Bacterial diet and weak cadmium stress affect the survivability of Caenorhabditis elegans and its resistance to severe stress.

Stress may have negative or positive effects in dependence of its intensity (hormesis). We studied this phenomenon in Caenorhabditis elegans by applying weak or severe abiotic (cadmium, CdCl2) and/or biotic stress (different bacterial diets) during cultivation/breeding of the worms and determining their developmental speed or survival and performing transcriptome profiling and RT-qPCR analyses to explore the genetic basis of the detected phenotypic differences. To specify weak or severe stress, developmental speed was measured at different cadmium concentrations, and survival assays were carried out on different bacterial species as feed for the worms. These studies showed that 0.1 μmol/L or 10 mmol/L of CdCl2 were weak or severe abiotic stressors, and that E. coli HT115 or Chitinophaga arvensicola feeding can be considered as weak or severe biotic stress. Extensive phenotypic studies on wild type (WT) and different signaling mutants (e.g., kgb-1Δ and pmk-1Δ) and genetic studies on WT revealed, inter alia, the following results. WT worms bred on E. coli OP50, which is a known cause of high lipid levels in the worms, showed high resistance to severe abiotic stress and elevated gene expression for protein biosynthesis. WT worms bred under weak biotic stress (E. coli HT115 feeding which causes lower lipid levels) showed an elevated resistance to severe biotic stress, elevated gene expression for the innate immune response and signaling but reduced gene expression for protein biosynthesis. WT worms bred under weak biotic and abiotic stress (E. coli HT115 feeding plus 0.1 μmol/L of CdCl2) showed high resistance to severe biotic stress, elevated expression of DAF-16 target genes (e.g., genes for small heat shock proteins) but further reduced gene expression for protein biosynthesis. WT worms bred under weak biotic but higher abiotic stress (E. coli HT115 feeding plus 10 μmol/L of CdCl2) showed re-intensified gene expression for the innate immune response, signaling, and protein biosynthesis, which, however, did not caused a higher resistance to severe biotic stress. E. coli OP50 feeding as well as weak abiotic and biotic stress during incubations also improved the age-specific survival probability of adult WT worms. Thus, this study showed that a bacterial diet resulting in higher levels of energy resources in the worms (E. coli OP50 feeding) or weak abiotic and biotic stress promote the resistance to severe abiotic or biotic stress and the age-specific survival probability of WT.

Introduction

Hormesis describes the intensity-dependent positive or negative effects of several stressors [1]. Low concentrations of the heavy metal copper cause, for instance, a higher reproduction rate in Campanularia flexuosa [2], and low quantities of polyphenol tannic acid (TA) have positive effects on the thermal tolerance and oxidative stress resistance of the model organism C. elegans [3]. Low concentrations of the heavy metal cadmium promote nitrate reductase activity in Hydrilla verticillata [4] but high cadmium concentrations may result in oxidative stress, protein damage, and disturbed DNA repair [5, 6]. Resistance mechanisms to severe cadmium stress include the antioxidative defense, for instance, by the glutathione redox system, protection and repair of proteins by heat shock proteins, binding of cadmium by metallothioneins, and export of cadmium by ATP binding cassette (ABC) transporters and associated proteins (glutathione, phytochelatin) [7, 8, 9].

The bacterial diet of C. elegans also bears the risk of ingesting pathogenic bacteria. However, feeding C. elegans with the pathogenic E. coli strain 536 revealed a hormetic effect (i.e., higher thermal tolerance due to the expression of small heat shock proteins; [10]). Even E. coli OP50 as standard bacterial diet of C. elegans in the laboratory [11] has been reported to be mildly pathogenic [12], especially for older worms [13, 14]. E. coli HT115 [15] as another bacterial diet of C. elegans in the laboratory (primarily for RNA interference assays; [16]) differs, for instance, from E. coli OP50 by a longer outer core of the lipopolysaccharides (LPS) in the outer membrane [17, 18, 19], which was reported to affect positively the life span of C. elegans wild type (WT) worms at a higher cultivation temperature (25 °C; [20]). However, Brooks et al. [21] did not find appreciable effects of different E. coli strains (including E. coli HT115) on the life span of WT at 20 °C, with a slightly higher maximum life span on E. coli OP50. These authors as well as Maier et al. [20] and Xiao et al. [22] showed a higher fat storage (triacylglycerol level) in WT on E. coli OP50 than on E. coli HT115. The feeding rate of WT was also higher on E. coli OP50 than on E. coli HT115 at 20 °C [22] and 25 °C [20]. The developmental speed of WT was reported to be either higher (at 25 °C; [20]) or lower (at 20 °C; [22]) on E. coli OP50 than on E. coli HT115. WT also produced a higher number of progeny on E. coli OP50 than on E. coli HT115 (at 25 °C; [20]). Thus, these two bacterial diets evoke substantial differences in the metabolism, behavior, development, and reproduction of WT, which seem to vary with temperature.

Important signaling and processing units for the cellular responses of C. elegans to abiotic or biotic stress are the PMK-1 and KGB-1 mitogen-activated protein kinase (MAPK) and DAF-2 insulin-like signaling pathways. The p38-type PMK-1 MAPK pathway promotes the heat or cadmium resistance of C. elegans [23, 24], with PMK-1 promoting in the latter case also DAF-16-mediated gene expression. PMK-1 signaling is also important for the innate immune response [25, 26, 27] by promoting the expression of immune response genes. PMK-1 target genes encode C-type lectins, CUB-like domain proteins, and ShK toxins [26]. The antimicrobial effect of C-type lectins was proven in mice [28] but C. elegans also showed an upregulation of several C-type lectins under pathogenic stress [29]. Even if possible functions or roles in pathogenic defense mechanisms of the CUB-like domain proteins and metridin-like ShK toxins are still under debate [26, 30, 31], an upregulation of corresponding genes was also detected in C. elegans under pathogenic stress [26]. The JNK-like KGB-1 MAPK pathway participates in the responses to heavy metal stress [32, 33, 34] via the transcription factor FOS-1 [35], and it is also involved in innate immune responses [25, 36]. Active DAF-2 signaling prevents the nuclear translocation of the transcription factor DAF-16. Inhibited DAF-2 signaling, however, results in the nuclear translocation of non-phosphorylated DAF-16 and DAF-16 target gene expression. These genes encode, for instance, small heat shock proteins (α-crystallins), metallothioneins (MTL-1/2), and antioxidant enzymes (superoxide dismutase 3) [37] or the lysozyme LYS-7 [26]. DAF-16 is also an essential life-prolonging transcription factor [38].

This study on C. elegans focused on the effects of weak abiotic and/or biotic stress on the age-specific survival probability and the resistance to severe abiotic or biotic stress. Weak as well as severe abiotic and biotic stressors were identified in developmental and survival assays. The effects of weak abiotic and/or biotic stress on the age-specific survival probability of adult wild type and two MAPK signaling mutants (kgb-1Δ, pmk-1Δ) were determined in long-term (over weeks) survival assays. The effects of weak abiotic and/or biotic stress during breeding on the later resistance of adult wild type to severe abiotic or biotic stress were determined in short-term (over days) survival assays. To explore the genetic basis of the effects of weak abiotic and/or biotic stress, transcriptome analyses were performed by RNA-Seq on wild type worms, which were bred under weak abiotic and/or biotic stress conditions. Results of the RNA-Seq study were checked by quantitative RT-PCR analyses. The results of this study showed that either a bacterial diet causing a higher level of energy resources in the worms or the exposure to weak abiotic and biotic stress significantly promote the age-specific survival probability and the resistance to severe abiotic or biotic stress of wild type.

Materials and methods

Experimental organisms

C. elegans of the N2 Bristol variety (wild type, WT) and the signaling mutants kgb-1Δ [KB3 kgb-1(um3) IV], pmk-1Δ [KU25, pmk-1(km25) IV], daf-2Δ [CB1370, daf-2(e1370) III], and daf-16Δ [CF1038, daf-16(mu86) I] as well as the Escherichia coli strain OP50 were obtained from the Caenorhabditis Genetics Center (CGC) (https://www.cbs.umn.edu/research/resources/cgc). The E. coli strain HT115 (empty vector L4440) was provided by Source BioScience LifeSciences (Nottingham, UK). Chitinophaga arvensicola was obtained from the DSMZ (DSM No. 3695; Braunschweig, Germany). The maintenance of C. elegans and the composition of C. elegans-specific media and buffers followed the instructions of [11]. Worms were bred at 20 °C on NGM (nematode growth medium) plates (petri dishes, ∅ = 9.2 cm) seeded with 1 mL of E. coli OP50 suspension [OD600nm = 1; LB (lysogeny broth) or OP50 medium, with the latter containing 0.01 g/L uracil (Sigma-Aldrich, Germany)]. All experiments were carried out at T = 20 °C as well. According to the German law, experiments carried out on the invertebrate C. elegans do not have to be announced or approved.

Development studies

WT worms were synchronized, and ten eggs were transferred onto NGM plates (∅ = 6 cm) seeded with 1 mL of E. coli OP50 suspension (OD600nm = 1; OP50 medium) and containing cadmium chloride (CdCl2 × 2½ H2O; Sigma-Aldrich, Germany) at final concentrations between 0 and 300 μmol/L. The number of worms of each developmental stage (L1, L2, L3, L4 larval stages, young adult and adult worms) was determined every day (always at the same time of day) by visual observation and counting.

