Endoplasmic Reticulum Stress Pathway Required for Immune Homeostasis Is Neurally Controlled by Arrestin-1

Background: GPCRs function in the C. elegans nervous system to control immunity. Results: Arrestin-1, the only GPCR adaptor in C. elegans, functions in the nervous system to control immunity. Conclusion: Neural arrestin-1 signal regulates immunity. Significance: Our data underscore the importance of the nervous system in the control of longevity and immune homeostasis and suggest that overlapping and distinct neural circuits control these processes. In response to pathogen infection, the host innate immune system activates microbial killing pathways and cellular stress pathways that need to be balanced because insufficient or excessive immune responses have deleterious consequences. Recent studies demonstrate that two G protein-coupled receptors (GPCRs) in the nervous system of Caenorhabditis elegans control immune homeostasis. To investigate further how GPCR signaling controls immune homeostasis at the organismal level, we studied arrestin-1 (ARR-1), which is the only GPCR adaptor protein in C. elegans. The results indicate that ARR-1 is required for GPCR signaling in ASH, ASI, AQR, PQR, and URX neurons, which control the unfolded protein response and a p38 mitogen-activated protein kinase signaling pathway required for innate immunity. ARR-1 activity also controlled immunity through ADF chemosensory and AFD thermosensory neurons that regulate longevity. Furthermore, we found that although ARR-1 played a key role in the control of immunity by AFD thermosensory neurons, it did not control longevity through these cells. However, ARR-1 partially controlled longevity through ADF neurons.

Increasing evidence indicates that activation of the innate immune system accounts for the major physiological, metabolic, and pathological responses to infections (1). The response to microbial infection is accompanied by increased demand for protein folding in the endoplasmic reticulum (ER), 2 which must be alleviated by UPR pathways (2)(3)(4)(5)(6)(7)(8). The nervous system, which can respond in milliseconds to different environmental stimuli, has several characteristics that make it an ideal partner with the innate immune system to regulate nonspecific host defenses (9 -11).
To provide insights into the neural mechanisms that regulate innate immunity, we have taken advantage of the simple immune system and the very well characterized nervous system of Caenorhabditis elegans. A unique advantage to the study of neural mechanisms involved in the control of immune responses in C. elegans is the unparalleled characterization of its nervous system at the cellular level. Each of the 302 neurons of an adult animal has a precise identity, and, most uniquely, C. elegans is the only animal for which the wiring diagram of the nervous system has been established based on anatomical reconstructions. NPR-1, a GPCR similar to mammalian neuropeptide Y receptors, participates in a neural circuit that controls the p38/PMK-1 MAPK pathway required for innate immunity (12). In addition, OCTR-1, which is a catecholamine GPCR for octopamine, controls not only the p38/PMK-1 MAPK pathway but also a canonical and a noncanonical UPR pathway that may play a key role to alleviate ER stress during immune responses against microbial pathogens (7,8).
The discovery that individual GPCRs function in the nervous system of C. elegans to control the immune system (7,8,12) offers a unique opportunity to analyze the role of the entire nervous system in immune homeostasis. To study the role of neural GPCR signaling in the regulation of UPR genes and its role in immune defense against bacterial infections, we took advantage of a unique component of GPCR signaling, the GPCR adaptor protein arrestin-1 (ARR-1). Arrestins block G protein-mediated signaling and can also function as signal transducers in their own right (13). The only member of the arrestin family in C. elegans, ARR-1, is expressed almost exclusively within the C. elegans nervous system (14), making it easy to manipulate neural ARR-1 signaling and to study its effect on innate immune responses against pathogen infection.
Here we show that ARR-1 functions in a number of sensory neurons that regulate UPR genes that are expressed in nonneuronal tissues and required for immune defense. ARR-1 also partially controls longevity through ADF neurons. The results highlight the importance of ARR-1 signaling in a subset of sensory neurons that control longevity and immune homeostasis during response to pathogen infection.
C. elegans Killing Assay-C. elegans wild-type N2 animals and mutants were maintained as hermaphrodites at 20°C, grown on modified nematode growth medium (NGM) agar plates (0.35% instead of 0.25% peptone), and fed with E. coli OP50 as described. The bacterial lawns used for C. elegans killing assays were prepared by placing a 15-l drop of an overnight culture of the bacterial strains on modified NGM agar on plates 3.5 cm in diameter. Full lawn plates used for C. elegans killing assays were prepared by spreading a 25-l drop of an overnight culture grown at 37°C of P. aeruginosa on the complete surface of modified NGM agar in 3.5-cm-diameter Petri plates. Plates were incubated at 37°C for 12ϳ16 h. Plates were cooled down at room temperature for at least 1 h before seeding with synchronized young adult animals. The killing assays were performed at 25°C, and live animals were transferred daily to fresh plates. Animals were scored at the times indicated and were considered dead when they failed to respond to touch.
