Cyclic di-nucleotides are important secondary signaling molecules in bacteria that regulate a wide range of processes. In this study, we found that Caenorhabditis elegans can detect and are attracted to multiple signal molecules produced by Vibrio cholerae, specifically the 3′,5′-cyclic diguanylate (c-di-GMP), even though this bacterium kills the host at a high rate. C-di-GMP is sensed through C. elegans olfactory AWC neurons, which then evokes a series of signal transduction pathways that lead to reduced activity of two key stress response transcription factors, SKN-1 and HSF-1, and weakened innate immunity. Taken together, our study elucidates the role of c-di-GMP in interkingdom communication. For C. elegans, bacterial c-di-GMP may serve as a cue that they can use to detect food. On the other hand, preexposure to low concentrations of c-di-GMP may impair their immune response, which could facilitate bacterial invasion and survival.
Bacteria are able to communicate with one another through a process known as quorum sensing (QS). This process mainly relies on the production and secretion of specific autoinducers (AIs), and allows the cells to make group decisions based on cell-density1. While the AIs are extracellular signaling molecules, bacteria also use a range of nucleotide-based intracellular signaling molecules known as second messengers to regulate physiological responses to cope with a changing environment. Molecules such as cyclic adenosine 3′,5′-monophosphate (cAMP) and guanosine pentaphosphate or tetraphosphate ((p)ppGpp) have been well studied for almost 50 years2,3. In the past two decades, the field of cyclic dinucleotides (CDNs) is expanding and has attracted more attention in different areas of research. 3′,5′-cyclic diguanylate (c-di-GMP) was the first identified CDN and has been extensively studied since its discovery. Initially characterized as an activator of cellulose synthase in Acetobacter xylinum4, c-di-GMP is now known as a ubiquitous bacterial signal that regulates a variety of physiological processes in Gram-negative bacteria such as motility, biofilm formation, virulence, and transmission between hosts5. Its role in modulating flagellar motility, exopolysaccharide production, hyphae formation, etc., has also been reported in a few Gram-positive bacteria6. The second CDN, 3′,5′-cyclic diadenylate (c-di-AMP), is mainly found in Gram-positive bacteria. It has been characterized as an essential signaling molecule that regulates cell wall homeostasis, potassium ion channels, DNA integrity, as well as biofilm formation and virulence7. Adenosine-guanosine-3′,3′-cyclic monophosphate (cGAMP) is the newest addition to the CDN list that has only been identified in very few bacteria. In Vibrio cholerae (V. cholerae), cGAMP plays a role in efficient intestinal colonization8, and in Geobacter, it controls exoelectrogenesis9.
As more research has been conducted, it was discovered that the role of these signaling molecules is not limited to communication among bacterial cells. Bacteria and their eukaryotic hosts can also communicate with each other via these signaling molecules. This so-called interkingdom communication has recently become an expanding field of research with broad implications. Bacteria can sense and respond to mammalian hormones. For example, the classic stress hormones adrenaline and noradrenaline can induce bacterial growth and virulence expression. Conversely, bacterial AIs (specifically the acyl-homoserine lactones), can enter the mammalian cells and modulate host immune response and promote apoptosis10. Bacterial CDNs also have immunomodulation functions. It was reported that Listeria monocytogenes secrets c-di-AMP through a multidrug efflux pump and activates host type I interferon11. Whether c-di-GMP or cGAMP is secreted by bacteria is not known at this time. However, these signals can be detected by the endoplasmic-reticulum-resident protein STING (stimulator of interferon genes) in humans and mediate the type I interferon immune response12,13.
In the aforementioned cases, bacterial signals need to enter the host cells in order to elicit a specific response. Recent studies of bacteria-C. elegans interactions illustrated that bacterial AIs could function as chemical cues or odors that affect worms’ behavior. Beale et al. first reported that C. elegans senses Pseudomonas aeruginosa acyl-homoserine lactones and are attracted by the signal. They also showed that C. elegans could acquire relatively long-lasting memory to avoid the AIs if they were pre-exposed to them for a short period14. V. cholerae, the etiologic agent of cholera, possesses multiple QS signaling molecules. CAI-1 and AI-2 are the two well-studied autoinducers. The CAI-1 molecule was found to chemoattract C. elegans, and it is sensed through the amphid AWCON neuron15. In this study, we show that in addition to CAI-1, c-di-GMP is another chemoattractant that can be sensed by C. elegans by both AWCON and AWCOFF neurons. Sensing c-di-GMP elicits a series of signal transduction pathways in the host, which leads to a reduced innate immune response and a shortened lifespan by affecting the activity of two key stress response transcription factors, SKN-1 and HSF-1.
C. elegans are attracted to V. cholerae by signals other than the QS autoinducers
In a previous study, we reported that water-soluble cranberry extract protects C. elegans from killing by different pathogenic bacteria, including V. cholerae, Pseudomonas aeruginosa, Salmonella typhimurium, Enterococcus faecalis, and Staphylococcus aureus16. During these killing assays, an interesting phenomenon caught our attention. Worms fed on V. cholerae seemed to be allured by this type of bacterium and remained in the bacterial lawn until death. In contrast, worms fed on the other four pathogens showed a scattered behavioral pattern on the test plates. It is known that C. elegans can sense and avoid bacterial pathogens, but in our study, the worms were lured to a pathogen that significantly decreases their lifespan. A simple choice index (CI) assay was performed with the N2 wild type C. elegans over the course of 2 h to test whether the worms prefer the pathogenic V. cholerae wild type (wt) strain C6706 rather than E. coli OP50, the common food source for C. elegans used in the lab. As shown in Fig. 1a, the worms were readily attracted to V. cholerae (wt) compared to E. coli OP50.