Short-term survival assay

Synchronized L1 larvae of WT and the signaling mutants were transferred onto NGM plates (∅ = 9.2 cm) seeded with 1 mL of E. coli HT115 suspension (OD600nm = 1; LB medium). After reaching adulthood, worms were washed off the NGM plates with M9 buffer and transferred onto empty NGM plates to get rid of bacteria. Then, 15 adult worms were transferred onto NGM plates (∅ = 3.5 cm) seeded with 0.2 mL of E. coli OP50, E. coli HT115, or C. arvensicola suspension (OD600nm = 0.5; LB medium). Flight behavior was minimized by a thin layer of 10 mg/mL of palmitic acid (BioXtra, Sigma-Aldrich, Germany) around the edge of the petri dishes [39]. Survival was tested after 1, 2, 3, and 4 d (always at the same time of day) by applying weak mechanical stimuli to evoke body movements (touch response), with the worms transferred onto new NGM plates after 3 d to avoid a lack of food or overcrowding. Worms that had buried themselves into the NGM or climbed up the edge of the petri dishes were not considered.

Long-term survival assay

Synchronized L1 larvae of WT, kgb-1Δ, and pmk-1Δ were transferred onto NGM plates (∅ = 9.2 cm) seeded with 1 mL of E. coli HT115 suspension (OD595nm = 0.7; LB medium). The adult worms were washed off the NGM plates with M9 buffer and transferred onto empty NGM plates to remove bacteria. Then, 15 adult worms were transferred onto NGM plates (∅ = 6 cm) seeded with 0.5 mL of E. coli OP50 or E. coli HT115 suspension (OD595nm = 0.7; LB medium) and containing 0 or 0.1 μmol/L of CdCl2. Flight behavior was minimized as described above. Survival was tested daily (always at the same time of day) by eliciting the touch response, with the worms transferred onto new NGM plates every second day. Worms within the NGM or on the edge of the petri dishes were not considered.

Resistance to severe abiotic or biotic stress after breeding on varied CdCl2 concentrations and bacterial diets

Synchronized L1 larvae of WT were transferred onto NGM plates (∅ = 9.2 cm) seeded with 1 mL of E. coli OP50 or E. coli HT115 suspension (OD600nm = 1; LB medium) and containing 0 or 0.1 μmol/L of CdCl2 for the abiotic stress (i.e., 10 mmol/L of CdCl2) or 0, 0.1, or 10 μmol/L of CdCl2 for the biotic stress (i.e., C. arvensicola feeding). Adult worms were washed off the NGM plates with M9 buffer and transferred onto empty NGM plates to get rid of bacteria. For testing the resistance to severe abiotic stress, 15 adult worms were transferred onto NGM plates (∅ = 6 cm) seeded with 0.5 mL of E. coli OP50 or HT115 suspension (OD600nm = 1; LB medium) and containing 0 or 10 mmol/L of CdCl2. Survival was tested after 1 d by eliciting the touch response. For testing the resistance to severe biotic stress, 15 adult worms were transferred onto NGM plates (∅ = 3.5 cm) seeded with 0.2 mL of C. arvensicola suspension (OD600nm = 0.5; LB medium). Survival was tested after 4 d by eliciting the touch response, with the worms transferred onto new NGM plates after 3 d. Flight behavior was minimized in both cases as described above. Worms within the NGM or on the edge of the petri dishes were not considered.

RNA-Seq

Transcriptome profiling was carried out under five experimental conditions. WT worms were incubated for three weeks on NGM plates (∅ = 9.2 cm) seeded with 1 mL of E. coli (i) OP50 (abbr. OP50) or (ii) HT115 (abbr. HT115_1) suspension (OD600nm = 1; LB medium) or on NGM plates seeded with 1 mL of E. coli HT115 suspension and containing (iii) 0 (abbr. HT115_2), (iv) 0.1 (abbr. 0.1 Cd), or (v) 10 (abbr. 10 Cd) μmol/L of CdCl2. The worms were transferred onto new NGM plates twice a week. Subsequently, synchronized L1 larvae were transferred onto 15 different NGM plates (∅ = 9.2 cm) per experimental condition (see above). After reaching adulthood, worms were washed off from the NGM plates and cleaned several times using purified water to get rid of bacteria. Five plates per experimental condition were combined to one biological sample. Thus, three biological replicates per experimental condition were further prepared for RNA-Seq.

The following procedure has already been described in [23]. Briefly, RNAiso-G (Segenetic, Borken, Germany) was added to the samples, before they were frozen in liquid nitrogen. After thermal disruption of worms (liquid nitrogen, 35 °C; three repetitions), chloroform extraction on ice (10 min), and centrifugation (12.000 × g, 4 °C, 15 min), DNA was digested and RNA purified with a RNase-free DNase set and RNeasy® mini kit (Qiagen, Hilden, Germany). Quality control was performed using an Agilent Bioanalyzer® (Agilent Technologies, Böblingen, Germany). After adding the samples to RNAstableTM matrix (Biomatrica, San Diego, CA, USA) and a subsequent vacuum centrifigution for drying (2 h), the samples from three experimental approaches were sent to the Beijing Genomics Institute (BGI) for RNA-Seq analysis (OP50 and HT115_1; HT115_2 and 10 Cd; 0.1 Cd).

Using Illumina HiSeq2000 technology, samples were sequenced at the BGI with a minimum of 10 megareads per sample and a sequencing quality of more than 98% clean reads. Sequences were mapped to Wormbase release WS257. Mean expression intensities of genes (i.e., means from the three biological replicates per experimental condition, after checking for substantial deviations between corresponding data; see below) and differential gene expression were calculated with the FPKM method (fragments per kilobase of exon per million fragments mapped) to normalize for sequencing depth and gene length. The NOISeq method [40] was used to determine the threshold of P (diverge probability ≥ 0.8) for the differentially expressed genes (DEGs). In case of OP50, HT115_2, 0.1 Cd, and 10 Cd, the three biological replicates at each test condition showed a high degree of similarity with regard to the expression intensity of corresponding genes, and mean expression intensities were calculated using the data from all three biological replicates. In case of HT115_1, however, one biological replicate differed significantly from the two others regarding the expression intensity of corresponding genes, and mean expression intensities were calculated here using the data from only two biological replicates.

Quantitative reverse transcription PCR (RT-qPCR)

Samples for RNA-Seq (before matrix embedment) were also used for RT-qPCR analyses (see [24] for details). Briefly, RNA concentrations were measured using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Darmstadt, Germany). After reverse transcription of 0.5 μg of total RNA per sample using oligo (dT)18-primers (First Strand cDNA Synthesis Kit; Fermentas, St. Leon-Rot, Germany) and RevertAid Reverse Transcriptase (Thermo Fisher Scientific), quantitative PCR was performed using a Real-Time PCR System (Eco Real-Time PCR; Illumina) and the PerfeCTa SYBR® Green SuperMix Kit (Quanta Biosciences, Beverly, MA, USA). Each sample contained 1 μL of template cDNA, 5 μL of SYBR Green SuperMix, 1 μL of the specific primer pair (10 mmol/L each; see Table 1 for the sequences), and 3 μL of purified water, resulting in an end volume of 10 μL. Then, the following procedure was applied: 2 min at 95 °C (initial phase), 40 cycles in the order 1 min at 95 °C, 10 s at 95 °C; 30 s at 60 °C, followed by a melting curve (between 60 °C and 95 °C). Data analysis was accomplished via ΔCq calculations [41]. Ratios were calculated between the mRNA levels of the genes of interest and the housekeeping gene cdc-42 [42].

Table 1

Primer sequences used in qRT-PCR. Primers were designed using the PerlPrimer tool [60].

Gene ID	Public name	Forward primer	Reverse Primer
WBGene00021580	clec-174	GTAACTACTTGGCAGTCTTCG	CTTGCCATTATCGTCCAGTG
WBGene00021224	clec-209	CTTCTTAGCATCAATCGTTCGT	TTTGTGAAATCAGGTGTGGAC
WBGene00020945	clec-219	AAGAAGAACCAACACCACCT	ACAGAAGCAAACATACACCAC
WBGene00018971	clec-67	TGATGGTGACAGTTCAAAGC	TTCCAAAAATGCCCGAGTAG
WBGene00016577	clec-3	ACTGTTCGATGTGAAGAATGC	GTAGCAGGGATCCTCATGAG
WBGene00013931	clec-97	AGACTTGCGATACTGGATGG	TGATCCAGATTGATTACCAAGG
WBGene00008916	clec-221	AATGGGATGATCAACTGACTG	AGGAAGTGAGCAATCTTTGTC
WBGene00008202	clec-197	GTGCATCAAGGTTTACCCAG	TCTTCCACATATCTCTTCTCGG
WBGene00000390	cdc-42	TTCGACAATTACGCCGTCAC	CCTGAGATCGACTTGAGTACC

Bioinformatics

David Bioinformatics Resources 6.8 (https://david.ncifcrf.gov/tools.jsp) was used for GO (gene ontology) analyses (i.e., functional annotation charts). Cluster analysis was carried out using the program Cluster 3.0 (http://bonsai.hgc.jp/∼mdehoon/software/cluster/software.htm). For this program, the common DEGs of the contrasts E. coli HT115 vs. E. coli OP50 feeding, 0.1 vs. 0 μmol/L of CdCl2, and 10 vs. 0.1 μmol/L of CdCl2, with E. coli HT115 as bacterial diet in the latter two cases, were initially determined. After that, gene clustering was performed, with the WormBase identification numbers (WB GeneIDs) and mean expression intensities of the 1607 common DEGs in the five differently treated worm groups (HT115_1, OP50, 0.1 Cd, 10 Cd, HT115_2; see above) as starting values. The data were log2-transformed, and the genes and arrays were separately centered (using the medians) and normalized. Hierarchical clustering of the genes and arrays was performed using the similarity metric ‘correlation (uncentered)’ and the clustering method ‘average linkage’. The dendrogram was created with the program Java TreeView 1.1.6r4 (https://sourceforge.net/projects/jtreeview/).