C. elegans Lifespan Assay-E. coli OP50 was grown as described above. A 50-l drop of the bacteria was plated on a 6-cm plate of modified NGM agar containing 40 g/ml fluorodeoxyuridine from Sigma. The assays were performed at 20°C.
Profile of Bacterial Accumulation in the Nematode Intestine-To determine the profiles of bacterial accumulation in the intestine, wild-type, arr-1(ok401), and octr-1(ok371) animals were synchronized. Synchronized L1 larvae were grown on E. coli OP50 at 20°C until they had reached the young adult stage. Wild-type, arr-1(ok401), and octr-1(ok371) animals were then transferred to plates seeded with P. aeruginosa expressing GFP (P. aeruginosa/GFP) cultured for 24 -40 h at 25°C. Animals were transferred to an NGM plate seeded with E. coli for 15 min and transferred again to a new NGM plate seeded with E. coli for 30 min to eliminate P. aeruginosa/GFP stuck to the body of the nematodes. Animals were visualized and imaged using a Leica MZ FLIII fluorescence stereomicroscope.
Quantification of Intestinal Bacterial Loads-For the quantification of cfu, wild-type, arr-1(ok401), AY103, AY104, AY105, AY106, AY107, and AY108 animals were synchronized by treatment of gravid adults with sodium hydroxide and bleach. Synchronized L1 larvae were grown on E. coli OP50 at 20°C until they had reached the young adult stage. Wild-type, arr-1(ok401), and transgenic animals were then transferred to plates seeded with P. aeruginosa/GFP for 24 h at 25°C. Animals were transferred to an NGM plate seeded with E. coli for 15 min to eliminate P. aeruginosa/GFP stuck to the body of the worms. Animals were transferred to a new NGM plate seeded with E. coli for 30 min to further eliminate external P. aeruginosa/ GFP. Ten nematodes/condition were transferred into 50 l of PBS plus 0.1% Triton X-100 and ground. Serial dilutions of the lysates (10 Ϫ1 , 10 Ϫ2 , 10 Ϫ3 , 10 Ϫ4 ) were plated onto LB/kanamycin to select for P. aeruginosa/GFP cells and grown overnight at 37°C. Single colonies were counted next day and represented as the number of bacterial cells or cfu per worm.
Confocal Microscopy-AY103 and AY104 worm strains were imaged using the 63ϫ lens of a Leica TCS SL confocal microscope. To construct the whole worm, the individual images were layered and flattened in Photoshop CS5.
RNA Isolation-Gravid adult wild-type, arr-1(ok401), and the transgenic animals were lysed using a solution of sodium hydroxide and bleach (ratio of 5:2) and washed, and the eggs were synchronized for 22 h in S basal liquid medium at room temperature. Synchronized L1 larval animals were placed onto NGM plates seeded with E. coli OP50 and grown at 20°C until the animals had reached the L4 larval stage. Animals were collected and washed with M9 buffer before transferring to NGM plates containing P. aeruginosa PA14 for 4 h at 25°C. After 4 h, animals were collected and washed with M9 buffer, and RNA was extracted using TRIzol reagent (Invitrogen). Residual genomic DNA was removed by DNase treatment (Ambion, Austin, TX).
Quantitative Real-time PCR (qRT-PCR)-Total RNA was obtained as described above. qRT-PCR was conducted using the Applied Biosystems One-Step Real-time PCR protocol using gene-specific Taqman assays on an Applied Biosystems 7900HT real-time PCR machine in 96-well plate format. Twenty five nanograms of RNA was used for each replicate. Relative -fold changes for transcripts were calculated using the comparative CT(2-⌬⌬CT) method (23) and normalized to actin-4. Cycle thresholds of amplification were determined by StepOnePlus software (Applied Biosystems). All samples were run in triplicate. Taqman assay information is available upon request and on the Applied Biosystems website.