Previously, Werner KM et al. reported that V. cholerae QS CAI-1 is a chemoattractant sensed by C. elegans amphid sensory neuron AWCON15. To test whether other signal molecules are involved in chemoattraction, we utilized three mutant strains: cqsA− (no CAI-1 production), luxS−(no AI-2 production), and a cqsA−/luxS− double deletion. Each mutant was tested against E. coli OP50 using the standard CI assay. C. elegans preference was observed towards all three mutant strains (Fig. 1a), conveying that signals other than the two autoinducers play a role. Interestingly, the choice index of any of the mutant strains over E. coli was comparable to that of the wild type over E. coli. This observation implies that rather than CAI-1, other signal molecules play a predominant role in attracting C. elegans. Next, we mixed the supernatant of the overnight culture of wild type V. cholerae and the cqsA−/luxS− double mutant with E. coli OP50, respectively, and tested worms’ preference over E. coli OP50. As expected, worms were attracted to E. coli OP50 mixed with the supernatant from either the wild type or the double mutant strain (Fig. 1b). This result confirms the existence of other chemoattraction signals present in the V. cholerae cell-free supernatant.
C-di-GMP is the major signaling molecule that attracts C. elegans
QS autoinducers and the cyclic dinucleotides are essential signaling molecules that regulate numerous physiological functions in bacteria. Unlike the QS autoinducers, which are secreted into the environment, the cyclic dinucleotides are generally known as intracellular second messengers. We questioned whether c-di-GMP and cGAMP, the two cyclic dinucleotides characterized in V. cholerae, could function as extracellular signals that attract C. elegans. To test this, we purchased pure solutions of c-di-GMP and cGAMP. CI assays were conducted as usual, except that E. coli OP50 was mixed with two concentrations (0.1 and 1 nM) of each cyclic dinucleotide to observe preference. These concentrations were tested because it was theorized that low levels of these molecules would be present outside of the cells. Figure 2a shows that at 1 nM, both c-di-GMP and cGAMP were able to trigger an attractive behavior in C. elegans. As a control, c-di-AMP (1 nM) was also tested in the same assay. In contrast to c-di-GMP and cGAMP, c-di-AMP caused an apparent avoidance in C. elegans. We also tested similar signaling molecules, GMP and cGMP, and showed that neither cGMP nor GMP could cause any chemoattraction or repulsion at concentrations similar to what elicited a response from c-di-GMP and cGAMP (Fig. 2a).
We next harvested the supernatant and the cell pellet from the wild type V. cholerae overnight culture (~ 109 CFU/ml) by centrifugation, and measured the concentration of c-di-GMP and cGAMP from the two portions using liquid chromatography-mass spectrophotometry (LC-MS) assay. Pure c-di-GMP and cGAMP solutions were used as standards for this assay. As shown in Fig. 2b, c-di-GMP was detected in both the cell lysate and the supernatant. The concentration was calculated as 177.3 ± 49.4 nM in the cell lysate and 19.3 ± 4.6 nM in the supernatant. No cGAMP was detected in the cell lysate or the supernatant. These results suggest that c-di-GMP may be the additional signal that attracts C. elegans towards V. cholerae.
To further confirm the role of c-di-GMP in chemoattraction, we sought to manipulate the concentration of c-di-GMP in V. cholerae. C-di-GMP is synthesized by the diguanylate cyclases containing a GGDEF domain and degraded by the phosphodiesterases containing either an EAL or HD-GYP domain17. pAT1568, a plasmid that overexpresses the phosphodiesterase (EAL domain) of the vieA gene from the l-arabinose-inducible pBAD promoter18, was introduced into the cqsA−/luxS− strain by electroporation. The resulting transformant was grown overnight in the presence of 0.2% l-arabinose to induce phosphodiesterase expression and was used in CI assays against E. coli OP50. As a control, the empty vector pBAD33 was also introduced into the cqsA−/luxS− strain and cultivated in a similar fashion. Figure 2c shows the CI result. With pAT1568 (i.e., decreased c-di-GMP level), C. elegans was no longer attracted to the cqsA−/luxS− strain (choice index ≈ 0), while with the empty vector pBAD33, a strong preference was still observed. This observation revealed the critical role of c-di-GMP production and presence in the chemoattractive behavior of C. elegans. Combining the results from Fig. 2a, b, we consider that c-di-GMP is the major signaling molecule that attracts C. elegans.
C. elegans senses c-di-GMP through the AWC neurons and the cGMP-gated TAX-2/TAX-4 channel
The chemosensory system in C. elegans allows the organism to detect food, develop, avoid danger, mate, etc. There are 11 pairs of amphid chemosensory neurons in C. elegans. Of these, the ASE gustatory neurons are known to sense salts and water-soluble attractants, and the AWA and AWC olfactory neurons are required to sense attractants with volatile odors. Different chemical signals are sensed by different G protein-coupled chemoreceptors in the chemosensory neurons, and then passed to two major signal transduction sensory channels, the cGMP-gated TAX-2/TAX-4 channel and the lipid-sensing OSM-9/OCR-2 TRPV channel19. It was reported that C. elegans senses V. cholerae CAI-1 through the AWCON neuron and uses the TAX-2/TAX-4 channel15. We hypothesized that the same neuronal sensory pathway could be involved in sensing c-di-GMP. To test this, different C. elegans mutant strains were assessed to determine if they were able to similarly respond to c-di-GMP as the wild type strain. The tax-2/tax-4 mutant was no longer attracted to c-di-GMP, while the osm-9/ocr-2 mutant was still attracted, signifying that the cGMP-gated TAX-2/TAX-4 channel is required (Fig. 3). Mutation to ceh-36 causes defects in developing functional AWC and ASEL neurons, and mutation to che-1 causes an inability to develop functional ASEL and ASER neurons. Mutations to nsy-5 and nsy-1 cause an inability to develop the AWCON and AWCOFF neurons, respectively. Figure 3 shows that loss of functional AWC or ASEL neuron (ceh-36) kept the worms from the attraction, and worms without the functional ASEL and ASER neuron (che-1) were still attracted, indicating the AWC neuron is involved. Further assay with the nsy-5 and nsy-1 mutants revealed that the AWCON and AWCOFF neurons are both necessary because no preference was observed when studying these mutants. Based on these results, it can be concluded that C. elegans senses c-di-GMP through the AWCON and AWCOFF neurons and the cGMP-gated TAX-2/TAX-4 channel.
