Functional characterization of thioredoxin 3 (TRX-3), a Caenorhabditis elegans intestine-specific thioredoxin

elegans moderately protects against infection. ABSTRACT Thioredoxins are a class of evolutionary conserved proteins that have been demonstrated to play a key role in many cellular processes involving redox reactions. We report here the genetic and biochemical characterization of C. elegans TRX-3, the first metazoan thioredoxin with an intestine-specific expression pattern. By using GFP reporters we have found that TRX-3 is expressed in both the cytoplasm and nucleus of intestinal cells with a prominent localization at the apical membrane. Although intestinal function, reproductive capacity, longevity and resistance of trx-3 loss of function mutants to many stresses is indistinguishable from that of wild-type animals, we have observed a slight reduction in size and a minor reduction in the defecation cycle timing of trx-3 mutants. Interestingly, trx-3 is induced upon infection by Photorhabdus luminescens and Candida albicans and TRX-3 overexpression provides a modest protection against these pathogens . Together, our data indicate that TRX-3 function in the intestine is dispensable for C. elegans development but may be important to fight specific


INTRODUCTION
The thioredoxin system is one of the most important systems to maintain redox homeostasis in all eukaryotes [1]. Thioredoxins (TRX) are a class of small multifunctional 12-kDa proteins that are characterized by the redox active site sequence Trp-Cys-Gly-Pro-Cys (WCGPC) and a compact threedimensional structure consisting of a central core of -sheets surrounded by -helices with the active site located in a protrusion of the protein [1]. In many cases, thioredoxin modules are found as part of multidomain proteins. Thioredoxins act as general protein disulfide reductases reducing many different substrates and becoming inactive in the process by the oxidation of the two cysteine residues at the active site. Oxidized thioredoxins are reactivated by thioredoxin reductases (TrxR) at the expense of the reducing power of NADPH [2]. The Caenorhabtidis elegans genome codes for several thioredoxin family members although only a few of them have been characterized to date. TRX-1 is expressed in the bilateral sensory neuron ASJ and has been shown to regulate dauer formation and dietary restriction-mediated lifespan extension [3][4][5]. TRX-2 is a mitochondrial thioredoxin of unknown function that is induced when mitochondrial unfolded protein response is activated [6]. Both TRX-1 and TRX-2 have been shown to modulate the function of CEP-1 (the worm orthologue of the tumor suppressor p53) in neuronal integrity and lifespan [7]. A third member of the family is PNG-1, a peptide: N-glycanase that has an N-terminal thioredoxin domain and regulates axon branching [8,9]. Finally, DNJ-27 is the worm ortholog of human ERdj5 which has been recently shown to mediate some pathological phenotypes of C. elegans models of human neurodegenerative diseases [10].
We report here the genetic, cellular and biochemical characterization of C. elegans TRX-3, the first example of a thioredoxin protein with an intestine-specific expression in eukaryotes. The intestine is one of the largest organs in the nematode Caenorhabditis elegans and its correct functioning is critical for many different physiological processes. For example, the intestine is responsible for food digestion, nutrient uptake, synthesis and storage of macromolecules and elimination of the end products of metabolism [11]. Moreover, the posterior part of the intestine also acts as pacemaker in the worm to regulate the defecation cycles [11]. Hence, a tight regulation of metabolite trafficking in the intestine is essential to control the energy metabolism of the organism [11]. Additional evidence implicates the intestine in the aging process. For instance, mutants of the daf-16/FOXO transcription factor have shorter lifespan, which is rescued to wild type levels when DAF-16 activity is restored in the intestine [12]. Also, mutants of the intestinal peptide transporter pept-1 have been shown to further extend the long lifespan of the insulin receptor daf-2 mutants [13]. Being directly exposed to the environment, the intestine also constitutes a first line of defense against the infection of pathogens that colonize this organ [14] and against toxicants and pollutants ingested with the food [11]. To cope with these stresses, C. elegans intestine is equipped with a number of enzymatic defensive systems including heat-shock proteins, TRX-3, a novel intestine-specific thioredoxin in C. elegans 4 chaperones, detoxifying and antioxidant proteins.
Using a trx-3 loss of function mutant we demonstrate that TRX-3 is not essential for intestinal morphogenesis or for key cellular processes where the role of the intestine is crucial, further supporting the notion that the thioredoxin system is dispensable for C. elegans development. However, we have uncovered a potential role of TRX-3 as protective system against some pathogen infections.

MATERIALS AND METHODS
C. elegans strains and culture conditions -The standard methods used for culturing and maintenance of C. elegans were as described previously [15]. The strains used in this work are described in Supplementary Table 1. All experiments were performed at 20ºC unless otherwise noted. All VZ strains were 6x backcrossed with N2 wild type.
