Elucidating the Function of an Ancient NF-κB p100 Homologue, CrRelish, in Antibacterial Defense

Abstract

The family of NF-κB transcription factors essentially regulates immune-related gene expression. Recently, we isolated and characterized the classical NF-κB/inhibitor κB (IκB) homologues from a “living fossil,” the horseshoe crab, Carcinoscorpius rotundicauda. Interestingly, this ancient species also harbors another class I NF-κB p100 homologue, C. rotundicauda Relish (CrRelish). Similar to Drosophila Relish and the mammalian p100, CrRelish contains both the Rel-homology domains (RHD) and the IκB-like domain. In this study, we found that the RHD of CrRelish can recognize horseshoe crab and human κB response elements and activate the downstream reporter in vitro, thereby suggesting the evolutionary conservation of this molecule. Pseudomonas aeruginosa infection transcriptionally upregulates CrRelish, which exhibits a dynamic protein profile over the time course of infection. Surprisingly, secondary infection reinduced an upsurge in CrRelish protein expression to a level which overrode the protein degradation at 12 h postinfection. These observations strongly suggest the involvement of CrRelish in antibacterial defense. Secondary infection causes (i) the maintenance of a favorable expression-competent sequence context of the CrRelish gene and/or (ii) the derepression or stabilization of the CrRelish transcript resulting from the primary infection to enable the more rapid expression and accumulation of the CrRelish protein, reflecting apparent signal/immune priming in a repeated infection.

The NF-κB signaling pathway plays a major role in inflammation and innate immunity by regulating the expression of numerous genes involved (4, 6). Up to now, two pathways for NF-κB activation have been described, the classical pathway and the alternative pathway (2). In the classical pathway of NF-κB activation, NF-κB proteins, which are normally sequestered in the cytoplasm by inhibitor κB (IκB), are freed upon IκB kinase β (IKKβ)-dependent IκB degradation (6). The degradation of IκB unmasks the nuclear localization signal (NLS) of the NF-κB protein, leading to its nuclear translocation. In the nucleus, the NF-κB transcription factor binds to the promoters of various genes with the consensus κB sequence, 5′-GGGRNNYYCC-3′, and upregulates the expression of these target genes (2). On the other hand, in the alternative pathway of NF-κB activation, the phosphorylation and processing of NF-κB p100 to p52 will allow its dimerization and nuclear translocation (17). The activation mechanism involves the action of NF-κB-inducing kinase and the IKK1-dependent phosphorylation of p100 (3), which triggers its proteasomal processing. This signaling pathway is activated in response to developmental signals transduced by lymphotoxin β receptor and RANK, which are required for lymph node genesis (17), or by BAFFR, CD40, and CD27, which regulate B-cell survival and proliferation (15, 24).

Recently, human p100 was demonstrated to be the fourth inhibitory IκB protein, which operates under pathological scenarios (1). This finding suggests possible cross talk between the classical and alternative NF-κB pathways in which the inflammatory and developmental signals converge. To further understand the functions of p100 in innate and adaptive immune responses, it would be crucial to define its evolutionary basis and functional significance in invertebrates which, presumably, harbor a more primitive immune system. However, other than evidence of NF-κB homologues in several dipteran insects, little is known about the NF-κB signaling pathway in other invertebrates (8). Furthermore, in Caenorhabditis elegans, the NF-κB transcription factor is absent and similar functional homologues (Traf, Pelle, and Cactus) found in C. elegans are not involved in the innate immune response (14). These observations are suggestive that the canonical NF-κB signaling pathway is not functional in the immune system of C. elegans (9).

The horseshoe crab, which has survived unchanged for about 550 million years, is the most ancient arthropod. It harbors a rich repertoire of innate immune molecules acting in the front line of defense and has been demonstrated to be a useful invertebrate model for dissecting innate immune mechanisms (12, 22, 25, 26). Our recent discovery and functional characterization of classical NF-κB and IκB in the horseshoe crab provides us with valuable information on the molecular evolution of NF-κB-mediated cell signaling pathways during host-pathogen interaction (23). The isolation of Carcinoscorpius rotundicauda Relish (CrRelish), which is a p100 homologue, from the horseshoe crab also prompted us to investigate the biological function of CrRelish in this “living fossil.” Mammalian p100 is involved in the development and maintenance of secondary lymphoid organs (2). On the other hand, the horseshoe crab is thus far reported to harbor only innate immunity. Therefore, the elucidation of the functional roles of this molecule upon bacterial challenge can greatly aid our understanding of the evolution of the alternative pathway of NF-κB signaling.