C-di-GMP suppresses the innate immunity of C. elegans
Since it has been shown that c-di-GMP can cause initial attraction of C. elegans over the course of a couple of hours, it was next of interest to investigate if detection of this signal molecule plays a role in the health of the host. Lifespan assays were performed in which NGM-FUDR media was supplemented with 1 nM of c-di-GMP to test the lifespan of C. elegans N2, when exposed to c-di-GMP throughout their entire life. The results in Fig. 4a and Table 1 show that the presence of c-di-GMP significantly affects the lifespan of C. elegans N2 in that worm populations typically are living much shorter compared to control. The mean number of days lived of N2 worms at 25 °C when exposed to c-di-GMP was 11.30 ± 0.48, whereas nonexposure was 13.66 ± 0.39, an 17.0% decrease (p < 0.001). The lifespan shortening effect is specific to the c-di-GMP molecule, as neither cGAMP nor c-di-AMP at 1 nM affected the normal lifespan of N2 worms (Fig. 4a and Table 1). We also found that the presence of live bacteria is requisite, as c-di-GMP only caused a slight decrease in the lifespan when the N2 worms were fed on heat-killed E. coli OP50 (Fig. 4b and Table 1).
Younger worms have robust grinders that can break down bacteria effortlessly, thus live bacteria are often absent from the lumen. There is a decline in grinder function as the worms age, which allows bacteria to easily escape to the lumen of the gut and proliferate20,21. Effective curb of intestinal bacterial accumulation is an indicator of strong gut immunity, and also an important causative factor of lifespan determination20,21. We therefore compared the gut colonization of E. coli OP50 from day-6 (equivalent to middle age) N2 worms in the presence and absence of c-di-GMP. A substantial increase of bacterial number was observed with the addition of 1 nM c-di-GMP (Fig. 4c). Incubating E. coli OP50 with the same concentration of c-di-GMP in vitro neither increases the growth rate nor enhances bacterial surface attachment (Supplementary Figs. 1 and 2). Moreover, when E. coli OP50 was cultured with 1 nM c-di-GMP overnight and then fed to N2 worms (no c-di-GMP added to the assay plates), the resulting lifespan remains the same as the control (Supplementary Fig. 3). These observations indicate that the increased gut colonization is largely attributed to a weakened gut immunity in the host.
We speculated that the expression of C. elegans innate immune response genes might be inhibited by c-di-GMP based on previous results. To examine this, we selected a few innate immune genes (C23G10.1, clec-46, clec-71, col-41, dct-5, fmo-2, pqn-5, and dod-22) that are reported to be upregulated during bacterial infections22,23 and analyzed their expression by qRT-PCR. Figure 5 shows that when synchronized L4-stage N2 worms were exposed to 1 nM c-di-GMP for only 10 min, expression of these genes was generally reduced 2-fold to 5-fold except for dod-22, which decreased by 1.6-fold. This was consistent with our hypothesis that the presence of this signaling molecule is modulating innate immunity.
C-di-GMP acts through SKN-1 and HSF-1 to impact immunity and lifespan
The FOXO family protein DAF-16, the SKN-1 protein, and the heat shock protein HSF-1 are the three major stress response transcription regulators in C. elegans. DAF-16, the downstream transcription factor of the insulin/insulin-like growth factor-1 signaling (IIS) pathway, plays a key role in modulating longevity and immunity24,25,26. SKN-1 is the downstream effector of the major immune-signaling p38 mitogen-activated protein kinase (MAPK) pathway and controls numerous genes involved in stress response and lifespan regulation27. HSF-1 is another versatile transcription factor that regulates a multitude of genes involved not only in stress response but also in development, metabolism, as well as lifespan and immunity modulation28,29,30. To test whether c-di-GMP acts through these regulators, lifespan assays in the presence and absence of c-di-GMP were conducted in mutant daf-16 and hsf-1 strains. In the case of skn-1, RNA interference (RNAi) was utilized to knock down the gene expression, and the empty vector pL4440 was used in parallel to serve as the control. We reasoned that if c-di-GMP acts through a particular transcriptional regulator to shorten the lifespan, mutating or knocking down expression of the gene will eliminate the effect. Figure 6a–c and Table 1 show the results. Supplementation of 1 nM of c-di-GMP still reduced the lifespan of the daf-16 mutant strain, and the percentage of decrease is comparable to that in the wild type N2. Lifespan was not affected by c-di-GMP in the hsf-1 mutant and the skn-1 RNAi knock-down strains, meaning that SKN-1 and HSF-1 are required for c-di-GMP to exert its effect.
Next, two representative target genes of DAF-16 (C10G11.5 and F52H3.5), two of SKN-1 (gcs-1 and gst-4), and two of HSF-1 (hsp-16.2 and hsp-70) were measured for their expression levels in the presence and absence of c-di-GMP. Figure 6d shows that c-di-GMP triggered downregulation of the SKN-1 and HSF-1 target genes but not the DAF-16 targets. These results suggest that sensing c-di-GMP by the chemoreceptors in the sensory neuron elicited signal transduction, that eventually led to reduced activity of SKN-1 and HSF-1.
p38 MAPK(PMK-1) and RTK-Ras-ERK MAPK (MPK-1) are required in the c-di-GMP-elicited signal transduction pathway
MAP kinases are central components of a series of signal transduction pathways that control many vital cellular processes. MAP kinases in mammals are grouped into three families: p38/SAPKs (stress-activated protein kinases), JNKs (Jun amino-terminal kinases), and ERKs (extracellular-signal-regulated kinases)31. The three corresponding MAPKs in C. elegans are PMK-1, JNK-1, and MPK-1. Activity and cellular localization of SKN-1 and HSF-1 are modified by phosphorylation. Phosphorylated SKN-1 and HSF-1 are translocated into the nucleolus where they activate their target genes. Previous studies reported that either PMK-132 or MPK-133 possess the ability to directly phosphorylate SKN-1 at the same site. The kinases that phosphorylate HSF-1 remains elusive. Using the pmk-1 and jnk-1 mutant strains as well as the mpk-1 RNAi, we examined whether these MAPKs are involved in the signal transduction pathway elicited by c-di-GMP. Lifespans of these mutant worms were compared in the presence and absence of 1 nM c-di-GMP and shown in Fig. 7a–c and Table 1. In the pmk-1 mutant and the mpk-1 RNAi worms, adding c-di-GMP in the medium no longer reduced the lifespan as compared to that in the N2 or N2/pL4440 worms. While in the jnk-1 mutant, a 10% decrease in lifespan was observed. Our data suggest that the two MAP kinases, PMK-1 from the p38 MAPK pathway and MPK-1 from the RTK-Ras-ERK pathway, are required in the c-di-GMP-elicited signal transduction.