RNA extraction and RT-PCR analysis -For total RNA extraction, gravid hermaphrodites were washed off the plates with M9 buffer and dissolved in 5M NaOH bleaching solution. Embryos were collected and washed several times with M9 buffer. RNA was extracted from embryos using the NucleoSpin RNA II (Macherey-Nagel) kit following the manufacturer's instructions. The total RNA was DNase-treated using the Amplification Grade DNase I (Sigma) and 1µg of DNase-treated RNA was reverse transcribed in a 20 µL reaction mixture. cDNA was generated using the iScript™ cDNA Synthesis Kit (Biorad). 1µg of cDNA was used for RT-PCR reactions using MBL-Taq DNA Polymerase (Dominion-MBL) along with the corresponding primers for trx-3 and ama-1 genes (Supplementary Table 2) at a final concentration of 0.1 µM.
Recombinant protein expression and purification -trx-3 cDNA from N2 wild type and trx-3(tm2820) mutant was amplified with the forward primer 5´-CAGGGATCCGCTAAGAACTTTTTCTCCGG-3´ and the reverse primer 5´-GGCTGAATTCTTATTATGCACGGATTCTCTCG-3´ and cloned into the BamHI and EcoRI restrictions sites of the pGEX-4T-1 vector to generate the constructs GST-CeTRX-3 and GST-CeTRX-3, respectively. These constructs were used to transform the E. coli HMS174 strain and recombinant protein expression was induced by adding 1mM IPTG to a 100 ml LB medium bacteria culture of 0.5-0.7 OD supplemented with 0.1 mg/ml ampicillin and further incubating the cells at 25ºC and 200 rpm during 5 hours. Cells were collected by centrifugation, immediately resuspended in 5 ml PBS containing 3 mg lysozyme, 0.5mg DNase and 0,5mg RNase and incubated for 10 min at room temperature with gently shaking. Next, the preparation was sonicated for 15 min on ice and the cell free extract was obtained by centrifugation at 8000 x g during 30 min at 4ºC. Recombinant GST-CeTRX-3 and GST-CeTRX-3 proteins were purified from the cell free extract using a Glutatione Sepharose TM 4B Affinity column (GE healthcare) equilibrated with PBS and eluted with 50 mM Tris-HCl, pH 8, 10 mM GSH. Finally, the purified protein was concentrated using Nanosep ® Centrifugal Devices equipped with a

TRX-3, a novel intestine-specific thioredoxin in C. elegans
CeTRX-3 proteins was tested by their ability to reduce bovine insulin A and B chains (Sigma, St. Louis, Missouri) either using DTT or NADPH (Sigma, St. Louis, Missouri) and rat thioredoxin reductase-1 (IMCO, Stockholm, Sweden) as electron donors, as previously described [16], with slight modifications.
Briefly, for the DTT assay, 25 l of a reaction mix (composed of 40 l Tris-HCl 1M pH 7.5, 10 l EDTA 0.2 M and 200 l of bovine insulin 10 mg/ml) were mixed with the protein preparation in a final assay volume of 200 l. The reaction was initiated by adding 2l of DTT 100 mM and the thioredoxin activity was measured by monitoring the increase of absorbance at 595 nm due to insulin free-B chain precipitation over time. For the NADPH and thioredoxin reductase assay, 20 l of a reaction mix Developmental analyses -For the E blastomere cell lineage analysis, 4D-microscopy was carried out using standard live-animal mounting techniques on a Leica DM6000 microscope fitted with DIC optics.
The use of DIC optics allows cells tracing without using any dye or fluorescent marker that might alter the cell cycle progression. Embryonic cell lineage was determined as described [17]. In summary, gravid hermaphrodites were dissected and 2-to 4-cell stage embryos were mounted on 4% agar pads in water, and sealed with Vaseline. Imaging was performed at 25ºC. The multi-focal time-lapse microscopy of the samples was controlled with the open source software Micro-manager (www.micro-manager.org).
