, respectively. The number of plaques (biological replicates) within each experimental condition within blocks varied between 2 and 13 (median 3).  Mean and standard deviation of mortality were estimated for each infection condition, nematode strain and experimental block. Data were analyzed using a meta-analysis approach for continuous data, with infection status considered as a fixed factor, using Cohen’s d (±1 standard error) as a normalized estimation of the effect size, and restricted maximum-likelihood for parameter’s estimation. These analyses were done using SPSS version 29.0.1.1 (IBM Corp).

(g) Quantification of alg-1 and alg-2 expression

To quantify expression levels of alg-1 and alg-2, wild-type control and infected animals were used. Nematodes were heat-shocked as described above and samples were taken before heat-shock, right after the heat-shock, and 2 and 8 hphs. Animals were collected with PBS 0.05% Tween and washed 3 times before freezing in liquid N₂. RNA was extracted using Trizol as previously described. RT-qPCRs were performed using Power SYBR Green PCR Master Mix (Applied Biosystems) on an ABI StepOne Plus Real-time PCR System (Applied Biosystems). Ten ng of total RNA were loaded and samples were normalized to expression of cdc-42. Primers used can be found in table 1. Relative expression levels were computed using the DDC_T method.

Fold-change values for each mutant genotype were independently fitted to a generalized linear model with a Normal distribution and identity function with infection status and hpsh incorporated as orthogonal random factors. The interaction between the two factors is reported in the text as an evaluation of the differences in temporal expression dynamics. These analyses were also done with SPSS version 29.0.1.1.