Using the V. cholerae-C. elegans interaction model, we discovered that c-di-GMP acts as an interkingdom communication signaling molecule to attract C. elegans and suppress the host’s innate immunity. Sensing c-di-GMP is through the olfactory AWCON and AWCOFF neurons and the cGMP-gated TAX-2/TAX-4 channel. This is different from the case of sensing V. cholerae CAI-1, which only requires the AWCON neuron15. Unlike the gustatory neurons, olfactory neurons are able to detect long-range signals, usually at nanomolar concentrations. In consensus with this, the concentration of c-di-GMP from the overnight V. cholerae culture was measured to be around 20 nM.
Though the two species may not intersect in nature, we think our findings could have universal implications. Being a soil-dwelling nematode, C. elegans feeds on bacteria and has coevolved with bacteria over the years, and thus has developed a particularly intimate relationship with bacteria. We believe that the production of c-di-GMP by bacteria, or its presence outside of the bacterial cell, acts as a cue attracting C. elegans towards a food source. That initial response and attraction from the worms by c-di-GMP could signal a favorable environment. While this attraction may decrease immune response initially, degradation and weakening of this signal over time could lead to a return to baseline immunity under noninfectious conditions in vivo. However, when this signal is produced by a pathogen, such as V. cholerae, this initial attraction could be all it takes for C. elegans to begin feeding and thus become infected. Pathogens such as V. cholerae and P. aeruginosa possess a powerful signaling network that utilizes c-di-GMP for its success. C-di-GMP has been considered to be prevalent in Gram-negative bacteria, but its existence and function in Gram-positive bacteria have been gradually revealed in recent years, especially in those commonly found in the soil, such as Bacillus spp.34,35,36,37, Streptomyces coelicolor38, Clostridium difficile39, and Mycobacterium smegmatis40. With the attraction of C. elegans towards low concentrations of c-di-GMP, this can allow it to find numerous bacteria in the environment, Gram-positive or Gram-negative, pathogen or not.
Our observation seems contradictory to the current knowledge that CDNs trigger the innate immune system by activating interferon genes12. However, we do not think it is a contradiction. C. elegans lacks the c-di-GMP receptor STING41 and the robust interferon system, therefore c-di-GMP will not trigger the canonic STING pathway in the C. elegans model. Only within the last decade has significant research been conducted investigating the binding of bacterial CDNs to the human STING receptor. Initially, this was implicated in the innate immune response to pathogens, but more recently, it has been shown to detect tumor-derived DNA and direct antitumor immunity through the versatile type I IFN42. The human STING receptor has a higher binding affinity with cGAMP and c-di-GMP than it does to c-di-AMP43. Its affinity is even higher when binding endogenous mammalian cGAMP, which possesses a 2′-5′ phosphodiester linkage rather than the microbial 3′-5′ linkage13. Through these studies, it was found that STING has the affinity to bind to the bacterial c-di-GMP at a concentration of about 4.4 μM, which initiates the downstream signal transduction. Efforts have been made in attempts to modify STING for higher affinity, and while this proved successful, the concentration of c-di-GMP needed for binding was still high, 0.461 ± 0.053 μM43.
In our study, we found that c-di-GMP, as low as 1 nM, is being sensed by C. elegans, and this decreases its immune response. We also found that this behavioral response occurred at similar concentrations towards microbial cGAMP. Although not uncovered in this work, there is a possibility that C. elegans may possess a receptor with a higher affinity for CDNs than the mammalian STING. If this were the case, finding this receptor and studying its structure will provide valuable information to the STING research field. The ability to bind CDNs at nanomolar concentrations would remarkably improve STING’s sensitivity to foreign and cancer cell-derived DNA. This is an exciting new area of research, and it has the possibility to be far-reaching, especially in the fight against cancer and improving immunotherapies.
We showed that c-di-GMP acts through two key stress response transcription factors in C. elegans, SKN-1 and HSF-1. SKN-1 is the downstream effector of the major immune-signaling p38 MAPK pathway and controls numerous genes involved in stress response and lifespan regulation27. HSF-1 is a multifaceted transcription factor involved in stress response, development, metabolism, as well as lifespan and immunity modulation28,29,30. The possible MAP kinases (PMK-1, JNK-1, and MPK-1) that regulate SKN-1 and HSF-1 activity and nucleus localization were investigated. PMK-1 from the p38 MAPK pathway and MPK-1 from the RTK-Ras-ERK pathway are required in the c-di-GMP-elicited signal transduction. It is known that both PMK-1 and MPK-1 can phosphorylate SKN-1 to modulate its function32,33, while the kinase in C. elegans to phosphorylate HSF-1 is still not clear. Our data propose the possibility that PMK-1 or MPK-1 might be the upstream kinase for HSF-1. Of course, further studies are needed to address this interesting question.
Although we believe that we have successfully defined this new role for bacterial second messengers and their effect on eukaryotes, more work still needs to be done. Communications between bacteria and their hosts likely have played an indispensable role during their coevolution over millions of years. Bacteria have evolved to release signaling molecules to improve their survival and spread by interfering with the host’s behavior and immune response. In conjunction, the hosts have developed receptors for these signals to trigger immune responses to defend themselves. This has contributed to the symbiotic relationship that exists between bacteria and their hosts. Manipulation of such signaling can provide valuable insight to researchers looking to develop novel methods to treat infection and combat persistence. The better we understand this “language”, the better we can start to control the conversation rather than reacting.