Statistics and computations

Data are given as mean ± standard deviation (sd) or as mean of means ± standard error (se). n indicates the number of biological replicates and N the number of animals used. SigmaPlot 11.0 (Systat Software, Erkrath, Germany) was used for graph preparations and statistical analysis. Two-way analysis of variance (anova, including the Holm-Sidak all pairwise multiple comparison procedure) or t-tests were used to identify significant differences (P ≤ 0.05, 0.01, or 0.001) in survival or mRNA level.

Results

Abiotic and biotic stress influence the developmental speed and short-term survival probability of worms

The developmental speed of wild type (WT) from egg to adult worm was studied on nematode growth medium (NGM) plates containing CdCl2 at concentrations between 0 and 300 μmol/L (Fig. 1a–j). A CdCl2 concentration of 50 μmol/L already induced a slight developmental delay (Fig. 1b) in comparison to the control (Fig. 1a). At CdCl2 concentrations of more than 175 μmol/L, the worms rarely or never reached adulthood (Fig. 1g–j). These results also suggest that an exposure to a CdCl2 concentration of 0.1 μmol/l causes only a weak abiotic stress.

Fig. 1

The effects of different CdCl2 concentrations or bacterial diets on the developmental speed or short-term survival probability of worms. The developmental speed of wild type from egg to adult worm was determined (a–j) at different CdCl2 concentrations (0–300 μmol/L of CdCl2) in the NGM plates at 20 °C, using E. coli OP50 as bacterial diet (per CdCl2 concentration, n = 3 biological replicates on N = 10 worms each). The diameter of the symbols marks the relative number of worms of a specific developmental stage, which was counted from day 0 (egg deposition) to day 8. The diameter of the symbols at day 0 corresponds to 100% (i.e., 10 worms). The survival probability of synchronized adult (k) wild type, (l) kgb-1Δ, (m) pmk-1Δ, (n) daf-2Δ, and (o) daf-16Δ worms, which were fed with E. coli OP50 (gray circles), E. coli HT115 (white circles), or C. arvensicola (red diamonds), was determined over a period of four days at 20 °C (mean ± sd; per strain and bacterial diet, n = 9–12 biological replicates on N = 15 worms each). Dashed curves show the survival probability of wild type on C. arvensicola as bacterial diet (from k). Asterisks mark significant differences after four days of incubation (∗P ≤ 0.05, ∗∗P ≤ 0.01, P∗∗∗ ≤ 0.001; two-tailed t-test).

Feeding synchronized adult wild type and the synchronized signaling mutants kgb-1Δ, pmk-1Δ, daf-2Δ, and daf-16Δ with E. coli OP50, E. coli HT115, or C. arvensicola revealed negative effects of C. arvensicola feeding on the survival probability after four days of incubation (Fig. 1k–o). The mutants kgb-1Δ (Fig. 1l) and pmk-1Δ (Fig. 1m) additionally showed a severely reduced survival probability in comparison to wild type (Fig. 1k) on any bacterial diet. The negative effect of C. arvensicola feeding was lowest in daf-2Δ, whose survival probability was maximum on E. coli OP50 (Fig. 1n). The survival probability of daf-16Δ (Fig. 1o) was slightly below that of wild type at any bacterial diet. These results suggest that C. arvensicola feeding causes a severe biotic stress, whereas E. coli HT115 feeding may represent a weak biotic stress in comparison to E. coli OP50 feeding.

Weak abiotic and/or biotic stress affect the age-specific survival probability of worms

Determining the survival probability of synchronized adult wild type (Fig. 2a–e), kgb-1Δ (Fig. 2f–j), and pmk-1Δ (Fig. 2k–o) at CdCl2 concentrations of 0 or 0.1 μmol/L in the NGM and on E. coli OP50 or E. coli HT115 as bacterial diet revealed differential effects of these experimental conditions on the age-specific survival probability during the remaining life span of these strains. Although there were distinct strain-specific differences in life span, the different experimental conditions had no effect on maximum life span.

Fig. 2

CdCl2 concentration and bacterial diet influence the age-specific survival probability of wild type, kgb-1Δ, and pmk-1Δ. The age-specific survival probability of synchronized adult (a–e) wild type, (f–j) kgb-1Δ, and (k–o) pmk-1Δ worms, which were incubated on NGM plates containing either no CdCl2 (circles and/or triangles) or 0.1 μmol/L of CdCl2 (squares and/or triangles) and fed with E. coli OP50 (gray symbols) or E. coli HT115 (white symbols) was determined over periods of up to 30 (wild type), 18 (kgb-1Δ) or 21 (pmk-1Δ) days at 20 °C (mean ± sd; per strain, CdCl2 concentration, and bacterial diet, n = 3–6 biological replicates on N = 15 worms each). Asterisks mark significant differences between whole curves (∗∗P ≤ 0.01; P∗∗∗ ≤ 0.001; two-way anova), with triangle symbols indicating significant differences on specific days of incubation (P ≤ 0.05; Holm-Sidak all pairwise multiple comparison procedure).

The age-specific survival probability of wild type worms was higher on E. coli OP50 than on E. coli HT115 (Fig. 2a) and on E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM than on E. coli HT115 without CdCl2 in the NGM (Fig. 2c). Conversely, the age-specific survival probability of kgb-1Δ worms was lower on E. coli OP50 than on E. coli HT115 (Fig. 2f). It was also lower on E. coli OP50 without or with 0.1 μmol/L of CdCl2 in the NGM than on E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM (Fig. 2i, j). The age-specific survival probability of pmk-1Δ worms was also lower on E. coli OP50 than on E. coli HT115 (Fig. 2k), but it was higher on E. coli OP50 without or with 0.1 μmol/L of CdCl2 in the NGM than on E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM (Fig. 2n, o). In contrast to wild type, pmk-1Δ showed a higher age-specific survival probability on E. coli HT115 (without CdCl2 in the NGM) than on E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM. Thus, the differences in cadmium exposure and/or bacterial diet caused quite different effects on the age-specific survival probability of WT, kgb-1Δ, or pmk-1Δ.

Breeding under weak abiotic and/or biotic stress influence the stress resistance of wild type

After breeding wild type from egg to adult worm at different CdCl2 concentrations in the NGM (0, 0.1, or 10 μmol/L) on E. coli OP50 or E. coli HT115 as bacterial diet, the adult worms were exposed for one day to severe abiotic stress (10 mmol/L of CdCl2, with the respective bacterial diet remaining unchanged) or for four days to severe biotic stress (C. arvensicola feeding). The resistance to severe abiotic stress was higher on E. coli OP50 than on E. coli HT115 but also higher after breeding at 0.1 μmol/L of CdCl2 in the NGM in case of E. coli HT115 feeding (Fig. 3a). The resistance to severe biotic stress was maximum after breeding at 0.1 μmol/L of CdCl2 and lower after breeding at 0 or 10 μmol/L of CdCl2, with the positive effect of 0.1 μmol/L of CdCl2 particularly strong after breeding on E. coli HT115 (Fig. 3b). Thus, differences in cadmium exposure and/or bacterial diet during breeding affected the stress resistance of WT.

Fig. 3

Breeding on varied CdCl2 concentrations and bacterial diets affect the resistance of wild type to severe abiotic or biotic stress. Wild type developed from egg to adult worm on NGM plates containing 0, 0.1, or 10 μmol/L of CdCl2, with either E. coli OP50 (gray bars) or E. coli HT115 (white bars) as bacterial diet. Afterwards, the stress resistance of the adult worms (mean ± sd) was determined after (a) one day on NGM plates containing 10 mmol/L of CdCl2 (seeFig. 1a–j), with the specific bacterial diet remaining unchanged (per CdCl2 concentration and bacterial diet during breeding, n = 9 biological replicates on N = 15 worms each) or (b) four days with C. arvensicola as bacterial diet (seeFig. 1k–o) (per CdCl2 concentration and bacterial diet during breeding, n = 10–16 biological replicates on N = 15 worms each) at 20 °C. Asterisks mark significant differences (∗P ≤ 0.05, ∗∗P ≤ 0.01, P∗∗∗ ≤ 0.001; two-tailed t-test).