Pictures on 30 focal planes (1micron/section) were taken every 30 seconds for 12 h. Embryo lineages were analyzed with the software SimiBiocel [18]. For worm size determination, eggs from animals grown at 25°C were harvested and allowed to hatch overnight in the absence of food. The synchronized larvae were transferred to NGM plates seeded with OP50 and kept 48 hours at 25°C. Then, the animals on the plates were analyzed with the Union Biometrica COPAS automated sorter following the manufacturer's instructions. The size of the worm (also known as Time Of Flight: TOF) is generated in arbitrary, but constant units and indicates the time required for the animal to travel through the measurement cell. Rotor-Gene 6000 system. The results were normalized to act-1 and then the relative expression was calculated. Control and experimental conditions were tested in the same run as technical triplicates. At least 2 independent experiments were performed. Longevity assays -Lifespan assays were performed at 25ºC as previously described [22] with slight modifications. Tightly synchronized embryos from bleached gravid adult hermaphrodites were allowed to develop through the L4 larval stage and then transferred to fresh NMG plates in groups of 25 worms per plate for a total of 100 individuals per experiment. The day animals reached the L4 larval stage was used as t = 0. Nematodes not carrying the fer-15 (b26) mutation were transferred to fresh plates daily until progeny production ceased and after that were transferred every second to third day but monitored daily for dead animals. Nematodes that did not respond to gentle prodding and displayed no pharyngeal pumping were scored as dead. Animals that crawled off the plate or died due to internal hatching or extruded gonad were censored and incorporated as such into the data set. All longevity assays were repeated two times.
Briefly hypochlorite-purified eggs of each worm strain were spotted onto 60 mm plates and incubated at 20ºC and 25ºC. OP50 bacteria were not spread to the edges of the plates in order to minimize the number of dauer larvae crawling off the plate. Dauer and non-dauer larvae were then counted as the first nondauers reached egg-laying age.
Pathogen killing assays -For Pseudomonas aeruginosa PA14, killing assays were performed as previously described [24], with slight modifications. We plated 10 ml of a saturated culture of P.
aeruginosa PA14 in LB modified NGM plates and incubated for 24 h at 37ºC followed by another incubation at 25ºC for 24 h. Killing assays were carried out in triplicate and performed by manually TRX-3, a novel intestine-specific thioredoxin in C. elegans transferring L4-staged animals from E. coli OP50 plates to pathogen plates. The killing plates were incubated at 25ºC and worm mortality was monitored at various points along the time course of death. For Serratia marcescens strain Db10 and Photorhabdus luminescens strain Hb, killing assays were performed as described [25] and repeated three times. For Candida albicans strain SN148 killing assays, young adult C. elegans were transferred from E. coli OP50 plates onto unseeded Brain Heart Infusion (BHI) plates, then onto BHI plates containing 7.7µM Kanamycin and 100µM FUDR seeded with a lawn of C. albicans.
C. elegans survival was monitored at 25 o C. We used the GraphPad Prism Software Package for graphical display and statistical analysis.
TRX-3, a novel intestine-specific thioredoxin in C. elegans

C. elegans TRX-3 is a functional thioredoxin specifically expressed in the intestine -The C. elegans
genome encodes for nine members of the thioredoxin system with the conserved active site sequence WCGPC (Supplementary Table 3 ). The tm2820 allele is a 340 bp deletion that completely removes the third exon ( Figure 1A) and an RT-PCR analysis identified a truncated mRNA species in extracts from the tm2820 mutant ( Figure   1B). The conceptual translation of the tm2820 mRNA variant produces an aberrant protein (TRX-3) that lacks amino acid residues 51 to 142, which correspond to those composing exon 3 (Supplementary Figure   1B), whose absence is predicted to dramatically alter the three-dimensional structure of the native TRX-3 protein (Supplementary Figure 1C). Consistent with this prediction, recombinant TRX-3 is devoid of enzymatic activity both in the DTT and the NADPH/thioredoxin reductase assays while native TRX-3 is fully active ( Figure 1C and 1D). Therefore, we conclude that trx-3(tm2820) is a loss of function allele. However, subtle phenotypes were identified in trx-3(tm2820) animals such as a smaller size ( Figure 2E) and a shorter defecation cycle ( Table 1). One of the pathways that regulates body growth in C. elegans is the TGF- Sma/Mab pathway. Worms unable to synthesize the ligand DBL-1 or any of the downstream TRX-3, a novel intestine-specific thioredoxin in C. elegans components of the pathway are dwarf. However, animals that overexpress DBL-1 become long.
Moreover, DBL-1 has also been shown to regulate the development of male tail structures and worm innate immunity [26,27]. Thus, we asked whether the small size of trx-3(tm2820) mutants could be caused by regulation of trx-3 expression by TGF- Sma/Mab signaling. For that purpose, we crossed the Ptrx-3::mCherry reporter strain with a dbl-1 null mutant and a dbl-1 overexpressing strains. As shown in Figure 2F, both genetic backgrounds cause a downregulation of the Ptrx-3::mCherry reporter, suggesting that proper DBL-1 levels are needed to keep a regular trx-3 expression.