Plasmids, bacterial strains, and growth conditions
The plasmid pBAD33 (pACYC184 araC ParaBAD, Cmr)44 was obtained from Coli Genetic Stock center, pAT1568 (pBAD33::NTF3vieA-His6) was provided by Dr. Andrew Camilli (Tufts University)18. The V. cholerae strains used in this study were all derived from the wild-type C6706 strain (O1 serotype El Tor isolated from Peru)45. The cqsA− (cqsA deletion), luxS− (luxS deletion), cqsA−/luxS− (cqsA and luxS double deletion) strains were obtained from Dr. Jay Zhu (University of Pennsylvania)46,47,48. E. coli OP50 came from the laboratory stocks. E. coli OP50 and V. cholerae were cultured in Lysogeny broth (LB) with 100 μg/mL of streptomycin. Chloramphenicol (5 μg/mL) was used for the selection of pBAD33 and pAT1568 in V. cholerae. l-arabinose was added at a final concentration of 0.2% to induce expression from ParaBAD.
C. elegans strains and growth conditions
All C. elegans strains were obtained from the Caenorhabditis Genetics Center, University of Minnesota, USA. Strains used in this study were: N2 Bristol (wild type), AU3 (nsy-1(ag3)), FK100/PR678 (tax-2(ks10)/tax-4(p678)), FG125 (ocr-2(ak47), osm-9(ky10), ocr-1(ak46)), FK311 (ceh-36(ks86)), PR674 (che-1(p674)), CX6161 (inx-19(ky634), previously nsy-5), GR1370 (daf-16 (mgDf50)), PS3551 (hsf-1(sy441)), KU25 (pmk-1(km25)), and VC8 (jnk-1(gk7)),
All strains were maintained at their permissive temperature on nematode growth medium (NGM) seeded with E. coli OP50 feeding strain. NGM for maintenance was prepared via standard protocol in 60 mm plates. E. coli OP50 was dropped on the center of the plates the night prior to transfer of worms. Bacterial strains were dropped onto the center of the plates 2 h prior to worm transfer. NGM-FUDR 35 mm plates were used for lifespan assays. Fluorodeoxyuridine (FUDR) is an inhibitor of DNA synthesis, and at a concentration of 50 μg/mL it prevents the reproduction and development of progeny but doesn’t interfere with the lifespan and the physiology of adult C. elegans49,50. Five times concentrated bacterial cultures were dropped in 100 μL aliquots onto the center of plates 2 h prior to worm transfer. NGM or NGM-FUDR plates with 1 nM c-di-GMP were created by added correct concentrations of stock c-di-GMP (1 mM) to warm NGM mixtures after autoclaving.
Preparation of cyclic di-nucleotides
Cyclic diadenosine monophosphate (c-di-AMP), cyclic diguanosine monophosphate (c-di-GMP), and cyclic adenosine monophosphate– guanine monophosphate (cGAMP) were purchased from BioLog Life Science Institute (catalog numbers are C 088, C 057, and C 117, respectively). Stock concentrations of these sodium salt compounds were made by dissolving in ddH2O. The stock samples were stored in −20 °C freezer. Serial dilutions were made from frozen stocks to obtain desired concentration before experiments were conducted.
Liquid chromatography-mass spectrophotometry (LC-MS) assay
V. cholerae C6706 was grown shaking at 37 °C for 22 h in 100 mL of LB broth. The culture was centrifuged for 10 min at maximum speed to separate the supernatant and the cell pellet. The supernatant and the cell lysate were tested for the presence of CDNs, corresponding to the extracellular and intracellular concentrations, respectively. The supernatant was treated with final concentration of 10 mM EDTA for 10 min and then passed through 0.45-µ filter. The filtrate was freeze-dried, resuspended in 500 μL of sterile ddH2O, and centrifuged in Amicon Ultra-0.5 mL centrifugal filter device to remove any DNA or protein molecules larger than 3 KDa. The sample was lyophilized and stored at room temperature before LC-MS assay. The cell pellet was suspended in 1 mL of 25 mM Tris-HCl + 10 mM EDTA buffer. Cells were then lysed by sonication at 20 W for 60 s (three 20 s pulses). Tricholoacetic acid at a final concentration of 12% was used to precipitate cellular macromolecules and centrifugation was used to separate the precipitate. The supernatant was freeze-dried, resuspended in 500 μL of sterile ddH2O, and centrifuged in Amicon Ultra-0.5 mL centrifugal filter devices to remove any remaining DNA or protein molecules larger than 3 KDa. The sample was lyophilized and stored at room temperature before LC-MS assay.
C-di-GMP and cGAMP in supernatant and cell pellet were quantified by HPLC with online tandem mass spectrometry (MS). The CDNs were resolved on a C18 column (150 × 2 mm, ProdigyTM 5 μm ODS-2 150 Å) (Phenomenx) using a gradient generated between 0.2% formic acid in water (A) and 0.2% formic acid in methanol (B) at a flow rate of 0.25 mL/min starting from 0% B over 2 min, then linearly to 20% B over 4 min, then to 100% B over 0.5 min, followed by this solution for 3 min, then back to 0% B. CDNs were analyzed on a Shimadzu 8050 triple-quadrupole mass spectrometer interfaced to a Shimadzu UHPLC multiplexing system using electrospray ionization in positive-ion mode with multiple reaction monitoring of parent and characteristic daughter ions m/z 675.10 → 136.10 for cGAMP and m/z 691.10 → 152.10 for c-di-GMP. Several other parent and characteristic daughter ions m/z 675.10 → 524.00, 675.10 → 330.05 and m/z 691.10 → 539.90, 691.10 → 248.05 were used as reference to confirm the defined cGAMP and c-di-GMP, respectively. The MS parameters were optimized by using their respective standards. Series concentrations of CDN standards were monitored by the above UHPLC/MS to determine retention times and regression equations to calculate the concentrations in supernatant and cell pellet.
Choice index (CI) assays
C. elegans were grown at 25 °C on E. coli OP50 under well-fed and un-crowded conditions. CI assays were performed on standard LB agar plates. The plates were divided in half to reveal the center point in each plate. Bacterial strains that were used for this experiment were grown in a shaking incubator overnight in the proper growth medium and temperature at approximately 120 rpm. Overnight cultures were then seeded onto each end of the plate 6 cm apart, and a 2 cm radius is drawn around each lawn (Fig. 8). The lawns are allowed to dry for 2 h before experiments were conducted. After the lawns were able to dry, 1 µL of 10 mM sodium azide (NaN3) was dropped on the center of each lawn to paralyze C. elegans to make sure, they did not change bacterial lawns once one was initially chosen. Between 50 and 150 well-fed N2 worms were then placed in the center of the assay plate to begin the choice experiment. Worms present within the 2 cm radius of each bacterial lawn were counted after 1 and 2 h. Each assay was performed independently and in at least triplicate. Choice index was calculated as follows:
*Choice Index = (# of worms on test strain - # of worms on E. coli OP50) / Total # of worms.