Breeding under weak abiotic and/or biotic stress affect gene expression in wild type

To explore the genetic basis of these results, transcriptome profiling by RNA-Seq was performed on synchronized adult wild type worms. For this, worms were long-term incubated and bred with E. coli OP50 or E. coli HT115 as bacterial diet to study effects of the different feeding conditions or with 0, 0.1, or 10 μmol/L of CdCl2 in the NGM and E. coli HT115 as bacterial diet to study effects of the different cadmium concentrations. More than 18,000 genes were identified and carried a WormBase identification number (WB GeneID). Genes termed as differentially expressed genes (DEGs) showed a diverge probability of greater than 0.8. To identify higher regulated genes or DEGs, the mean expression intensities (mean FPKM values) of corresponding genes or DEGs from different experimental conditions, which carried identical WB GeneIDs, were plotted against each other (x-y plots) (Fig. 4 and Supplementary Fig. S1). Higher regulated groups of genes or DEGs showed substantial deviations from the 45-degree diagonal line in double linear plots (left graphs) or distinct log2-fold differences in expression intensity in double log2 plots (right graphs), with the two scale types chosen to emphasize genes with higher (double linear plots) or lower expression intensity (double log2 plots). The function of these groups of genes or DEGs was studied by gene ontology (GO) analyses (functional annotation chart; David Bioinformatics Resources 6.8) (Table 2 and Supplementary Table S1).

Fig. 4

Breeding on varied CdCl2 concentrations and bacterial diets affect gene expression in wild type. Transcriptome profiling was performed by RNA-Seq to determine gene expression intensities in differently treated synchronized adult wild type worms (mean FPKM values; per CdCl2 concentration and bacterial diet, n = 2–3 biological replicates, see Materials and methods). For this, an initial long-term incubation (for three weeks) and the subsequent development from egg to adult worm took place on NGM plates (at 20 °C) containing either (a, b) no CdCl2, with E. coli OP50 or E. coli HT115 as bacterial diet, or (c, d) 0 (control) or 0.1 μmol/L of CdCl2, (e, f) 0 (control) or 10 μmol/L of CdCl2, and (g, h) 0.1 μmol/L or 10 μmol/L of CdCl2, with E. coli HT115 as bacterial diet in c–h. The x-y plots show the mean expression intensities of all genes, which carried a WormBase identification number (WB GeneID), under two different experimental conditions as double linear (left graphs) or double log2 (right graphs) plots to emphasize genes with higher or lower expression intensity. Highly regulated groups of genes with clearly different expression intensities under the two experimental conditions can be found in the gene clusters, which deviate substantially from the 45-degree diagonal line in the double linear plots and show expression intensities of greater than 1,000 (red or green circles) or differ in their log2-fold expression intensities by values of greater than 2 (red circles) or lesser than -2 (green circles) between control (x-axis) and test (y-axis) conditions in the double log2 plots and show log2-fold expression intensities of lesser than 10. Continuous lines mark the 45-degree diagonal lines, with the dashed lines (right graphs) indicating offsets of 2 or -2.

Table 2

Gene ontology analysis on groups of identified genes. Gene ontology (GO) analyses (functional annotation chart; David Bioinformatics Resources 6.8) were carried out on the highly regulated groups of identified genes shown in the X-Y plots of Fig. 4 (red and green circles). The columns show (from left to right): section (S) of the table, type of contrast, number of identified genes (or unknown genes), annotation (A) source (G, GOTERM_BP_Direct; I: Interpro; U, UP_Keywords), term (Greek characters indicate specific gene groups; see Results), number (count) of genes assigned to this term, the percentage share of these genes, Fisher Exact P-value (EASE Score), Q-value (Benjamini). Background colors indicate the respective group of genes. Only dominant terms are shown.