Statistical Analysis-Animal survival was plotted as a nonlinear regression curve using the PRISM (version 4.00) computer program. Survival curves were considered different from the appropriate control indicated in the main text when p values were Ͻ0.05. Prism uses the product limit or Kaplan-Meier method to calculate survival fractions and the log rank test, which is equivalent to the Mantel-Heanszel test, to compare survival curves. A two-sample t test for independent samples was used to analyze cfu and qRT-PCR results; p values Ͻ0.05 are considered significant. All experiments were repeated at least three times unless otherwise indicated.

ARR-1 Signaling Regulates Pathogen Resistance and Lifespan
Extension by Targeting Different Pathways-We studied whether ARR-1 is required for C. elegans defense against bacterial infections by exposing arr-1(ok401) mutant animals to P. aeruginosa and by comparing their survival with that of wildtype animals. arr-1(ok401) animals exhibited an enhanced resistance to P. aeruginosa (ERP) phenotype (Fig. 1A), suggesting that the lack of inhibition of GPCR signaling in the entire nervous system due to a lack of ARR-1 activity enhances immunity. Wild-type and arr-1(ok401) animals exhibited similar pumping rates on P. aeruginosa, indicating that they are exposed to a comparable dose of pathogen (data not shown). To determine whether the enhanced immune response caused by mutation in the arr-1 gene is specific to P. aeruginosa, we exposed arr-1(ok401) animals to S. enterica and Y. pestis, two Gram-negative pathogens known to kill C. elegans (24,25). As shown in Fig. 1, C and D, arr-1(ok401) animals also exhibited an enhanced resistance to these pathogens.
Because pathogen avoidance is part of the C. elegans defense response to P. aeruginosa that is controlled by the nervous system (12), we examined the ERP phenotype of arr-1(ok401) animals on agar plates that were completely covered in bacteria, a condition that eliminates pathogen avoidance. arr-1(ok401) animals died at a slower rate than wild-type animals (Fig. 1B), indicating that pathogen avoidance does not play a role in the ERP phenotype of arr-1(ok401) animals. Consistent with this conclusion, arr-1(ok401) animals are resistant to S. enterica (Fig. 1C), a pathogen that does not elicit an avoidance behavior (26).
The extended lifespan of arr-1(ok401) animals grown on bacterial pathogens may simply be a consequence of their extended lifespan when grown on nonpathogenic E. coli (27). However, arr-1(ok401) mutants were more susceptible to the Gram-positive bacterium Enterococcus faecalis than wild-type animals (data not shown), making this possibility unlikely. In addition, although inhibition of the FOXO family transcription factor DAF-16 that controls longevity fully suppressed the extended lifespan of arr-1(ok401) animals (Fig. 1E), it only slightly suppressed the ERP phenotype of arr-1(ok401) animals (Fig. 1F). These results suggest that ARR-1 regulates different pathways involved in pathogen resistance and lifespan extension.
Neural ARR-1 Regulates Innate Immunity against Bacterial Pathogens-Expression of ARR-1::GFP under its own promoter restored ARR-1 expression in the nervous system (Fig. 2C) and rescued the ERP phenotype of arr-1(ok401) animals ( Fig. 2A), providing evidence for the role of neuronal ARR-1 signaling in the regulation of innate immunity. ARR-1::GFP expressed under the regulation of a pan-neuronal promoter unc-119 (Fig.  2F) also fully rescued the ERP phenotype of arr-1(ok401) animals (Fig. 2D). Native or pan-neuronal expression of ARR-1 also rescued the reduced accumulation of P. aeruginosa phenotype exhibited by arr-1(ok401) animals (Fig. 2, B and E).
To study the role of ARR-1 in ASH and ASI neurons to regulate immune homeostasis, we studied arr-1(ok401) animals carrying a deletion in octr-1. The ERP phenotype of arr-1(ok401);octr-1(ok371) animals was stronger than that of arr-1(ok401) or octr-1(ok371) animals (Fig. 3A). In addition, unlike octr-1(ok371) animals, arr-1(ok401) animals are more resistant to bacterial accumulation ( Fig. 3C and supplemental Fig. S2), indicating that the ERP phenotype of arr-1(ok401) animals is mediated through GPCRs other than OCTR-1. If OCTR-1 were the only GPCR blocked by ARR-1 in ASH and ASI neurons, the rescue of ARR-1 expression in these neurons would further enhance the ERP phenotype of arr-1(ok401) animals and the up-regulation of abu genes. However, no differences in survival or bacterial accumulation were observed between arr-1(ok401) and arr-1(ok401) animals expressing ARR-1 under the regulation of the sra-6 promoter (Fig. 3, B and C), which drives ARR-1 expression to OCTR-1-expressing neurons ASH and ASI. Neither did the expression of ARR-1 further enhance the up-regulation of abu genes in arr-1(ok401) animals (Fig. 3D). These results suggest that there may be another GPCR in ASH and ASI neurons whose signaling promotes immune homeostasis.