*Positive values indicate preference towards the test strain, while negative values indicate preference towards E. coli OP50. If there were no preference, the equation would yield a value close to 0.
Average and standard error were calculated from three independent experiments.
Unpaired Student’s T-test (between two groups) and one-way balanced ANOVA (for more than two groups) were used for statistical analysis. A p-value < 0.05 was accepted as statistically significant.
C. elegans lifespan assay
Lifespan assays were carried out at 25 °C. Worms were synchronized by transferring 20 gravid worms to 60 mm NGM plates seeded with E. coli OP50 2 days prior to the start of assays. Worms were allowed to lay eggs for 4 h, and then parent worms were removed from plates, leaving only synchronized eggs. Worms were then incubated until L4 stage. Overnight bacterial cultures were concentrated by centrifugation and removal of 50% of supernatant. Cultures were then resuspended in remaining medium. When necessary, E. coli OP50 was heat killed at this stage (1 h at 80 °C). Aliquots of 100 μL of E. coli OP50 were dropped on the center of NGM-FUDR 35 mm plates (with and without 1 nM c-di-GMP) and allowed to dry for 2 h. Worms in L4 stage were transferred to assay plates; 20 worms per plate. If the worms were transferred onto plates with heat-killed E. coli OP50, then prior to transferring to the assay plates they were allowed to roam on LB gentamycin (50 µg/mL) plates for 1 h to rid external live bacteria. Plates were subsequently checked daily. Dead worms were removed from the plates after being counted and recorded. The day of transfer was defined as day zero. Statistical analysis was carried out through SPSS software under the Kaplan–Meier lifespan analysis. The p-values were determined via log rank test, and a p < 0.05 was accepted as statistically significant.
Gut colonization of C. elegans
E. coli OP50 was grown overnight in LB broth with 100 µg/mL of streptomycin broth. Aliquots of 100 µL of E. coli OP50 was dropped on NGM with 1 nM c-di-GMP and control NGM plates. Five C. elegans N2 gravid worms were transferred onto each plate and incubated at 20 °C for 4 h. Synchronized egg populations were allowed to grow to L4 stage (~40 h), and then transferred to NGM-FUDR and NGM-FUDR with 1 nM c-di-GMP plates with 100 µL of respective overnight cultures at 2× concentration. This was designated as Day 0, and tests were conducted on Day 6 aged worms on each bacterial species +/− c-di-GMP. Ten worms were randomly chosen from each plate and transferred to LB gentamycin (50 µg/mL) plates for one hour to rid N2 worms of external bacteria. Worms were then transferred to Eppendorf tubes containing 400 mg of St. Si Carbide beads and 1 mL of M9. Tubes were vortexed for 1 min and then centrifuged at max rpm for 2 min. After, tubes were then quickly vortexed and serially diluted in M9 and plated on LB streptomycin (100 μg/mL) plates. Colony-forming units (CFUs) were then determined after overnight incubation at 37 °C. Tests were conducted in three independent trials, between three independent populations and CFUs were averaged per N2 worm. Unpaired student’s T-test (between two groups) was used for analysis, and a p-value < 0.05 was accepted as statistically significant.
RNA interference (RNAi) of skn-1 and mpk-1
The RNAi constructs targeting skn-1 was obtained from the C. elegans ORFeome RNAi library v1.151. The RNAi constructs targeting mpk-1 was constructed by PCR amplification of exon 6 (forward primer: 5′-ATGGCTAGCCACGTGATTACATAGTTCAATGCCTGAT-3′ and reverse primer: 5′-TCG ATGTCGATACGATTATGTGGATTGAAAGTG-3′) and exon 7 (forward primer: 5′-ATAATCGTATCGACATCGAGCAAGCA-3′ and reverse primer: 5′-GGTCGACGGTATCGATACTAAACAGGATTCTGCCCTC-3′) of the mpk-1 gene and cloned between the NheI and SalI restriction sites of the pL4440-DEST vector. The skn-1 and mpk-1 RNAi plasmids were introduced into E. coli HT115(DE3) by transformation. The resulting transformants were inoculated in LB plus tetracycline (12.5 μg/mL) and carbenicillin (25 μg/mL) and grown at 37 °C with shaking for 8–12 h. Thereafter, isopropyl β-d-1-thiogalactopyranoside (IPTG) was added into the growth culture with a final concentration of 2 mM to induce the expression of the interfering RNA. Four hours after the induction, the bacteria were harvested by centrifugation. Ten-fold concentrated RNAi bacteria were seeded onto RNAi plates containing 25 μg/mL of carbenicillin. E. coli HT115(DE3) with the empty vector pL4440-DEST was cultured the same way and seeded onto the control plates. Synchronized worm populations were acquired by allowing 10–15 hermaphrodites to lay eggs on NGM plates seeded with E. coli HT115(DE3) with the RNAi plasmid or the empty vector (EV) for 4 h at permissive temperature, and then the parents were removed. Eggs left on the plate would reach L4 after around 40 h. The L4 worms were then transferred to corresponding RNAi or EV E. coli seeded NGM-FUDR plates (with and without 1 nM of c-di-GMP). Lifespan assays were conducted and analyzed as described above.