Importantly, the absence of any developmental or morphological defect in trx-3(tm2820) mutants is not a consequence of redundancy with any of the two other described worm thioredoxins, trx-1 and trx-2 [3,6]. Hence, double mutants trx-1(ok1449); trx-3(tm2820) and trx-2(tm2720); trx-3(tm2820) are viable, do not display any obvious developmental or morphological phenotype and have a lifespan comparable to that of the single trx-1 or trx-2 mutants (Supplementary Figure 2). Furthermore, we failed to identify any developmental or morphological synthetic interaction of trx-3(tm2820) mutants when treating them with RNAi targeting all known members of the thioredoxin, and closely related glutaredoxin and peroxiredoxin protein family members in C. elegans (Supplementary Table 4), making it unlikely that trx-3 is functionally redundant with one of these genes. However, it remains possible that RNAi did not completely ablate the expression of these genes or that several of these genes may be functionally redundant with trx-3.
Given the prominent expression of TRX-3 in the apical intestinal membrane, we next asked whether TRX-3 function is required to maintain its structural integrity or the proper localization of other apical markers. As shown in Figure 3A-D, the continuity of the apical membrane is maintained in a trx-3(tm2820) mutant background, demonstrated by the normal apical distribution of well-known apical markers such as VHA-6, a vacuolar ATPase [28], or CAV-2, a member of the caveolin family [29].
Interestingly, caveolins have been shown to interact with the thioredoxin system in mammals [30]. Given that both TRX-3 and CAV-2 are located at the intestinal apical membrane, we asked whether trx-3 and cav-2 mutants display any synthetic genetic interaction. However, trx-3(tm2820); cav-2(tm394) double mutants had no obvious phenotype except for the slight brood-size reduction already reported for the cav-2(tm394) single mutant (data not shown) [29]. An important function of the intestine is the uptake of nutrients [11] and, because of the apical localization of TRX-3, we asked whether the luminal content uptake was compromised in trx-3(tm2820) animals. To this purpose, we fed worms with the fluorescent marker Texas Red BSA [29] and demonstrated that trx-3(tm2820) mutants incorporate the luminal marker at a similar rate as the wild type control ( Figure 3E-F), suggesting that TRX-3 is not essential for luminal endocytosis and that the apical membrane functionality is not compromised in these mutants.
Consistently, fat content and distribution in trx-3(tm2820) mutants was comparable to those of wild type control worms ( Figure 3G-H). We therefore conclude that TRX-3 is not essential for the intestinal TRX-3, a novel intestine-specific thioredoxin in C. elegans 13 function under normal growth conditions.

trx-3 does not play a protective role against chemical or heat stress in C. elegans -The C. elegans
intestine is constantly exposed to chemical stressors and toxins as well as pathogens [11] and thioredoxins are well known antioxidant proteins which have been shown to function as protective systems against different types of stresses [31]. Thus, we decided to test whether TRX-3 plays a role as anti-stress defense mechanism in the C. elegans intestine. With this objective, we first studied whether trx-3(tm2820) mutants modified the induction or subcellular distribution patterns of well-known GFP markers of stress such as DAF-16, HSP-16.2 or GST-4. DAF-16 is a foxo type transcription factor that normally resides inactive in the cytoplasm but translocates into the nucleus upon stress where it mediates the transcription of many genes involved in stress defense [32]. Importantly, DAF-16 function in intestine has been linked to fat storage and longevity [12]. HSP-16.2 is a small heat-shock protein, direct target of DAF-16, that is activated upon thermal and oxidative stress [33]. GST-4 is a glutathione S-transferase that has been shown to be induced upon exposure to diverse toxic chemical such as acrylamide or methylmercury [21,34]. As shown in Figure 4A, heat shock or acrylamide treatments did not alter the nuclear translocation of DAF-16 or induction of HSP-16.2 and GST-4 GFP reporters in trx-3(tm2820) animals as compared with the age-matched wild type controls. In addition, trx-3(tm2820) mutants subjected to heat-shock and chemically induced oxidative stress by paraquat and juglone did not show enhanced lethality either ( Figure 4B-D). As shown in Figure 4E Consistent with these results, the redox status of trx-3(tm2820) mutants is not significantly different from that of wild type controls, as determined by the levels of oxidized versus reduced PRDX-2 [35,36] or by using the genetically encoded fluorescent biosensor HyPer [37] (Supplementary Figure 3). As a whole, it seems unlikely that this lack of enhanced sensitivity to stress associated with loss of trx-3 is due to a compensatory mechanism involving other closely related stress-response genes. Instead, our data indicate that TRX-3 does not play a key role in the general anti-stress response in C. elegans.
daf-2 dependent phenotypes are not regulated by trx-3 -Compromised nutrient uptake has been associated to increased lifespan. This is illustrated by the extended longevity of eat-2 mutants that have reduced food intake due to decreased pharyngeal pumping [38]. In this context, reduced function of nhx-2, a nutrient transporter located in the intestinal apical membrane, promotes lifespan extension [39].