S	Contrast	Genes	A	Term	Count	%	P	Q
E. coliHT115 vs.E. coliOP50
a	Fig. 4a, red circles (up. on E. coli OP50)	137	U	Ribosomal proteinα	77	56.2	7.7E-133	6.3E-131
			U	Ribonucleoproteinα	77	56.2	1.0E-124	4.0E-123
			G	GO:0006412: Translationα	77	56.2	1.7E-96	3.1E-94
			G	GO:0000003: Reproduction	96	70.1	7.6E-44	6.7E-42
			G	GO:0002119: Nematode larval development	90	65.7	5.9E-41	3.5E-39
			G	GO:0009792: Embryo development ending in birth or egg hatching	101	73.7	6.2E-36	2.7E-34
			G	GO:0006915: Apoptotic process	49	35.8	1.8E-34	6.2E-33
			G	GO:0002181: Cytoplasmic translationα	14	10.2	2.5E-21	7.5E-20
			G	GO:0018996: Molting cycle, collagen and cuticulin-based cuticle	23	16.8	8.2E-13	2.1E-11
			U	RNA-binding	14	10.2	6.5E-11	1.8E-09
			G	GO:0008340: Determination of adult lifespan	36	26.3	9.3E-11	2.0E-09
			U	Collagen	14	10.2	1.7E-10	3.5E-09
b	Fig. 4b, green circles (up. on E. coli HT115)	1013	I	IPR016187: C-type lectin foldβ	53	5.2	9.9E-21	3.4E-18
			I	IPR016186: C-type lectin-likeβ	47	4.6	1.0E-17	1.8E-15
			I	IPR001304: C-type lectinβ	36	3.6	1.5E-11	1.7E-09
			I	IPR003326: TRA-1 regulated	14	1.4	4.2E-09	3.5E-07
			I	IPR006583:PAN-3 domain	15	1.5	1.5E-07	1.0E-05
	.		G	GO:0007606: Sensory perception of chemical stimulusγ	22	2.2	6.5E-08	1.4E-05
			G	GO:0050907: Detection of chemical stimulus involved in sensory perceptionγ	27	2.7	1.4E-06	1.5E-04
			G	GO:0050911: Detection of chemical stimulus involved in sensory perception of smellγ	22	2.2	1.1E-05	5.9E-04
			G	GO:0042048: Olfactory behaviorγ	23	2.3	9.5E-06	6.7E-04
			G	GO:0043401: Steroid hormone mediated signaling pathway	22	2.2	3.6E-05	1.5E-03
			G	GO:0007186: G-protein coupled receptor signaling pathwayγ	38	3.8	5.6E-05	2.0E-03
			I	IPR019422: 7TM GPCR, serpentine receptor class h (Srh)γ	23	2.3	5.0E-04	2.8E-02
			I	IPR000536: Nuclear hormone receptor, ligand-binding, core	22	2.2	7.0E-03	2.9E-01
c	Fig. 4b, red circles (up. on E. coli OP50)	500 (3)	G	GO:0042048: Olfactory behaviorγ	18	3.6	6.4E-07	7.7E-05
			G	GO:0050911: Detection of chemical stimulus involved in sensory perception of smellγ	17	3.4	1.3E-06	7.7E-05
			I	IPR019428: 7TM GPCR, serpentine receptor class r (Str)γ	19	3.8	5.7E-07	1.1E-04
			I	IPR001304: C-type lectinβ	16	3.2	3.6E-05	3.4E-03
			G	GO:0050907: Detection of chemical stimulus involved in sensory perceptionγ	16	3.2	9.2E-05	3.7E-03
			I	IPR016187: C-type lectin foldβ	17	3.4	8.9E-05	5.6E-03
			I	IPR016186: C-type lectin-likeβ	16	3.2	1.2E-04	5.9E-03
			G	GO:0007186: G-protein coupled receptor signaling pathwayγ	23	4.6	3.5E-04	1.0E-02
			G	GO:0043401: Steroid hormone mediated signaling pathway	12	2.4	2.8E-03	6.5E-02
E. coli HT115 with 0.1 μmol/L of CdCl2vs. E. coli HT115
d	Fig. 4c, red circles (up. on E. coli HT115)	141 (4)	U	Ribosomal proteinα	73	51.8	1.2E-121	8.3E-120
			U	Ribonucleoproteinα	73	51.8	4.1E-114	1.4E-112
			G	GO:0006412: Translationα	70	49.6	4.5E-87	6.1E-85
			G	GO:0002119: Nematode larval development	82	58.2	2.9E-37	2.0E-35
			G	GO:0000003: Reproduction	82	58.2	5.9E-34	2.7E-32
			G	GO:0006915: Apoptotic process	42	29.8	3.0E-28	1.0E-26
			G	GO:0009792: Embryo development ending in birth or egg hatching	86	61.0	8.4E-27	2.3E-25
			G	GO:0002181: Cytoplasmic translationα	14	9.9	7.3E-22	1.6E-20
			G	GO:0018996: Molting cycle, collagen and cuticulin-based cuticle	23	16.3	1.1E-13	2.2E-12
			U	Collagen	15	10.6	1.8E-11	4.1E-10
			I	IPR002486: Nematode cuticle collagen, N-terminal	15	10.6	1.7E-10	4.5E-08
			G	GO:0008340: Determination of adult lifespan	25	17.7	2.0E-05	3.5E-04
e	Fig. 4d, green circles (up. at 0.1 μmol of CdCl2)	255	I	IPR027401: Myosin-like IQ motif-containing domain	3	1.2	2.8E-03	1.7E-01
			I	IPR019420: 7TM GPCR, serpentine receptor class bc (Srbc)γ	6	2.4	1.1E-03	1.9E-01
			I	IPR004009: Myosin, N-terminal, SH3-like	3	1.2	2.2E-03	1.9E-01
			I	IPR002928: Myosin tail	3	1.2	2.2E-03	1.9E-01
			G	GO:0050907: Detection of chemical stimulus involved in sensory perceptionγ	9	3.5	2.1E-03	2.2E-01
			I	IPR003598: Immunoglobulin subtype 2	4	1.6	1.1E-02	3.0E-01
f	Fig. 4d, red circles (up. on E. coli HT115)	2136 (2)	I	IPR000535: Major sperm protein	79	3.7	1.2E-57	5.8E-55
			I	IPR008962: PapD-like	79	3.7	1.0E-56	2.4E-54
			I	IPR002486: Nematode cuticle collagen, N-terminal	85	4.0	2.8E-47	4.5E-45
			I	IPR008160: Collagen triple helix repeat	67	3.1	8.6E-34	1.0E-31
			G	GO:0008360: Regulation of cell shape	41	1.9	7.9E-23	2.5E-20
			G	GO:0006470: Protein dephosphorylationδ	54	2.5	1.6E-22	2.5E-20
			G	GO:0018105: Peptidyl-serine phosphorylationδ	41	1.9	1.5E-17	1.7E-15
			G	GO:0035335: Peptidyl-tyrosine dephosphorylationδ	34	1.6	2.9E-14	2.3E-12
			I	IPR000242: Protein-tyrosine phosphatase, receptor/non-receptor typeδ	34	1.6	2.3E-13	2.2E-11
			I	IPR003595: Protein-tyrosine phosphatase, catalyticδ	32	1.5	2.1E-11	1.4E-09
			I	IPR006186: Serine/threonine-specific protein phosphatase/bis (5-nucleosyl)-tetraphosphataseδ	22	1.0	1.9E-11	1.5E-09
			G	GO:0006468: Protein phosphorylationδ	62	2.9	3.4E-09	2.2E-07
E. coli HT115 with 10 μmol/L of CdCl2vs. E. coli HT115
g	Fig. 4e, red circles (up. on E. coli HT115)	117 (5)	U	Ribosomal proteinα	58	49.6	2.2E-93	1.5E-91
			U	Ribonucleoproteinα	58	49.6	7.9E-88	2.8E-86
			G	GO:0006412: Translationα	56	47.9	1.5E-68	2.1E-66
			G	GO:0002119: Nematode larval development	67	57.3	5.6E-31	3.9E-29
			G	GO:0000003: Reproduction	68	58.1	2.4E-29	1.1E-27
			G	GO:0006915: Apoptotic process	32	27.4	1.2E-20	4.2E-19
			G	GO:0009792: Embryo development ending in birth or egg hatching	68	58.1	1.8E-20	5.2E-19
			G	GO:0018996: Molting cycle, collagen and cuticulin-based cuticle	17	14.5	1.4E-09	3.3E-08
			G	GO:0008340: Determination of adult lifespan	22	18.8	1.8E-05	3.7E-04
h	Fig. 4f, green circles (up. at 10 μmol of CdCl2)	477	I	IPR016187: C-type lectin foldβ	26	5.5	1.4E-10	3.7E-08
			I	IPR016186: C-type lectin-likeβ	24	5.0	8.7E-10	1.1E-07
			I	IPR001304: C-type lectinβ	22	4.6	3.5E-09	3.0E-07
			I	IPR006149: EB domain	8	1.7	4.2E-08	2.7E-06
			G	GO:0030968: Endoplasmic reticulum unfolded protein responseε	11	2.3	3.8E-07	6.0E-05
			I	IPR003582: Metridin-like ShK toxinβ	12	2.5	1.4E-05	7.2E-04
			I	IPR003366: CUB-like domainβ	8	1.7	4.9E-05	2.1E-03
			G	GO:0006952: Defense responseε	11	2.3	3.8E-05	3.0E-03
			I	IPR002347: Glucose/ribitol dehydrogenase	9	1.9	1.7E-04	6.0E-03
			G	GO:0045087: Innate immune responseβ	16	3.4	3.2E-04	1.7E-02
i	Fig. 4f, red circles (up. on E. coli HT115)	724 (1)	G	GO:0007186: G-protein coupled receptor signaling pathwayγ	45	6.2	3.9E-14	5.1E-12
			G	GO:0050911: Detection of chemical stimulus involved in sensory perception of smellγ	25	3.5	3.8E-11	1.7E-09
			G	GO:0050907: Detection of chemical stimulus involved in sensory perceptionγ	28	3.9	3.0E-11	2.0E-09
			G	GO:0042048: Olfactory behaviorγ	25	3.5	1.5E-10	5.0E-09
			G	GO:0007606: Sensory perception of chemical stimulusγ	18	2.5	5.0E-08	1.3E-06
			I	IPR019428: 7TM GPCR, serpentine receptor class r (Str)γ	23	3.2	5.3E-07	1.2E-04
			I	IPR019422: 7TM GPCR, serpentine receptor class h (Srh)γ	20	2.8	5.2E-05	6.2E-03
			I	IPR012885: F-box associated domain, type 2	15	2.1	4.3E-04	3.4E-02
			I	IPR001810: F-box domain, cyclin-like	22	3.0	1.2E-03	4.5E-02
			I	IPR017452: GPCR, rhodopsin-like, 7TMγ	25	3.5	1.1E-03	4.9E-02
			I	IPR019427: 7TM GPCR, serpentine receptor class w (Srw)γ	13	1.8	1.0E-03	6.0E-02
E. coli HT115 with 10 μmol/L of CdCl2vs. E. coli HT115 with 0.1 μmol/L of CdCl2
j	Fig. 4g, green circles (up. at 10 μmol of CdCl2)	129	U	Ribosomal proteinα	71	55.0	1.2E-120	1.3E-117
			U	Ribonucleoproteinα	71	55.0	2.2E-113	2.3E-110
			G	GO:0006412: Translationα	72	55.8	1.3E-90	1.5E-87
			G	GO:0002119: Nematode larval development	81	62.8	9.6E-36	1.1E-32
			G	GO:0000003: Reproduction	83	64.3	1.5E-34	1.8E-31
			G	GO:0006915: Apoptotic process	45	34.9	1.4E-31	1.6E-28
			G	GO:0009792: Embryo development ending in birth or egg hatching	89	69.0	3.9E-29	4.6E-26
			G	GO:0018996: Molting cycle, collagen and cuticulin-based cuticle	22	17.1	1.3E-12	1.5E-09
			U	RNA-binding	14	10.9	3.0E-11	3.2E-08
			U	Collagen	14	10.9	8.0E-11	8.4E-08
			I	IPR002486: Nematode cuticle collagen, N-terminal	14	10.9	1.2E-09	1.6E-06
			G	GO:0008340: Determination of adult lifespan	29	22.5	1.9E-07	2.2E-04
k	Fig. 4h, green circles (up. at 10 μmol of CdCl2)	1799 (1)	I	IPR002486: Nematode cuticle collagen, N-terminal	83	4.6	5.6E-49	2.5E-46
			I	IPR000535: Major sperm protein	69	3.8	1.1E-47	2.5E-45
			I	IPR008962: PapD-like	69	3.8	5.7E-47	8.6E-45
			I	IPR008160: Collagen triple helix repeat	68	3.8	6.7E-38	7.6E-36
			G	GO:0006470: Protein dephosphorylationδ	53	2.9	6.1E-23	1.8E-20
			G	GO:0008360: Regulation of cell shape	39	2.2	9.2E-22	1.4E-19
			G	GO:0035335: Peptidyl-tyrosine dephosphorylationδ	37	2.1	1.0E-17	1.0E-15
			I	IPR000242: Protein-tyrosine phosphatase, receptor/non-receptor typeδ	37	2.1	1.8E-17	1.6E-15
			G	GO:0018105: Peptidyl-serine phosphorylationδ	39	2.2	8.5E-17	8.4E-15
			I	IPR003595: Protein-tyrosine phosphatase, catalyticδ	34	1.9	2.6E-14	2.0E-12
			I	IPR003582: Metridin-like ShK toxinβ	34	1.9	6.1E-12	3.9E-10
			G	GO:0045087: Innate immune responseβ	49	2.7	2.3E-09	1.4E-07
			I	IPR001304: C-type lectinβ	43	2.4	3.2E-08	1.8E-06
			I	IPR016186: C-type lectin-likeβ	46	2.6	3.7E-08	1.8E-06
			I	IPR016187: C-type lectin foldβ	48	2.7	5.8E-08	2.6E-06
			G	GO:0006468: Protein phosphorylationδ	55	3.1	2.1E-07	1.1E-05
			I	IPR011009: Protein kinase-like domainδ	68	3.8	5.1E-07	2.1E-05
			I	IPR000719: Protein kinase, catalytic domainδ	58	3.2	5.2E-06	2.0E-04
			I	IPR017441: Protein kinase, ATP binding siteδ	40	2.2	2.1E-05	7.4E-04
l	Fig. 4h, red circles (up. at 0.1 μmol of CdCl2)	375 (1)	G	GO:0050907: Detection of chemical stimulus involved in sensory perceptionγ	14	3.7	1.9E-05	2.0E-03
			G	GO:0050911: Detection of chemical stimulus involved in sensory perception of smellγ	12	3.2	4.8E-05	2.5E-03
			G	GO:0042048: Olfactory behaviorγ	12	3.2	8.7E-05	3.1E-03
			G	GO:0007186: G-protein coupled receptor signaling pathwayγ	19	5.1	9.8E-05	2.6E-03
			I	IPR019428: 7TM GPCR, serpentine receptor class r (Str)γ	12	3.2	3.7E-04	7.0E-02
			I	IPR001436: Alpha crystallin/Heat shock proteinε	4	1.1	1.6E-03	1.4E-01
			I	IPR002068: Alpha crystallin/Hsp20 domainε	4	1.1	1.8E-03	1.1E-01
			I	IPR008978: HSP20-like chaperoneε	4	1.1	3.8E-03	1.7E-01
			I	IPR019422: 7TM GPCR, serpentine receptor class h (Srh)γ	10	2.7	6.8E-03	2.3E-01
			G	GO:0009408: Response to heatε	5	1.3	6.8E-03	1.4E-01