Herein, we demonstrated that the transactivation mechanism of this alternative NF-κB pathway has been evolutionarily entrenched since several hundred million years ago. CrRelish recognized the promoter sequence of horseshoe crab and human κB-responsive genes and activated their transcription. Challenge of horseshoe crab with Pseudomonas aeruginosa intimately regulated CrRelish expression in vivo, suggesting a tight regulation of the molecule in a highly dynamic manner. Furthermore, we showed that secondary infection caused a more rapid induction of the CrRelish protein. Therefore, CrRelish is likely a target of the acute phase of the immune signaling pathway(s).

MATERIALS AND METHODS

Bacterial challenge.

Horseshoe crabs (C. rotundicauda) were collected from Kranji Estuary, Singapore. The animals were handled according to national and institutional guidelines (National Advisory Committee for Laboratory Animal Research, Singapore). Gram-negative bacteria (P. aeruginosa [ATCC 27853]) were cultured overnight in tryptic soy broth (Difco) at 37°C. Bacteria were pelleted at 5,000 × g for 10 min at 4°C, washed, and resuspended in 0.9% saline to a density of 1 × 10⁷/ml. The horseshoe crabs were injected with 1.2 × 10⁷ CFU/kg of body weight (11). Hemocytes were collected at indicated times postinfection (0 to 72 h) for further analysis. Control experiments included mock infection with injections of equal volumes of saline at the respective time points.

Construction of expression vectors.

The insect expression vector pAc5.1/V5-HisA and the mammalian expression vector pcDNA3.1/V5/HisA (Invitrogen) were used for fusion expression of full-length and truncated CrRelish proteins (CrRelish, the Rel-homology domain [RHD] with the NLS and the linker between the NLS and ankyrin [ANK; RHD+NLS+L], RHD+NLS, and the RHD) (Fig. 1C). The cloning sites chosen were ApaI and KpnI. All fusion proteins are tagged with His and V5. Except for RHD, all other constructs contained the NLS.

FIG. 1.

Sequence analysis of CrRelish and schematic representation of expression constructs. (A) Schematic representation of full-length CrRelish and sequence alignment with the characteristic motifs of other NF-κB proteins. Full-length CrRelish contains an N-terminal RHD (gray box) with a conserved DNA-binding motif, followed by an NLS (black box) and a C-terminal IκB-like domain which contains six ANK repeats (white boxes). The amino acid sequences of NLSs from different organisms were aligned, and the first and last amino acids of the DNA-binding motif and the NLS are indicated. (B) Amino acid identity of RHD and ANK repeats between representative members of the p100 protein family. For a list of GenBank accession numbers of the sequences used, see Table S2 in the supplemental material. (C) Schematic representation of the expression vectors used in transfection experiments: full-length CrRelish, RHD+NLS+L (RHD with the NLS and the linker between the NLS and ANK), RHD+NLS (RHD with the NLS), and RHD (RHD only).

Cell culture, transfection, and luciferase assays.

Plasmids were isolated using the EndoFree plasmid maxi kit (Qiagen) according to the manufacturer's instruction. Drosophila Schneider S2 cells were maintained at 25°C in Drosophila serum-free medium (Invitrogen, Carlsbad, CA) supplemented with 20 mM l-glutamine and 5% fetal bovine serum (Invitrogen). Twelve hours prior to transfection, the cells were seeded in a six-well plate at 1.2 × 10⁶ cells per well. Transfections were conducted using Cellfectin (Invitrogen) according to the manufacturer's recommendation.