Quantitative real-time PCR
All RNA samples were prepared using RNAzol RT reagent (Molecular Research Center, INC.) and stored at −80 °C. Complementary DNA was synthesized using the BioRad iScript cDNA synthesis kit (Bio-Rad). For C.elegans immune response genes, gravid worms were synchronized on NGM plates seeded with E. coli OP50. Worms were allowed to lay eggs for 4 h and then gravid adults were killed leaving only synchronous egg population. Once nematodes reached young adult stage, M9 buffer was used to wash collected worms. Worm populations were washed at least twice to remove any bacterium that was left in the media or on the worms. After this was completed, the population was split into 50 μg populations into separate sterile tubes. This is done to ensure an appropriate amount of RNA in each sample, as well as a high purity. One sample was treated with appropriate final concentration of 1 nM c-di-GMP, and the other sample served as a control.
qRT-PCR was performed using SensiFAST SYBR No-Rox Kit (Bioline) and the CFX96 real-time PCR detection system according to the manufacturer’s suggested protocol (Bio-Rad). The qPCR conditions were: 95 °C for 3 min, followed by 40 cycles of 10 s at 95 °C and 30 s at 60 °C. Transcription levels of act-1 were used as internal controls to normalize the expression levels of target transcripts. Relative fold changes are then to be calculated using the comparative CT (2−∆∆CT) method52. Cycle thresholds of amplification were determined by Light Cycler software (Bio-Rad). Each qPCR experiment was repeated three times using independent RNA preparations. The data were pooled and analyzed using unpaired Student’s T-test (between each target gene and the control act-1), and a p-value < 0.05 was accepted as statistically significant. The primers used in this study are listed in Table 2.
Statistics and reproducibility
Each assay was conducted in at least three biological replicates. One biological replicate is defined as using one batch of synchronized C. elegans population and/or one batch of freshly prepared bacterial culture. For the choice index, LC-MS, gut colonization, and qPCR assays, average and standard error were calculated from data points collected from all three biological replicates. Unpaired Student’s T-test (between two groups) and one-way balanced ANOVA (for more than two groups) were used for statistical analysis. A p-value < 0.05 was accepted as statistically significant. For the life span assays, statistical analysis was carried out through SPSS software under the Kaplan–Meier lifespan analysis for each replicate. The p-values were determined via log rank test, and a p < 0.05 was accepted as statistically significant. Analyses from all three biological replicates were compiled and similar results were obtained.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The authors declare that all data supporting the findings of this study are available within this article, Supplementary material, and Supplementary data set. Source data underlying the graphs in figures are provided in Supplementary data set.
Gray, K. M. Intercellular communication and group behavior in bacteria. Trends Microbiol. 5, 184–188 (1997).
Botsford, J. L. & Harman, J. G. Cyclic AMP in prokaryotes. Microbiol. Rev. 56, 100–122 (1992).
Zhu, M., Pan, Y. & Dai, X. p)ppGpp: the magic governor of bacterial growth economy. Curr. Genet. 65, 1121–1125 (2019).
Ross, P. et al. Regulation of cellulose synthesis in Acetobacter xylinum by cyclic diguanylic acid. Nature 325, 279–281 (1987).
Jenal, U., Reinders, A. & Lori, C. Cyclic di-GMP: second messenger extraordinaire. Nat. Rev. Microbiol 15, 271–284 (2017).
Purcell, E. B. & Tamayo, R. Cyclic diguanylate signaling in Gram-positive bacteria. FEMS Microbiol. Rev. 40, 753–773 (2016).
Fahmi, T., Port, G. C. & Cho, K. H. c-di-AMP: an essential molecule in the signaling pathways that regulate the viability and virulence of gram-positive bacteria. Genes 8, 197 (2017).
Davies, B. W., Bogard, R. W., Young, T. S. & Mekalanos, J. J. Coordinated regulation of accessory genetic elements produces cyclic di-nucleotides for V. cholerae virulence. Cell 149, 358–370 (2012).
Nelson, J. W. et al. Control of bacterial exoelectrogenesis by c-AMP-GMP. Proc. Natl Acad. Sci. USA 112, 5389–5394 (2015).
Hughes, D. T. & Sperandio, V. Inter-kingdom signalling: communication between bacteria and their hosts. Nat. Rev. Microbiol. 6, 111–120 (2008).
Woodward, J. J., Iavarone, A. T. & Portnoy, D. A. c-di-AMP secreted by intracellular Listeria monocytogenes activates a host type I interferon response. Science 328, 1703–1705 (2010).
Burdette, D. L. et al. STING is a direct innate immune sensor of cyclic di-GMP. Nature 478, 515–518 (2011).
Ablasser, A. et al. cGAS produces a 2′-5′-linked cyclic dinucleotide second messenger that activates STING. Nature 498, 380–384 (2013).
Beale, E., Li, G., Tan, M. W. & Rumbaugh, K. P. Caenorhabditis elegans senses bacterial autoinducers. Appl. Environ. Microbiol. 72, 5135–5137 (2006).
Werner, K. M., Perez, L. J., Ghosh, R., Semmelhack, M. F. & Bassler, B. L. Caenorhabditis elegans recognizes a bacterial quorum-sensing signal molecule through the AWCON neuron. J. Biol. Chem. 289, 26566–26573 (2014).
Dinh, J. et al. Cranberry extract standardized for proanthocyanidins promotes the immune response of Caenorhabditis elegans to Vibrio cholerae through the p38 MAPK pathway and HSF-1. PLoS ONE 9, e103290 (2014).
Ryjenkov, D. A., Tarutina, M., Moskvin, O. V. & Gomelsky, M. Cyclic diguanylate is a ubiquitous signaling molecule in bacteria: insights into biochemistry of the GGDEF protein domain. J. Bacteriol. 187, 1792–1798 (2005).
Tischler, A. D. & Camilli, A. Cyclic diguanylate (c-di-GMP) regulates Vibrio cholerae biofilm formation. Mol. Microbiol. 53, 857–869 (2004).
Bargmann, C. I. WormBook, ed. The C. elegans research community. WormBook https://doi.org/10.1895/wormbook.1.123.1 (2006).
Garigan, D. et al. Genetic analysis of tissue aging in Caenorhabditis elegans: a role for heat-shock factor and bacterial proliferation. Genetics 161, 1101–1112 (2002).
Portal-Celhay, C., Bradley, E. R. & Blaser, M. J. Control of intestinal bacterial proliferation in regulation of lifespan in Caenorhabditis elegans. BMC Microbiol. 12, 49 (2012).
Sahu, S. N. et al. Genomic analysis of immune response against Vibrio cholerae hemolysin in Caenorhabditis elegans. PLoS ONE 7, e38200 (2012).
Irazoqui, J. E. et al. Distinct pathogenesis and host responses during infection of C. elegans by P. aeruginosa and S. aureus. PLoS Pathog. 6, e1000982 (2010).