The gene or DEG groups identified in the double linear plots (Figs. 4 and S1, left graphs, colored symbols) commonly comprised genes or DEGs for translational processes, with annotation terms like ribosomal protein or translation (Tables 2 and S1; terms marked with the Greek character α). Genes or DEGs for translational processes were negatively affected by E. coli HT115 feeding in comparison to E. coli OP50 feeding (Figs. 4a and S1a) and particularly by 0.1 μmol/L of CdCl2 but also by 10 μmol/L of CdCl2 in comparison with E. coli HT115 feeding without CdCl2 in the NGM (Fig. 4c, e and S1c). Contrasting the two different cadmium concentrations showed that the expression of genes or DEGs for translational processes was more negatively affected by 0.1 than by 10 μmol/L of CdCl2 (Figs. 4g and S1g).

The gene or DEG groups identified in the double log2 plots (Figs. 4 and S1, right graphs, colored symbols) comprised genes, which were frequently assigned to terms (see Tables 2 and S1, terms marked with Greek characters) that are likely related to innate immune response (β) (e.g., C-type lectin, CUB-like domain, metridin-like ShK toxin), sensory perception (γ) (e.g., detection of chemical stimulus, G-protein coupled receptor signaling, olfactory behavior, 7-TM GPCR), signaling (δ) (e.g., de-/phosphorylation), or stress response (ε) (e.g., α-crystallin, ER unfolded protein response). Genes or DEGs associated with innate immune response were upregulated on E. coli HT115 (Table 2: b; Table S1: b, g, h), E. coli OP50 (Table 2: c), and E. coli HT115 with 10 μmol/L of CdCl2 (Table 2: h, k; Table S1: i, l, m). Genes or DEGs associated with signaling were upregulated on E. coli HT115 (Table 2: f; Table S1: b, h) and E. coli HT115 with 10 μmol/L of CdCl2 (Table 2: k; Table S1: l). Genes or DEGs associated with sensory perception were upregulated on E. coli HT115 (Table 2: b, i), E. coli OP50 (Table 2: c), and E. coli HT115 with 0.1 μmol/L of CdCl2 (Table 2: e, l). Genes or DEGs associated with stress responses were upregulated on E. coli HT115 (Table S1: b, h), E. coli HT115 with 0.1 (Table 2: l; Table S1: e, o) or 10 μmol/L of CdCl2 in the NGM (Table 2: h; Table S1: l).

Thus, genes or DEGs likely involved in the innate immune response or signaling were primarily upregulated on E. coli HT115 and E. coli HT115 with 10 μmol/L of CdCl2. Genes or DEGs likely involved in sensory perception were never upregulated on E. coli HT115 with 10 μmol/L of CdCl2. Genes or DEGs likely involved in the stress response were never upregulated on E. coli OP50. The most intensely upregulated DEGs underpinned this result (Fig. 5a–d). DEGs for the immune defense (C-type lectins, fungus-induced proteins) were frequent on E. coli HT115 (Fig. 5a) but also on E. coli HT115 with 10 μmol of CdCl2 in the NGM (Fig. 5c, d). Stress genes such as DEGs for heat shock proteins (HSP-70, small heat shock proteins) or the cadmium defense (CDR-1, MTL-1/2) were more frequent on E. coli HT115 with 0.1 or 10 μmol of CdCl2 in the NGM (Fig. 5b, c).

Fig. 5

Highly upregulated genes from transcriptome profiling and the mRNA level of different C-type lectins at cadmium exposure. Log2-fold changes in the expression of differentially expressed genes (DEGs), which showed the highest level of upregulation, are shown for the contrasts (a) E. coli HT115 vs. E. coli OP50 feeding or (b) 0.1 μmol/L of CdCl2 vs. 0 μmol/L of CdCl2 (control), (c) 10 μmol/L of CdCl2 vs. 0 μmol/L of CdCl2 (control), and (d) 10 μmol/L of CdCl2 vs. 0.1 μmol/L of CdCl2, with E. coli HT115 as bacterial diet in b–d. Several gene groups are highlighted by different colors: gray bars (immune defense genes), black bars (genes for heat shock proteins), dark-gray bars (cadmium defense genes). The relative quantity of transcripts for different C-type lectins was measured by qRT-PCR (using cdc-42 as housekeeping gene) in the RNA-Seq samples from synchronized adult wild type worms, which were bred on NGM plates containing 0 or 10 μmol/L of CdCl2, with E. coli HT115 as bacterial diet (mean ± se; per CdCl2 concentration during breeding, n = 3 biological replicates). Asterisks mark significant differences between control and cadmium-treated worms (∗P ≤ 0.05, ∗∗P ≤ 0.01; two-tailed t-test).

As transcriptome profiling frequently showed an upregulation of C-type lectins in the analyzed contrasts, we determined the relative expression level of several C-type lectin mRNAs by qRT-PCR using the respective samples for RNA-Seq. Actually, there was a significant upregulation of several C-type lectin mRNAs in worms exposed to 10 μmol/L of CdCl2 in comparison to control worms, with E. coli HT115 feeding in both cases (Fig. 5e).

A cluster analysis on the 1607 common DEGs of five experimental conditions (HT115_1, OP50, 0.1 Cd, 10 Cd, HT115_2; see Materials and methods) was carried out (Supplementary Fig. S2). GO analyses on the DEG clusters a0-a22 and b0-b22 especially revealed the following annotation terms (in a partly different sequence): reproduction (GO:0000003), embryo development ending in birth or egg hatching (GO:0009792), receptor-mediated endocytosis (GO:0006898), cellular response to DNA damage stimulus (GO:0006974), and locomotion (GO:0040011) in case of a0-a22 (Supplementary Table S2; terms marked with the Greek character α) and translation (GO:0006412), nematode larval development (GO:0002119), reproduction (GO:0000003), and embryo development ending in birth or egg hatching (GO:0009792) in case of b0, b1, and b11 (Supplementary Table S2; terms marked with the Greek character β). However, the DEGs collected in these five or four terms encode quite different proteins, with helicases, protein kinases, or proteins containing armadillo or WD40 repeats dominating in all mentioned terms from a0-a22, and ribosomal proteins, translational and proteasomal proteins dominating in all mentioned terms from b0, b1, and b11. The highest mean expression intensities (centered and normalized log2-transformed values; see Material and methods) were found under the experimental condition 0.1 Cd (or HT115_1) in case of a0-a22 or OP50 in case of b0, b1, and b11 (boldfaced abbreviated worm treatments). GO analyses on the DEG clusters b2 and b21 especially revealed the annotation terms flavonoid biosynthetic process (GO:0009813) and innate immune response (GO:0045087) (Supplementary Table S2; terms marked with the Greek character γ), with the DEGs collected in these terms primarily encoding either UDP-glucuronosyl/UDP-glucosyltransferases or proteins with CUB-like or Gcn5-related N-acetyltransferase (GNAT) domains. The highest mean expression intensities were found under the experimental condition HT115_2 or 10 Cd in case of b2 or b21. Thus, E. coli OP50 feeding caused maximal mean expression intensities in the DEG clusters b0, b1, or b11, which primarily encode translational proteins. E. coli HT115 feeding or E. coli HT115 feeding with 10 μmol/L of CdCl2 in the NGM caused maximal mean expression intensities in the DEG clusters b2 or b21, which frequently encode proteins for flavonoid biosynthetic processes or the innate immune response.