For expression in mammalian cells, HEK 293 cells were routinely grown in complete Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37°C in a humidified atmosphere of 5% CO₂ and 95% air. Twelve hours prior to transfection, the HEK 293 cells were plated in six-well tissue culture plates (Nunc) at a cell density of 5 × 10⁵ cells per well. The transfection was conducted using Lipofectamine 2000 (Invitrogen). Cells were harvested for examination of protein expression and dual luciferase assays by using the dual luciferase reporter assay system (Promega). Cells were cotransfected with NF-κB-luciferase (Luc; Stratagene, La Jolla, CA) and pRL-cytomegalovirus (Promega) to obtain normalized reporter activity.

EMSA.

The electrophoretic mobility shift assay (EMSA) was performed in the presence of 1 × 10⁵ cpm/pmol ³²P-labeled oligonucleotide and 2 μg of poly(dI-dC) at 25°C for 30 min in binding buffer (50 mM NaCl, 2 mM MgCl₂, 2 mM dithiothreitol, 1 mM EDTA, 10% glycerol, and 10 mM HEPES, pH 7.8) before electrophoresis on a 6% native polyacrylamide gel electrophoresis gel (acrylamide-to-bisacrylamide ratio of 79:1). In competition assays, 100× cold probes were used in each reaction mixture incubated for 30 min before the addition of the [³²P]DNA probe. The probes were labeled by annealing the upper and lower strands of the oligonucleotides with γ-³²P (GE Healthcare) by using T4 polynucleotide kinase (NEB). For the sequences of the oligonucleotide probes, see Table S1 in the supplemental material.

RT-PCR.

After we infected horseshoe crab with P. aeruginosa (11), the hemocytes were collected at indicated time points. Total RNA was prepared using TRIzol (Invitrogen). Reverse transcription-PCR (RT-PCR) was performed by using the Invitrogen RT-PCR kit with 3 μg of total RNA and oligo(dT). Semiquantitative RT-PCR was performed with rapid heating to 95°C for 3 min, followed by 19 to 25 cycles of 56°C for 30 s, 72°C for 1 min, and 95°C for 30 s. The primers used for RT-PCR were Cr-RelF (5′-CTCTAGCCCACCATTACATC-3′) and Cr-RelR (5′-CTCCGGAACAAACAAACTCTCTCCCTGA-3′).

Immunofluorescence localization of CrRelish.

Immunocytochemistry was performed with transiently transfected HEK 293 cells. The cells were plated onto 12-mm-diameter glass coverslips (Heinz) at a density of 2.5 × 10⁴ cells per coverslip and allowed to attach overnight at 37°C. At 24 h posttransfection, the growth medium was completely removed and cells were rinsed with phosphate-buffered saline (PBS), fixed for 15 min in 4% paraformaldehyde, and permeabilized by 0.1% Triton X-100-PBS for 2 min at room temperature. They were then incubated with mouse monoclonal antibody against V5 (1:500 dilutions; Invitrogen) overnight at room temperature. After washing with PBS, the cells were incubated with Alexa 594 rabbit anti-mouse immunoglobulin G antibody (1:500 dilutions; Molecular Probes) for 2 h at room temperature. These cells were counterstained with DAPI (4′,6′-diamidino-2-phenylindole) (1 μg/milliliter; Molecular Probes-Invitrogen) for 1 min and rinsed three times with PBS prior to fluorescence microscopy (Olympus BX60).

Western blot analysis.

Anti-CrRelish RHD and Anti-CrRelish ANK antibodies were raised in rabbits against keyhole limpet hemocyanin-conjugated peptides (ITEADKNQLKKDAE for the CrRelish RHD and IVTSPKAKSQDLKS for CrRelish ANK) by BioGenes (Berlin). The antibodies were affinity purified using specific peptide as a ligand. CrRelish RHD antibody was tested for specificity by Western blot analysis using a recombinant CrRelish RHD. Densitometric scanning of the immunodetected bands was performed using a GS-800 densitometer (Bio-Rad). The intensity of the bands was quantified using Quantity One software.

Nucleotide sequence accession number.

Nucleotide sequence data reported are available in GenBank databases under the accession number CrRelish (DQ34578).