Sun, X., Chen, W. D. & Wang, Y. D. DAF-16/FOXO transcription factor in aging and longevity. Front. Pharm. 8, 548 (2017).
Murphy, C. T. et al. Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424, 277–283 (2003).
Garsin, D. A. et al. Long-lived C. elegans daf-2 mutants are resistant to bacterial pathogens. Science 300, 1921 (2003).
Oliveira, R. P. et al. Condition-adapted stress and longevity gene regulation by Caenorhabditis elegans SKN-1/Nrf. Aging Cell 8, 524–541 (2009).
Brunquell, J., Morris, S., Lu, Y., Cheng, F. & Westerheide, S. D. The genome-wide role of HSF-1 in the regulation of gene expression in Caenorhabditis elegans. BMC Genom. 17, 559 (2016).
Singh, V. & Aballay, A. Heat shock and genetic activation of HSF-1 enhance immunity to bacteria. Cell Cycle 5, 2443–2446 (2006).
Singh, V. & Aballay, A. Heat-shock transcription factor (HSF)-1 pathway required for Caenorhabditis elegans immunity. Proc. Natl Acad. Sci. USA 103, 13092–13097 (2006).
Morrison, D. K. MAP kinase pathways. Cold Spring Harb. Perspect. Biol. https://doi.org/10.1101/cshperspect.a011254 (2012).
Inoue, H. et al. The C. elegans p38 MAPK pathway regulates nuclear localization of the transcription factor SKN-1 in oxidative stress response. Genes Dev. 19, 2278–2283 (2005).
Okuyama, T. et al. The ERK-MAPK pathway regulates longevity through SKN-1 and insulin-like signaling in Caenorhabditis elegans. J. Biol. Chem. 285, 30274–30281 (2010).
Gao, X. et al. Functional characterization of core components of the Bacillus subtilis cyclic-di-GMP signaling pathway. J. Bacteriol. 195, 4782–4792 (2013).
Fagerlund, A. et al. Cyclic diguanylate regulation of Bacillus cereus group biofilm formation. Mol. Microbiol. 101, 471–494 (2016).
Fu, Y. et al. c-di-GMP regulates various phenotypes and insecticidal activity of gram-positive. Front. Microbiol. 9, 45 (2018).
Yang, Y., Li, Y., Gao, T., Zhang, Y. & Wang, Q. C-di-GMP turnover influences motility and biofilm formation in Bacillus amyloliquefaciens PG12. Res. Microbiol. 169, 205–213 (2018).
Hull, T. D. et al. Cyclic Di-GMP phosphodiesterases RmdA and RmdB are involved in regulating colony morphology and development in Streptomyces coelicolor. J. Bacteriol. 194, 4642–4651 (2012).
Peltier, J. et al. Cyclic diGMP regulates production of sortase substrates of Clostridium difficile and their surface exposure through ZmpI protease-mediated cleavage. J. Biol. Chem. 290, 24453–24469 (2015).
Gupta, K. R., Baloni, P., Indi, S. S. & Chatterji, D. Regulation of growth, cell shape, cell division, and gene expression by second messengers (p)ppGpp and cyclic Di-GMP in Mycobacterium smegmatis. J. Bacteriol. 198, 1414–1422 (2016).
Wu, X. et al. Molecular evolutionary and structural analysis of the cytosolic DNA sensor cGAS and STING. Nucleic Acids Res. 42, 8243–8257 (2014).
Chen, Q., Sun, L. & Chen, Z. J. Regulation and function of the cGAS-STING pathway of cytosolic DNA sensing. Nat. Immunol. 17, 1142–1149.
Ouyang, S. et al. Structural analysis of the STING adaptor protein reveals a hydrophobic dimer interface and mode of cyclic di-GMP binding. Immunity 36, 1073–1086 (2012).
Guzman, L. M., Belin, D., Carson, M. J. & Beckwith, J. Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J. Bacteriol. 177, 4121–4130 (1995).
Thelin, K. H. & Taylor, R. K. Toxin-coregulated pilus, but not mannose-sensitive hemagglutinin, is required for colonization by Vibrio cholerae O1 El Tor biotype and O139 strains. Infect. Immun. 64, 2853–2856 (1996).
Zhu, J. et al. Quorum-sensing regulators control virulence gene expression in Vibrio cholerae. Proc. Natl Acad. Sci. USA 99, 3129–3134 (2002).
Vance, R. E., Zhu, J. & Mekalanos, J. J. A constitutively active variant of the quorum-sensing regulator LuxO affects protease production and biofilm formation in Vibrio cholerae. Infect. Immun. 71, 2571–2576 (2003).
Zhu, J. & Mekalanos, J. J. Quorum sensing-dependent biofilms enhance colonization in Vibrio cholerae. Dev. Cell 5, 647–656 (2003).
Hosono, R. Sterilization and growth inhibition of Caenorhabditis elegans by 5-fluorodeoxyuridine. Exp. Gerontol. 13, 369–374 (1978).
Mitchell, D. H., Stiles, J. W., Santelli, J. & Sanadi, D. R. Synchronous growth and aging of Caenorhabditis elegans in the presence of fluorodeoxyuridine. J. Gerontol. 34, 28–36 (1979).
Rual, J. F. et al. Toward improving Caenorhabditis elegans phenome mapping with an ORFeome-based RNAi library. Genome Res. 14, 2162–2168 (2004).
Schmittgen, T. D. & Livak, K. J. Analyzing real-time PCR data by the comparative C(T) method. Nat. Protoc. 3, 1101–1108 (2008).
We would like to thank Dr. Jay (Jun) Zhu for generously providing all the V. cholerae strains and Dr. Andrew Camilli for providing the pAT1568 plasmid. This work was supported by the Clemson University Creative Inquiry fund and the Clemson SEED grant to M.C. and the National Institutes of Health and the Office of Dietary Supplements, R01HL130819, to Z.W.
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Angeloni, J., Dong, Y., Wang, Z. et al. Bacterial second messenger 3′,5′-cyclic diguanylate attracts Caenorhabditis elegans and suppresses its immunity. Commun Biol 3, 700 (2020). https://doi.org/10.1038/s42003-020-01436-9