Discussion

Weak stress

This study aimed to address the effects of weak abiotic and/or biotic stress on the survival probability, stress resistance and gene expression of C. elegans. To specify a weak abiotic stress or stressor, the developmental speed of wild type (WT) was determined at different cadmium concentrations (Fig. 1a–j), which revealed a developmental delay already at 50 μmol/L of CdCl2. Accordingly, we reduced the CdCl2 concentration in the nematode growth medium (NGM), on which worms were cultivated, to levels below one hundredth of this concentration (i.e., 0.1 μmol/L of CdCl2), which was hereinafter classified as weak abiotic stress. This choice also considered a previous report, which has shown positive effects of CdCl2 concentrations between 0.01 and 1 μmol/L on enzyme activities [4]. In addition, several experiments were carried out at a significantly higher cadmium concentration in the NGM (10 μmol/L of CdCl2).

To specify a weak biotic stress/stressor, we tested the effects of two different E. coli strains as bacterial diet. Short-term survival assays showed a tendentiously but not significantly higher survival probability of adult WT worms on E. coli OP50 (Fig. 1k) but a significantly higher survival probability of adult daf-2Δ worms on E. coli OP50 than on E. coli HT115 (Fig. 1n). Long-term survival assays showed no differences in maximum life span but a higher age-specific survival probability of adult WT worms on E. coli OP50 than on E. coli HT115 (Fig. 2a). Thus, we hereinafter classified E. coli HT115 as weak biotic stressor, either as bacterial diet of lower quality or as mildly pathogenic bacterium.

Severe stress

From the data on developmental speed at different cadmium concentrations (Fig. 1a–j), it is evident that a CdCl2 concentration of 10 mmol/L represents a severe abiotic stress/stressor (see also [24]). With regard to a severe biotic stress/stressor, preliminary experiments in our laboratory suggested to use C. arvensicola as animal feed. Accordingly, we tested C. arvensicola as bacterial diet and found significant negative effects on the short-term survival probability of adult worms of all tested C. elegans strains (Fig. 1k–o). Therefore, C. arvensicola can be considered as a severe biotic stressor for C. elegans. The negative effect of C. arvensicola feeding on survival probability was much lower in the stress-tolerant mutant daf-2Δ (Fig. 1n) but only slightly higher in the stress-sensitive mutant daf-16Δ (Fig. 1o) than in WT. Blocked DAF-2 signaling protected daf-2Δ against C. arvensicola by a stress-independent reinforcement of nuclear translocations of the non-phosphorylated transcription factor DAF-16. Inhibiting effects of C. arvensicola on the intact DAF-2 signaling in WT (to trigger DAF-16 nuclear translocations) were obviously not strong because WT and daf-16Δ did not much differ in their survival probability.

Effects of weak abiotic and/or biotic stress on the age-specific survival probability of wild type and different mutant strains

Further experimental series served to detect in identically bred adult worms the effects of different experimental conditions (weak abiotic and/or biotic stress) on their further age-specific survival. In accordance with previous results [43, 44], the signaling mutants kgb-1Δ and pmk-1Δ (Figs. Fig. 1, Fig. 2f–j, k–o) showed a lower age-specific survival probability than WT (Figs. 1k, 2a–e) under each tested experimental condition.

Determining the age-specific survival probability of adult WT, kgb-1Δ, and pmk-1Δ worms showed negative effects of weak biotic stress (Fig. 2a) but positive effects of weak abiotic and biotic stress (Fig. 2c) in WT worms, positive effects of weak biotic stress and weak abiotic and biotic stress (Fig. 2f, i, j) in kgb-1Δ worms, and also positive effects of weak biotic stress (Fig. 2k) but negative effects of weak abiotic and biotic stress (Fig. 2m, n, o) in pmk-1Δ worms.

As the level of fat storage (i.e., available energy resources) is higher in WT on E. coli OP50 than on E. coli HT115 (e.g., [21]), increases in the age-specific survival probability of WT seem to result from either a bacterial diet causing higher levels of energy resources (E. coli OP50) or a combination of weak abiotic and biotic stress (0.1 μmol/L of CdCl2 and E. coli HT115 feeding). Positive effects on the age-specific survival probability of kgb-1Δ were achieved by weak biotic stress alone, independent of the absence or presence of weak abiotic stress. The weak biotic stress also promoted the age-specific survival probability of pmk-1Δ, however, weak abiotic and biotic stress reduced it severely.

The results in kgb-1Δ and pmk-1Δ indicate a higher cadmium sensitivity of pmk-1Δ. This finding is in accordance with previous results [24], which showed that PMK-1 signaling strongly promotes the cadmium resistance of C. elegans mostly via SKN-1 and DAF-16 target gene expression. Thus, while the cadmium resistance of pmk-1Δ is impaired by the reduction in DAF-16 target gene expression, there were no specific effects of KGB-1 signaling on DAF-16 activity under hyperosmotic stress [44]. kgb-1Δ even showed an improved hyperosmotic tolerance likely due to an elevated level of membrane transporters. Accordingly, the absent negative effects of weak abiotic stress in kgb-1Δ may be explained by cadmium exports via membrane transports and/or non-reduced DAF-16 activity. The promotion of the age-specific survival probability of kgb-1Δ or pmk-1Δ by weak biotic stress indicates that an alternative stress mechanism exists, which is independent of KGB-1 or PMK-1 but upregulated on E. coli HT115. Possibly, elevated DAF-16 activity, not induced by PMK-1 signaling but by a signal related to the bacterial diet, constitutes such a mechanism.

Wild type has the immunosuppressive KGB-1 and PMK-1 pathways [25] at its disposal, which also provide resistance to several abiotic stressors [23, 24, 44]. However, the use of these more stress-specific signaling cascades may entail higher energy demands for signal processing (e.g., for the numerous phosphorylation reactions in the MAPK signaling network) than the rather unspecific DAF-2 pathway, where the stress-mediated inhibition of the DAF-2 receptor simply shuts off DAF-2 signaling, which inevitably results in the nuclear translocation of the non-phosphorylated transcription factor DAF-16. This would imply an energetic support for MAPK signaling on a bacterial diet causing higher levels of energy resources (E. coli OP50). Fully active cellular processes, including MAPK signaling, on E. coli OP50 as bacterial diet may then have promoted the stress resistance and age-specific survival probability of WT. The more stress-sensitive kgb-1Δ and pmk-1Δ mutants may predominantly use elevated DAF-16 target gene expression, triggered in this case by weak biotic stress (E. coli HT115), to achieve a higher stress resistance and age-specific survival probability. The more stress-insensitive WT, however, may need an additional weak abiotic stress, besides the weak biotic stress, as signal to promote DAF-16 activity and reach a higher stress resistance and age-specific survival probability via elevated DAF-16 target gene expression. These assertions are assessed in further detail below.

Effects of weak abiotic and/or biotic stress during breeding on the resistance of wild type to severe stress

The resistance of adult WT worms to severe stress was determined after breeding on E. coli OP50 or E. coli HT115 with 0, 0.1, or 10 μmol/L CdCl2 in the NGM (Fig. 3). Especially E. coli OP50 feeding but also weak abiotic and biotic stress (0.1 μmol/L of CdCl2 and E. coli HT115 feeding) during breeding improved the resistance to severe abiotic stress (10 mmol/L CdCl2). Especially weak abiotic and biotic stress but also weak abiotic stress (0.1 μmol/L of CdCl2, with E. coli OP50 as bacterial diet) or weak biotic stress (E. coli HT115 feeding) during breeding improved the resistance to severe biotic stress (C. arvensicola feeding). Thus, a bacterial diet causing higher levels of energy resources or weak abiotic and biotic stress during breeding promoted the resistance to severe abiotic stress, whereas weak abiotic and biotic stress or weak abiotic/biotic stress alone during breeding promoted the resistance to severe biotic stress.

These experiments and the transcriptomic experiments (see below) served to detect the effects of different experimental conditions (weak abiotic and/or biotic stress) during breeding on stress resistance and gene regulation in adult WT. Other survival experiments (Fig. 2), however, were carried out on identically bred adult worms to study the effects of different experimental conditions on their age-specific survival probability. Thus, the results from the former (stress resistance, gene regulation) and the latter (age-specific survival probability) experiments have to be compared with some care, although the results of both types of survival experiments were quite similar (higher stress resistance and age-specific survival probability on E. coli OP50 or E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM).

Effects of weak abiotic and/or biotic stress during breeding on gene expression in wild type

To explore the genetic basis of these results, transcriptome profiling by RNA-Seq was performed. In this context, however, it must be remembered that mRNA levels and the abundance or, moreover, the activity of proteins are not necessarily linked to each other. Although changes in mRNA level mostly take place with the aim of inducing changes in protein level, which frequently take place on different time scales [45, 46], at least repressed mRNA expression may not result in changes of the corresponding protein profile [47]. The reason is that a multitude of processes downstream of mRNA synthesis (e.g., posttranscriptional regulation, protein degradation, protein phosphorylation) may affect the total amount and/or activity of proteins in the cell [48]. Stressor-induced impairments of translational efficiency or protein structure may also require an upregulation of mRNA expression to maintain the protein level. Accordingly, the performed RNA-Seq study can be assessed as starting point for further mechanistic studies.