RESULTS

Sequence analysis of horseshoe crab p100 homologue CrRelish.

CrRelish was cloned from hemocytes, which are the major immune-responsive cells in the horseshoe crab. By amino acid sequence comparison, CrRelish shows high homology to Drosophila Relish and human p100 (Fig. 1A; see Fig. S1 in the supplemental material). The full-length CrRelish contains an N-terminal RHD with a conserved DNA-binding motif, followed by the NLS and a C-terminal IκB-like domain. Like Drosophila Relish, CrRelish also contains six ANK repeats in its C-terminal IκB-like domain. The sequence identity between the CrRelish RHD and that of other homologues ranged from 40 to 43%, while the sequence identity of ANK repeats between these proteins ranged from 31 to 38% (Fig. 1B). The analysis clearly shows that CrRelish shares a common ancestor with Relish and p100 proteins, which is consistent with their high sequence similarity. The complete alignment of CrRelish with human p100 and Drosophila melanogaster (DmRelish) showed that like the human p100, p105, and insect Relish, CrRelish is a mosaic protein which contains serine-rich regions, the DNA-binding motif, the RHD, the NLS, ANK repeats, and a death domain (see Fig. S1 in the supplemental material). A close examination of the CrRelish sequence suggests that it lacks the potential caspase cleavage site L-E-x-D (20) in the linker region. In DmRelish, Asp in the four-amino-acid motif is critical for recognition by caspase, which cleaves after the Asp residue (see Fig. S1 in the supplemental material). In CrRelish, Asp is mutated to Asn (in L-E-L-N at positions 632 to 635). This mutation would have abolished the potential recognition by caspase.

The RHD of CrRelish binds the κB response element.

To examine the functional roles of these motifs, several expression constructs were designed to express truncated CrRelish in Drosophila S2 cells (Fig. 1C). The structural motifs in CrRelish were computationally predicted. The truncated constructs and full-length CrRelish were tagged with V5. To investigate whether CrRelish can bind to the κB motif, the cell extracts were prepared from S2 cells transfected with the CrRelish RHD plasmid and used in the EMSA. The C. rotundicauda factor C (CrFC) promoter κB response element (positions −143 to −133), previously shown to be recognized by human NF-κB and the Drosophila dorsal protein (21), was used as the DNA probe.

The EMSA results showed that the CrRelish RHD can interact with the κB response element in the CrFC promoter (Fig. 2A). Overexpression of the CrRelish RHD resulted in an intensive binding between the RHD and the DNA probe (Fig. 2A, lane 2). The integrity of the κB sequence in the DNA probe is critical to the binding since mutation of the 5′ end of the κB motif (positions −143 to −141), from GGG to ATT, abolished the binding of the overexpressed RHD (Fig. 2A, lane 4). The interaction is specific as the presence of a 100× excess of the cold probe diminished RHD binding to the labeled probe (Fig. 2A, lanes 7 and 8). The fusion V5 tag in CrRelish RHD enabled a supershift assay using an anti-V5 antibody. Figure 2A, lane 6, shows a supershift for RHD, further suggesting that the complexes were formed by the RHD of CrRelish. The supershift was also shown to be specific as the anti-V5 antibody alone did not cause any binding (lane 5).

FIG. 2.

DNA binding activity of CrRelish RHD with CrFC κB probe and human consensus NF-κB probe. (A) Binding of CrRelish RHD protein to CrFC κB motif. Extracts of cells transfected with pAc5.1 vector (control) and pAc5.1 RHD were used in this assay. The CrRelish RHD-κB complex is indicated by a filled arrowhead (lane 2). When the mutant probe was used, no binding was observed for RHD (lane 4). Anti-V5 antibody caused a supershift (lane 6), which indicates specificity for the overexpressed recombinant protein (open arrowhead). A 100× excess of cold probe competed for the binding (lanes 7 and 8). The unspecific binding is indicated by a black arrow. (B) EMSA results for CrRelish RHD with human κB consensus DNA probe. The empty pAc5.1 vector was used as the control (lane 1). The binding of the human κB consensus probe with CrRelish RHD is comparable to that of CrFC probe, as indicated by the filled arrowhead. The supershift produced by anti-V5 antibody (indicated by the open arrowhead) shows the specificity of the binding. Human p65 was used as a positive control (black arrow). For the sequences of all probes used in this assay, see Table S1 in the supplemental material. −, absence of; +, presence of.