Moreover, the longer lifespan of daf-2 insulin receptor mutants is further enhanced by mutations in pept-1, another peptide transporter [13]. This last study also showed that trx-3 mRNA is induced 5.65-fold in daf-2 single mutants and 7.56-fold in a pept-1; daf-2 double mutant background (Britta Spanier, personal TRX-3, a novel intestine-specific thioredoxin in C. elegans communication). For this reason, we decided to study if trx-3 had any impact on daf-2 dependent traits such as longevity, dauer formation or survival against Pseudomonas aeruginosa PA14 infection [40,41].
Surprisingly, despite the remarkable induction of trx-3 mRNA in a daf-2 background, trx-3(tm2820) mutants did not show any effect on any of these traits either alone or in combination with a daf-2(e1370) mutation ( Figure 5).

C. elegans trx-3 is specifically induced upon Photorhabdus luminescens and Candida albicans infection but it is not essential for pathogen infection survival -The restricted expression of trx-3 in C.
elegans intestinal cells prompted us to examine the effect of different pathogens on trx-3 mRNA expression (Supplementary Table 5). From this survey, we found that trx-3 mRNA is induced upon infection by the bacteria Photorhabdus luminescens Hb and the fungi Candida albicans [25,42]. These results, together with our own killing assays with Pseudomonas aeruginosa PA14 ( Figure 5B) suggest that trx-3 has a role in response to specific pathogens. To further explore the role of trx-3 in the infection by these pathogens we first confirmed by qPCR the trx-3 mRNA induction upon exposure to P.
luminescens but not Serratia marcescens ( Figure 6A) as previously reported [25]. This induction also occurs at the protein level ( Figure 6B) and appears to be specific for trx-3. Hence, the mRNA levels of other related members of the thioredoxin and glutaredoxin family expressed in the intestine such as trx-2, glrx-21 and glrx-22 are not altered either in wild type or trx-3(tm2820) mutant backgrounds ( Figure 6C).
It is important to note that trx-3 mRNA induction upon P. luminescens infection is similar to that of the lysozyme lys-2, a well-known marker for intestinal infection in C. elegans [43]. Indeed, lys-2 induction is further enhanced in a trx-3(tm2820) background, suggesting that worms lacking trx-3 are already sensitized to P. luminescens infection ( Figure 6C) while trx-3(tm2820) mutants grown on the nonpathogenic OP50 bacteria have 25% lower levels of lys-2 compared to the wild type control (data not shown). However, similar to what we have found with P. aeruginosa infection ( Figure 5B), trx-3(tm2820) mutants do not display enhanced sensitivity to killing by P luminescens or S. marcescens ( Figure 6D-E), probably due to the induction of lys-2 and other antimicrobial defenses. In turn, when overexpressed, TRX-3 significantly increased the survival of animals infected with P. luminescens but not to S. marcescens infection ( Figure 6F-G). [As no antibodies are available for C. elegans TRX-3, we were unable to evaluate the increase of total TRX-3 in vzEx96 overexpressing worms compared to wild type controls. However, we determined the trx-3 mRNA levels by qPCR, which resulted in a 70.92 ± 6.73 fold-induction (mean ± SEM, two independent experiments with three replicates each)].
trx-3 mRNA expression is also rapidly induced upon exposure to the fungus Candida albicans infection [42] but not upon infection by other fungi such as Drechmeria coniospora and Harposporium sp. [25](Supplementary Table 5). Although, in consonance with previous microarray analyses [42], we have confirmed that the induction of trx-3 mRNA expression occurs rapidly (4-6 hours) after exposure to TRX-3, a novel intestine-specific thioredoxin in C. elegans C. albicans ( Figure 7A), trx-3(tm2820) mutants do not display enhanced sensitivity to C albicans killing ( Figure 7B). However, similar to P. luminescens, there was a suggestion that overexpression of TRX-3 affords some slight protection against C. albicans-mediated killing. Although, in the case of C. albicans, this increased resistance was not statistically significant ( Figure 7C), as a whole, our data indicate that TRX-3 may protect against infection by certain, specific pathogens. Finally, we asked whether overexpression of TRX-3 might also have a protective effect against stress or worm longevity. As shown in Figure 8, high levels of TRX-3 produced a small, although not significant, increase in lifespan and resistance against heat stress but had no effect on paraquat resistance. This is consistent with other data ( Figures 6G and 7C) in suggesting that TRX-3 does not play a major stress-protective role in C. elegans.