ARR-1 Signaling in ADF Chemosensory Neurons Regulates Both Pathogen Resistance and Lifespan Extension-Because
ARR-1 regulates DAF-16-mediated lifespan extension (27) and ADF sensory neurons control DAF-16 activation (29), we studied whether ARR-1 functions in ADF neurons to control either the lifespan extension or the ERP phenotype of arr-1(ok401) animals. We found that although ARR-1 expression under the regulation of the srh-142 promoter, which drives ARR-1 expression to ADF neurons, slightly rescued the extended lifespan of arr-1(ok401) animals (Fig. 5A), it rescued the ERP phenotype of arr-1(ok401) animals more strongly (Fig. 5B). In contrast, we found that daf-16 mutation fully suppressed the extended lifespan of arr-1(ok401) animals (Fig. 1E) and slightly suppressed the ERP phenotype of arr-1(ok401) animals (Fig.  1F). Thus, ARR-1 signaling in ADF neurons appears to play a more important role in the regulation of immunity against bacterial infections than in the control of lifespan. ARR-1 expres-sion in ADF neurons also partially rescued the enhanced expression of abu genes and reduced bacterial accumulation of arr-1(ok401) animals (Fig. 5, C and D), further supporting the function of ARR-1 signaling in ADF neurons to control immune homeostasis.
Regulation of Pathogen Resistance via ARR-1 Signaling in AFD Thermosensory Neurons-We hypothesized that ARR-1 signaling might control longevity and immune response by acting in thermosensory AFD neurons, as these neurons have been implicated in the control of not only temperature sensation and thermotaxis but also longevity (30 -32). We found that although ARR-1 activity driven to AFD neurons by the gcy-8 promoter had no effect on the extended lifespan of arr-1(ok401) animals (Fig. 6A), it strongly rescued their ERP phenotype (Fig.  6B). In addition, ARR-1 activity in AFD neurons strongly rescued the enhanced expression of abu genes and reduced bacterial colonization of arr-1(ok401) animals (Fig. 6, C and D).
These results indicate that ARR-1 signaling in these sensory neurons plays a crucial role in the control of immune homeostasis during bacterial infections.
We found that, in addition to ASH and ASI, ADF and AFD sensory neurons controlled the noncanonical UPR pathway, highlighting the importance of the nervous system in the control of this pathogen-induced pathway that may help alleviate the ER stress caused by the increased demand for protein folding that takes place during immune activation. Notably, the noncanonical UPR genes are strongly expressed in the pharynx and the intestine, which are the primary interfaces between host cells involved in immune responses and bacterial pathogens (7). This is consistent with the idea that the nervous system acts as a master regulator of microbial killing pathways and cellular stress pathways that need to be balanced as insufficient or excessive immune responses have deleterious consequences in the infected organism.
The C. elegans nervous system transmits signals from thermosensory neurons AFD to control the activity of a steroidsignaling pathway that controls longevity (32). Our results show that ARR-1 signaling in AFD neurons did not control longevity (Fig. 6A), suggesting that GPCRs may not participate in thermosensory signaling cascades involved in the control of longevity. However, ARR-1 in AFD neurons appeared to play a crucial role in the control of immunity (Fig. 6, B-D), suggesting that GPCR signaling and thermosensation in these cells may play an important role in response to pathogen infections. Fever is an ancient immune mechanism used by metazoans in response to microbial infections, and warm temperatures activate a conserved pathway involving HSF-1 that helps C. elegans fight bacterial infections (33). Whereas only homeotherms are capable of internally increasing the body temperature, both homeotherms and poikilotherms exhibit warmth-seeking behavior when infected. Our results raise the interesting possibility that AFD neurons may participate in circuits that integrate inflammatory cues and thermosensation to fight infections.
In summary, our results indicate that the C. elegans nervous system utilizes a network of sensory neurons that controls longevity and a stress-induced pathway that plays a key role in cellular homeostasis during responses against bacterial infections. Whereas ARR-1 signaling in both AFD and ADF neurons controls immunity, it only partially controls longevity through ADF neurons. The identification of GPCRs in these cells capable of sensing pathogens or the damage caused by infections or aging is an important area for further study.