RNA-Seq was carried out on synchronized adult WT worms, which were bred on E. coli OP50 or E. coli HT115 (Figs. 4a–b and S1a–b) or on E. coli HT115, with 0, 0.1, or 10 μmol/L of CdCl2 in the NGM (Figs. 4c–h and S1c-h). The transcriptomic data were functionally analyzed by GO analyses with regard to up- or downregulated genes (Fig. 4, Table 2) or DEGs (Fig. S1, Table S1) between two different experimental conditions.

The regulated gene or DEG groups with higher expression intensity (Figs. 4 and S1, left graphs) were consistently related to terms referring to protein biosynthesis (Tables 2 and S1). The expression intensity of these genes or DEGs was higher on E. coli OP50 than on E. coli HT115, higher on E. coli HT115 than on E. coli HT115 with 0.1 or 10 μmol/L of CdCl2 in the NGM, and higher on E. coli HT115 with 10 than with 0.1 μmol/L of CdCl2 in the NGM. Cluster and GO analyses (Fig. S2 and Table S2) also showed maxima in mean expression intensity on E. coli OP50 and minima on E. coli HT115 with 0.1 μmol/L of CdCl2 in the NGM in the DEG clusters b0, b1, and b11, which primarily encode translational proteins. Consequently, gene expression for protein biosynthesis decreased in the order: E. coli OP50 feeding, E. coli HT115 feeding, E. coli HT115 feeding with 10 μmol/L of CdCl2, and E. coli HT115 feeding with 0.1 μmol/L of CdCl2 in the NGM. The higher level of gene expression for translational processes on E. coli OP50 indicates a higher level of available energy resources, as an increase in the number of translational machineries involves higher energy (ATP) demands during their operation [49]. The lower level of gene expression for protein biosynthesis under the other experimental conditions may be related to the negative effects of DAF-16 on the expression of an essential component (DAF-15/Raptor) of target of rapamycin (TOR) signaling, which promotes protein biosynthesis and growth [50]. Nuclear DAF-16 inhibits daf-15 expression [51], which inhibits in the long-term TOR signaling, with negative effects on gene expression for protein biosynthesis. This interpretation suggests that negative effects of nuclear DAF-16 on TOR signaling and protein biosynthesis during breeding may have increased in the sequence E. coli OP50 feeding, E. coli HT115 feeding, E. coli HT115 feeding with 10 μmol/L of CdCl2, and E. coli HT115 feeding with 0.1 μmol/L of CdCl2 in the NGM. As this interpretation matches several of the former results on WT and their interpretation (see above), we checked for further indications of lower nuclear DAF-16 levels on E. coli OP50 and higher nuclear DAF-16 levels and positively controlled DAF-16 target gene expression under the other experimental conditions (see below).

The regulated gene or DEG groups with lower expression intensity (Figs. 4 and S1, right graphs) were consistently related to terms referring to innate immune response, sensory perception, signaling, or stress response (Tables 2 and S1).

Genes or DEGs likely involved in the innate immune response or in signaling processes were primarily upregulated on E. coli HT115 and E. coli HT115 with 10 μmol/L of CdCl2 (see Results). The most intensely upregulated DEGs frequently encoded putative immune defense genes on E. coli HT115 (Fig. 5a). Several of these DEGs were also found on E. coli HT115 with 10 μmol of CdCl2 (Fig. 5c, d), and there was a significant upregulation of several C-type lectin mRNAs in worms exposed to 10 μmol/L of CdCl2 in comparison to control worms, with E. coli HT115 feeding in both cases (Fig. 5e). The immune response genes encoded C-type lectins, proteins carrying a CUB-like domain structure, fungus-induced proteins, and metridin-like ShK toxins. Most of them are target genes of PMK-1 signaling [26]. Thus, it is possible that the parallel upregulation of genes or DEGs for signaling and innate immune responses is due to an upregulated gene expression for PMK-1 signaling and associated immune responses.

Genes or DEGs likely involved in sensory perception were never upregulated on E. coli HT115 with 10 μmol/L of CdCl2 (see Results), which may be related to toxic effects, as a loss of neurons was shown in C. elegans already at relatively low cadmium concentrations (2.5 μmol/L of CdCl2 [52]). G-protein-coupled receptors (GPCR) can be activated by chemical stimuli and of 100 candidate chemoreceptors, 60 were found to be expressed in chemosensory neurons in C. elegans [53]. These receptors are predicted to contain seven transmembrane receptor domains (7-TM), including the serpentine receptors for which responses to odorants were suggested [54, 55]. As zebrafish larvae showed significantly lower behavioral responses to L-cysteine stimuli under external cadmium concentrations between 1 and 5 μmol/L of CdCl2 [56], the ability of C. elegans to respond to chemical signals and detect feed may also be impaired at higher cadmium concentrations.

Genes or DEGs likely involved in stress responses were never upregulated on E. coli OP50 (see Results). The most intensely upregulated DEGs encoded several heat shock proteins (HSP-70; small heat shock proteins/α-crystallins) (Fig. 5b) or cadmium defense proteins (CDR-1, MTL-1/2) (Fig. 5c, d) on E. coli HT115, with 0.1 or 10 μmol of CdCl2 in the NGM. The gene cdr-1 is an SKN-1 target gene controlled by PMK-1 signaling [57]. Other genes are DAF-16 target genes [37, 58], such as genes for small heat shock proteins (HSP-16.2/16.41) or metallothioneins (MTL-1/2). The terms referring to insulin-like signaling, which were found on E. coli HT115 (Table S1: h), included up to thirteen signaling substances (insulin-like peptides) for the DAF-2 receptor. As there are activating or inhibiting insulin-like peptides [59], it is not clear whether they can activate or inhibit DAF-2 signaling. But it is evident that E. coli HT115 feeding caused a greater involvement of DAF-2 signaling. The term ‘response to heat’, which was found on E. coli HT115 with 0.1 μmol of CdCl2 (Table 2: l; Table S1: e) included up to five genes associated with DAF-2 signaling and DAF-16 target gene expression (e.g., hsp-16.2/16.41). Thus, DAF-2 signaling and DAF-16 target gene expression seem to be involved in the stress responses of WT on E. coli HT115, especially when 0.1 or 10 μmol of CdCl2 was added to the NGM.

These results support the former suggestion (see above) of lower nuclear DAF-16 levels on E. coli OP50 and higher nuclear DAF-16 levels on E. coli HT115 with 0.1 or 10 μmol/L of CdCl2. As PMK-1 signaling is possibly involved in the stress and immune responses on E. coli HT115 and E. coli HT115 with 10 μmol/L of CdCl2 (see above), DAF-16 target expression may be particularly important on E. coli HT115 with 0.1 μmol/L of CdCl2.

Conclusion

PMK-1 signaling and DAF-16 target gene expression provide resistance to abiotic and/or biotic stress. On a bacterial diet causing higher levels of energy resources (E. coli OP50), the available energy appeared to be sufficient to operate MAPK signaling, protein biosynthesis, and other cellular processes at simultaneously high levels, which probably caused the high resistance to severe abiotic stress (10 mmol/L of CdCl2) and promoted the age-specific survival probability of WT.

In case, energy resources were more restricted (E. coli HT115 feeding), gene expression for protein biosynthesis was downregulated but PMK-1 signaling and the associated expression of immune response genes were obviously upregulated, triggered by a signal related to this bacterial diet. The upregulation of PMK-1 signaling under weak biotic stress likely caused the higher resistance of adult WT to severe biotic stress (C. arvensicola feeding). Additionally, there were first indications for a greater involvement of DAF-2 signaling under this experimental condition.

An additional weak abiotic stress (0.1 μmol/L of CdCl2), besides the weak biotic stress (E. coli HT115), led to a further reduction in gene expression for protein biosynthesis. The upregulation of DAF-16 target gene expression (e.g., genes for small heat shock proteins) under this stress condition probably caused the high (higher) resistance of adult WT to severe biotic (abiotic) stress and was possibly responsible for the higher age-specific survival probability of adult WT under this experimental condition.

A clearly higher additional abiotic stress (10 μmol/L of CdCl2), however, re-intensified in a certain degree gene expression for protein biosynthesis as well as PMK-1 signaling and associated gene expression (cdr-1, immune response genes). DAF-16 target gene expression was still operating, with a focus now on the expression of cadmium defense genes (e.g., genes for metallothioneins). This and a possibly lower DAF-16 activity under this stress condition than under weak abiotic stress (0.1 μmol/L of CdCl2) may have been responsible for the missing preparation of WT for the subsequent exposure to severe biotic stress.

Thus, feeding with bacteria providing higher energy resources or weak abiotic and biotic stress during breeding particularly positively affect the stress resistance of adult worms, which is likely based on non-restricted MAPK signaling and protein biosynthesis or DAF-16 target gene expression. Adult worms also benefit from these environmental conditions as they improve their age-specific survival probability.

Declarations

Author contribution statement

Ramona Dölling: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Martha N. Mendelski: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data.

Rüdiger J. Paul: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.

Funding statement

This work was supported by the Deutsche Forschungsgemeinschaft (Pa 308/13-1).

Competing interest statement

The authors declare no conflict of interest.

Additional information

Data associated with this study has been deposited at the NCBI Gene Expression Omnibus (GEO) under the accession number GSE123183.

Supplementary content related to this article has been published online at https://doi.org/10.1016/j.heliyon.2019.e01126.