TRX-3, a novel intestine-specific thioredoxin in C. elegans

DISCUSSION
Thioredoxins play a key role in the maintenance of redox homeostasis in most organisms and, therefore, this family of proteins is well conserved through evolution. As the organism complexity increases from bacteria to metazoa, three levels of diversity are identified within the thioredoxin family: 1) tissue/organ specific expression, 2) specific subcellular localizations and 3) proteins containing one or more thioredoxin modules within a multidomain protein organization. In mammals, tissue-specific thioredoxins have only been identified in spermatozoa and lung ciliated epithelial cells [44]. Germ-cell specific expression of thioredoxins has also been reported in Drosophila [45]. Interestingly, we have identified novel tissue-specific expression patterns in the C. elegans thioredoxin family. Hence, TRX-1 is exclusively expressed in ASJ neurons [3] and TRX-2 is expressed in the mitochondria of ASEL and

AIYL/R neurons and muscle cells under non-stress conditions and also in intestinal cells upon induction
of the mitochondrial unfolded protein response [6]. We report here the characterization of C. elegans TRX-3 which, to our knowledge, is the first metazoan thioredoxin with a tissue-specific expression pattern restricted to intestine.
trx-3(tm2820) loss of function mutants do not show any gross developmental or morphological phenotypes or possess any phenotypes related to intestinal function under normal growth conditions. A possible explanation for the absence of a major phenotype in trx-3(tm2820) mutants could be a functional redundancy with other thioredoxins or closely related proteins such as glutaredoxins [31]. We have ruled out this possibility for trx-1 and trx-2 as double mutants of these two genes with trx-3(tm2820) are also viable, with no apparent phenotype. Moreover, we failed to identify any synthetic interaction when downregulating all known members of the thioredoxin, glutaredoxin and peroxiredoxin protein family in a trx-3(tm2820) mutant background. As RNAi feeding penetrance can be highly variable depending on genetic backgrounds and tissue expression, combinations of other thioredoxin, glutaredoxin and peroxiredoxin system mutants with that of trx-3 will be needed to unequivocally identify the redundant system. In addition, we cannot rule out that the redundant system could be a more distant member of the thioredoxin/glutaredoxin family not included in our RNAi screen. Alternatively, the absence of phenotype of trx-3(tm2820) mutants could be also explained by the possibility that the aberrant TRX-3 protein produced by the trx-3(tm2820) allele could retain some of the wild type TRX-3 functionality despite of being inactive in enzymatic activity assays, in vitro. Finally, the lack of phenotype might reflect a nonessential function of trx-3 under non-stressed conditions. Therefore, we asked if trx-3 might be required when the animals are stressed. This idea made sense as the worm intestine (together with the hypodermis) is directly exposed to the variable and sometimes harsh environmental conditions [11] and because thioredoxins have been long known to have a protective function against different types of stress [31]. However, trx-3(tm2820) mutants did not show enhanced TRX-3, a novel intestine-specific thioredoxin in C. elegans sensitivity to heat shock or oxidative stress treatments and did not induce expression of intestinal stress markers. Intestinal cells are also the main target of bacterial and fungi pathogens and, indeed, trx-3 seems to be specifically upregulated upon infection by certain bacterial pathogens. such as Photorhabdus luminiscens Hb [25] but not by other bacteria like Pseudomonas aeruginosa PA14 [46], Serratia marcescens Db10, Enterococcus faecalis [25], Staphylococcus aureus RN6390 [47] or Bacillus thuringiensis NRRL B-18247 [43]. Similarly, trx-3 mRNA expression is induced upon Candida albicans infection [42] but not upon other fungi infection such as Drechmeria coniospora and Harposporium sp.
The induction of trx-3 expression after P. luminescens and C. albicans infection is very rapid ( [25,42] and Fig.6A, 7A) and, interestingly, in the case of C. albicans, this trx-3 upregulation is independent of the infective capacity of the pathogen, as heat-killed C. albicans (which is avirulent) induces trx-3 mRNA expression to a similar extent to live C. albicans [42]. These data point to the possibility that trx-3 is induced in response to the presence of specific molecules on the surface of the pathogen (named pathogen-associated molecular patterns, PAMPs). PAMPs recognition is the earliest event in the C.
elegans response against pathogen infection, which is triggered independently of its infective capabilities [48]. Indeed, we have found that lys-2 is induced in trx-3(tm2820) mutants ( Figure 6C), which might compensate for the loss of trx-3, along with other non identified protective genes.