As shown in Fig. 2B, the CrRelish RHD also recognizes the consensus human NF-κB binding motif, 5′-AGTTGAGGGGACTTTCCCAGGC-3′. Therefore, CrRelish may serve as a functional substitute for vertebrate p100 or vice versa. This shows cross-phylum recognition of the κB element by CrRelish and its homologues and indicates the ancient origin and functional conservation of p100.

Transactivation activity of CrRelish.

The observed cross-phylum functionality of CrRelish made it logical and feasible to express CrRelish and its deletion mutants in human cell lines. This functionality also circumvents the low levels of CrRelish protein detected in transfected Drosophila S2 cells, despite the fact that both the horseshoe crab and Drosophila are arthropods. CrRelish and its mutants were all expressed using the TnT-coupled transcription/translation systems (see Fig. S2A in the supplemental material). However, in the transfected human HEK 293 cells, only the expressions of the RHD and RHD+NLS were immunodetectable (see Fig. S2B in the supplemental material). This result suggests that the stabilities of CrRelish and its mutants probably differ significantly inside HEK 293 cells.

To examine the transcriptional activities of CrRelish and its mutants, HEK 293 cells were cotransfected with two reporter vectors, κB-Luc (the firefly luciferase under the control of κB motifs) and pRL-CMV (which constitutively expresses Renilla luciferase under the cytomegalovirus promoter). The relative luciferase activity in the transfected HEK 293 cells was measured 24 h after transfection. The RHD of CrRelish showed significant and reproducible activation of the NF-κB reporter, while full-length CrRelish and its other mutants showed either low or no activation (see Fig. S3 in the supplemental material). The RHD induced a 12-fold increase in transcriptional activation, while RHD+NLS, being the second most potent mutant, caused only a threefold increase. The extent of activation seems to correlate with the level of CrRelish and mutant expression in the transfected cells (see Fig. S2B in the supplemental material). Furthermore, the reporter activity increased linearly with increasing doses of the CrRelish RHD plasmid (Fig. 3A). Taken together, the transfection studies strongly suggest that the CrRelish is functional.

FIG. 3.

Functional activation by RHD of CrRelish. (A) Dose-dependent activation of reporter by CrRelish RHD. HEK 293 cells were transiently cotransfected with κB-Luc (experimental reporter, as a control of the inducible expression of firefly luciferase) and pRL-CMV (control reporter that gives constitutive expression of Renilla luciferase), together with increasing doses of the CrRelish RHD. (B) Cotransfection of cells with the CrRelish RHD and human p65. The reporter activity elicited by cotransfection is about 66-fold compared with that elicited by the control. The reporter activity is approximately the sum of the values for the RHD and human p65 used individually. Thus, there was no synergy between the CrRelish RHD and human p65. All experiments were conducted in triplicate. Data are presented as means ± standard errors (error bars) of results from three independent experiments.

In order to assess the extent of activation produced by CrRelish in comparison with that by the human NF-κB, cells were cotransfected with p65 (RelA) and the same reporter plasmids. As shown in Fig. 3B, human p65 caused a 58-fold increase in luciferase expression compared to the 12-fold increase caused by the CrRelish RHD. As both the CrRelish RHD and p65 activate the same κB reporter, we investigated the possible synergy or antagonism between the CrRelish RHD and p65. Our rationale for investigating the possible synergy or antagonism between CrRelish and p65 was based on the possibility that CrRelish “requires” a protein partner to dimerize with and, thence, becomes active in DNA binding. Thus, we chose p65 as a potential partner since many NF-κB dimers are reported to contain human p65 (10). Figure 3B shows that while cotransfection with the CrRelish RHD and human p65 increased the reporter activity to 66-fold, this transactivation is more likely to be additive rather than synergistic.