TRX-3 protein is composed of 158 amino acid residues, contains the conserved redox active site WCGPC and, surprisingly, appears to be restricted to the Caenorhabditis genus. Interestingly, a phylogenetic analysis demonstrates that TRX-3 clusters with invertebrate and vertebrate nucleoredoxins and nucleoredoxin-like proteins [49]. Nucleoredoxins are typically of similar size to that of C. elegans TRX-3 but differ in the sequence of their redox active site where the intervening residues between the two cysteines are different to those of classical thioredoxins. In turn, the previously described C. elegans TRX-1 and TRX-2 [3,6] are found clustering together with yeast, Drosophila and human thioredoxins in a different branch of the phylogenetic tree [49]. Thus, given the close phylogenetic proximity of C.
elegans TRX-3 with nucleoredoxins and related proteins, a possible role of vertebrate nucleoredoxins in TRX-3, a novel intestine-specific thioredoxin in C. elegans host defense against infection arises as an interesting hypothesis. This idea further strengthens the relevance of our findings in C. elegans by its potential application to human pathologies.
Nematodes produce reactive oxygen species (ROS) by the dual oxidase Ce-Duox1/BLI-3 as defensive mechanism against diverse pathogens infections [50,51]. Unavoidably, this ROS production also causes a cellular damage in the host tissue, which is counteracted by the induction of a general oxidative stress response mainly orchestrated by the SKN-1 transcription factor [52,53]. Despite thioredoxins are considered as general antioxidant enzymes [31], it is unlikely that trx-3 is part of this protective mechanism as trx-3(tm2820) mutants are as sensitive as wild type animals to the killing by P. luminescens and C. albicans, and trx-3(tm2820) mutants do not show enhanced sensitivity to ROS producing chemicals such as juglone or paraquat. Moreover, as mentioned above, avirulent, heat-killed C. albicans also induces trx-3 to a similar extent of that achieved by the infective, alive fungi. However, we can not rule out other roles for trx-3 in gut immunity such as the establishment of the necessary redox environment for an efficient immune response or a direct effect toward microbial virulence factors.
Several signaling pathways have been identified to mediate the intestinal response to pathogen infection. These pathways include the insulin-like, TGF-, Toll-receptor, P38, ERK and JNK MAPK pathways as well as pathways regulated by ELT-2 and HSF-1 transcription factors [54]. It is now well established that C. elegans elicits distinct responses depending on the pathogen to which they are exposed [55], although the mechanisms underlying this differential response are still poorly understood. For instance, C. elegans infection by C. albicans has been shown to mainly induce genes under the control of the p38 MAPK signaling pathway [42] while P. luminescens infection causes a significant elevation of genes under the control of both P38 MAPK and TGF- pathways [55]. We have found that trx-3 expression is regulated by dbl-1, a TGF- ligand involved in worm size control, development of male tail structures and worm innate immunity [26,27]. Interestingly, trx-3 appears to be tightly regulated by DBL-1 levels as both dbl-1 mutation and overexpression cause a significant trx-3 downregulation. Thus, it is plausible that trx-3 induction in response to P. luminescens or C. albicans exposure may be controlled by the TGF- signaling pathway.
In summary, we describe here the functional characterization of the first metazoan thioredoxin with an expression pattern restricted to intestinal cells. Although no major phenotype has been found in trx-3 RNAi downregulated worms or trx-3(tm2820) mutants, the fact that trx-3 is induced upon exposure to specific bacterial and fungal pathogens suggests that TRX-3 could be part of the molecular mechanism mounted to counteract this insult. Our results identify TRX-3 as a promising candidate to delve further into the molecular pathways governing the differential innate immune response of C. elegans and, by extrapolation, the role of mammalian thioredoxins in intestinal function and pathogen infection.    wild type and trx-3(tm2820) mutants were assayed for their resistance to different stress treatments such as heat shock at 37ºC (B), juglone 240 mM (C) and paraquat 4 mM (D). Graphs represent the average of 3 independent experiments. All treatments (except heat-shock) were carried out at 20ºC. Error bars indicate the standard error of the mean (SEM). daf-2 (e1370) and mev-1 (kn1) mutants were used as resistant and sensitive controls, respectively. Differences between N2 wild type and trx-3(tm2820) animals were found not significant in all cases by One-Way ANOVA tests (p > 0.05). (E) Quantification of trx-2, trx-3, glrx-21 and glrx-22 mRNA levels by qPCR of N2 wild type animals and trx-3(tm2820) grown on E. coli OP50. Bars represent the percentage of each mRNA species ± SEM in the two different genetic backgrounds calculated from 3 independent assays. Differences were non significant in all cases by two-tail Student's t-test (p > 0.05). backgrounds. The fer-15 (b26) mutation was included to prevent internal hatching of the daf-2 (e1370) progeny and has been shown not to impact the longevity [56]. Kaplan-Meier plots were used to show the fraction of animals that survive over time. Longevity and infection survival assays were performed twice, obtaining similar results, and the composite data is shown. The survival rate of trx-3(tm2820) animals was compared to that of their respective controls wild type for trx-3 using the log-rank (Mantel-Cox) test and the differences were found not significant in all cases (p>0.05). When comparing daf-2 backgrounds versus their corresponding non daf-2 controls, the differences were highly significant (*** p<0.001). (C) Effect of the trx-3(tm2820) allele on the formation of dauers by daf-2 (e1370) animals. Assays were performed at 20ºC and 25ºC on ten independent plates initiated with 100 to 150 eggs per plate. Numbers above the bars show the percentage of dauers and the numbers in parentheses indicate the total number of animals scored. Error bars indicate SD. Chi-squared test was used to determine statistical significance (ns = not significant; * p < 0.05). luminescens Hb and and trx-3(tm2820) animals grown on P. luminescens Hb. Bars represent the percentage of each mRNA species ± SEM in the two different genetic backgrounds calculated from 3 independent assays. Two-tail Student's ttest was used to determine statistical significance (* p < 0.05; *** p < 0.001)(D-E) The survival of the trx-3(tm2820) mutant to infection by S. marcescens Db10 (C) and P. luminescens Hb (D) compared to that of non-pathogenic E. coli OP50 was assayed at 25ºC. Kaplan-Meier plots were used to show the fraction of animals that survive against time. Infection survival assays were performed three times, obtaining similar results, and one representative experiment is shown. The survival rate of trx-3(tm2820) mutants was compared to that of the N2 wild type control using the log-rank (Mantel-Cox) test and was found not significant in both cases (p > 0.05). (F-G) The survival of the TRX-3 overexpressing strain to infection by S. marcescens Db10 (F) and P. luminescens Hb (G) was performed as described above for the trx-3(tm2820) mutant. Two independent experiments were performed and one representative experiment is shown. The survival rate of the TRX-3 overexpressing strains was compared to that of the N2 wild type control using the log-rank (Mantel-Cox) test and was found significantly different in both cases (p < 0.05).

Figure 7.
(A) Quantification of trx-3 mRNA levels by qPCR of N2 wild type animals grown on E. coli OP50 and Candida albicans. Bars represent the average fold-change expression of 2 independent assays ± SEM. Two-tail Student's t-test was used to determine statistical significance (ns = not significant; ** p < 0.01). (B-C) The survival of the trx-3(tm2820) mutant (B) and the TRX-3 overexpressing strain (C) to infection by Candida albicans compared to that of non-pathogenic E. coli OP50 was assayed at 25ºC. Kaplan-Meier plots were used to show the fraction of animals that survive against time. Infection survival assays were performed two times, obtaining similar results, and one representative experiment is shown. The survival rate of the trx-3(tm2820) mutants and TRX-3 overexpressing strain was compared to that of the N2 wild type control using the log-rank (Mantel-Cox) test and was found not significant in both cases (p > 0.05). Error bars indicate the standard error of the mean (SEM). Differences between strains were found not significant in all cases by One-Way ANOVA tests (p > 0.05). (C) Longevity of trx-3(tm2820) mutants and trx-3 overexpressing worms (OE) from the vzEx96 array on E. coli OP50 was assayed at 25ºC. Kaplan-Meier plots were used to show the fraction of animals that survive over time. Longevity assays were performed twice, obtaining similar results, and the composite data is shown. The survival rate was obtained using the log-rank (Mantel-Cox) test and the differences were found not significant in all cases (p > 0.05). The pBoc (peak of the posterior body muscle contraction) was monitored at 20ºC in L4 animals crawling on a lawn of OP50 bacteria on agar plates. pBoc rhythmicity of individual worms was determined by calculating the coefficient of variance (CV), which is defined as the standard deviation of the pBoc period expressed as percentage of the mean from 12 successive contractions. The data represent the CV mean ± SD of six independent trials each composed of 20 independent animals. c The length of the worms was determined using the COPAS Biosorter. The mean length ± SD of three independent trials with approximately 500 worms of each genotype is shown. This phenotype not caused by a delayed growth, as demonstrated by a similar egg-laying time between wt and tm2820 animals. d L1 synchronized larvae (generated by a 16-hours incubation in M9 buffer of hypochlorite-purified eggs) were grown at 20ºC in OP50 agar plates and the time at which the first eggs are laid was determined. The mean of the egg laying time ± SD from 30 worms of each genotype is shown.