CrRelish RHD resides in the cytoplasm.

The cellular localization of the CrRelish RHD was investigated in transfected HEK 293 cells to help understand its transactivation mechanisms. The transfected cells were stained with an anti-V5 antibody 24 h posttransfection. The CrRelish RHD was detected predominantly in the cytoplasm (see Fig. S4A). However, under the same conditions, a substantial fraction of p65 was found in the nucleus. The abundance of p65 in the nuclei of transfected HEK 293 cells compared with the low level of the CrRelish RHD might explain its higher reporter activity.

The subcellular localization of the CrRelish RHD was further quantified by Western blot analysis. Cytoplasmic and nuclear extracts were prepared from HEK 293 cells transfected with the RHD and analyzed using the anti-V5 antibody (see Fig. S4B in the supplemental material). Consistent with the results obtained by immunostaining, the CrRelish RHD was abundantly localized in the cytoplasm, with a low level in the nucleus. This small amount of the CrRelish RHD in the nucleus was probably sufficient to cause the observed reporter activity.

Biological significance of infection-dependent upregulation of CrRelish.

While the in vitro studies demonstrated CrRelish RHD binding to the κB promoter (Fig. 2A), leading to reporter expression (Fig. 3A), the physiological relevance of these properties remained unclear. The observation that in vitro, full-length CrRelish was not detected in transfected S2 or HEK 293 cells makes this question particularly acute (see Fig. S2B in the supplemental material). In an attempt to gain insight into this intrigue, we examined native CrRelish expression in hemocytes. To follow the fate of specific CrRelish domains, we raised antibody against the RHD (α-RHD) of CrRelish. Horseshoe crabs were injected with P. aeruginosa, and hemocytes were harvested at various time points and subjected to Western blot analysis. Using the α-RHD antibody, we observed a time-dependent regulation of CrRelish expression in hemocytes upon bacterial challenge (Fig. 4A). The size of the protein band corresponds to that of full-length CrRelish (∼130 kDa). Infection induced the CrRelish protein, which first became detectable 1 h postinfection (hpi) and peaked at 3 hpi. This result was followed by a sharp decline at 6 hpi, and the protein returned to near the basal level from 24 hpi onwards. The drastic loss of CrRelish protein in the hemocytes between 3 and 6 hpi indicates its rapid downregulation. This downregulation suggests an internal postinfection autoregulatory mechanism for CrRelish expression. The detailed mechanism underlying this autoregulation warrants further investigation.

FIG. 4.

Regulation of CrRelish expression by bacterial challenge. (A) Western blot analysis of CrRelish protein expression after bacterial infection. The antibody against the RHD was used to probe the protein from naïve and P. aeruginosa-challenged horseshoe crab hemocyte lysates. The naïve sample served as a control for the normal endogenous level of CrRelish protein. Horseshoe crab actin was a loading control. (B) RT-PCR for examination of CrRelish mRNA expression. The hemocytes were collected at the indicated time points up to 72 hpi. Horseshoe crab actin mRNA served as a loading control. (C) Quantitation of Western blot analysis and RT-PCR bands by densitometric scanning. The relative intensities were scaled for comparison across different time points.

To determine the expression pattern of CrRelish upon P. aeruginosa infection at the transcription level, RT-PCR was performed with total RNA from hemocytes collected at the indicated time points after challenge. No CrRelish mRNA was detected in hemocytes before infection. Over an infection period of 1 to 72 h with P. aeruginosa, the expression of CrRelish mRNA was induced within 0.5 hpi and the level increased gradually (Fig. 4B) and was maintained throughout the 72 hpi. This result is in contrast to the protein expression profile, which was restricted to the early phase (1 to 6 hpi) of infection (Fig. 4C). This phenomenon suggests multiple mechanisms of regulation of the CrRelish expression during an infection.

Since CrRelish protein was hardly detectable between 12 and 24 hpi, we investigated the effect of a secondary infection on its expression over the same period. The secondary challenge was carried out at 12 h after the primary infection in order to avoid overlapping with the protein expressed after the primary challenge. When horseshoe crabs were rechallenged with an equal dose of P. aeruginosa at 12 h after the first infection, it was striking to observe that CrRelish expression was provoked effectively and more rapidly, with the protein being detectable within 0.5 h following secondary infection. In a manner similar to that for the protein profile of primary infection, the secondary infection also decreased the CrRelish protein sharply at 6 hpi (Fig. 5A). Additionally, through investigating the mRNA expression over the primary and secondary infections, we observed a positive correlation between the bacterial challenge and CrRelish mRNA expression (Fig. 5B). Quantifying both the mRNA and protein profiles showed that while the mRNA remained highly induced, the secondary infection advanced the appearance of the protein to 0.5 hpi, bringing about a second cycle as in the primary infection (Fig. 5C). Repeated experiments produced consistent and reproducible expression profiles of CrRelish mRNA and protein induced by the primary and secondary infections. This result is consistent and logical in view of a repeat infection where the demand on the NF-κB pathway would be extremely critical. Our results strongly suggest the involvement of CrRelish in antibacterial defense, and CrRelish expression is regulated at both the transcriptional and translational levels and perhaps even posttranslationally by the dynamics of the protein loss.

FIG. 5.

Regulation of CrRelish expression by secondary bacterial challenge. (A) Western blot analysis of CrRelish protein expression after first bacterial infection and secondary bacterial infection. Twelve hours after the first infection, the horseshoe crabs were rechallenged with the same dose of P. aeruginosa used in the primary infection and the hemocyte lysates were collected at the indicated time points up to 12 hpi after both the primary and secondary infections. Horseshoe crab actin served as a loading control. (B) RT-PCR for examination of CrRelish mRNA expression at time points after primary and secondary infections was as indicated for the Western blot. Horseshoe crab actin mRNA was the loading control. (C) Quantitation of Western blot analysis and RT-PCR bands by densitometric scanning. The relative intensities for primary and secondary infections were presented separately.

DISCUSSION

Since the discovery of NF-κB transcription factors (16), a wealth of information on the mechanisms that operate in the NF-κB signaling pathway and the roles of NF-κB in various diseases has been generated for the mammalian systems (7). However, other than evidence of Toll homologues in several insect species, an IKKβ homologue in an oyster, and NF-κB homologues in several dipteran insects (8), little is known about the NF-κB signaling pathway in invertebrates. The discovery of CrRelish, which is an ancient NF-κB p100 homologue in the horseshoe crab, pushes the limits of its evolutionary origin to several hundred million years, implying the conservation of such an important alternative NF-κB signaling molecule.

In this study, we found that despite the huge evolutionary distance between the horseshoe crab and the vertebrates, CrRelish displays signature motifs similar to those found in the vertebrate p100 homologues, notably the DNA binding motif, the NLS, and inhibitory ANK repeats. Our in vitro investigation showed that the RHD of CrRelish can bind both the horseshoe crab and the human κB response elements (Fig. 1B), indicating the functional conservation of the CrRelish across the phylum. Surprisingly, the RHD alone is sufficient to transactivate the κB-reporter in HEK 293 cells (see Fig. S3 in the supplemental material) in a dose-dependent manner (Fig. 3A), even though the RHD construct was devoid of an NLS. Immunostaining and Western blot analysis also showed that there was a small amount of CrRelish RHD protein present in the nucleus (see Fig. S4A and B in the supplemental material). A plausible explanation for the entry of a small amount of the RHD into the nucleus is its overexpression in these cells (see Fig. S2B in the supplemental material).

Like Drosophila and mosquito Aedes aegypti Relish (5, 18), the expression of CrRelish protein was induced significantly during the course of infection and it reached a maximum level at 3 hpi (Fig. 4A). The short induction time suggests the involvement of this protein in the acute phase of the immune response. Although the mRNA level of CrRelish appeared from 0.5 h and increased steadily to 72 hpi (Fig. 4B), the protein expression decreased sharply from 6 hpi onwards (Fig. 4A). The cleavage of a Relish homologue was reported for Drosophila (19), and it is dependent on caspase (20). The mutation of the predicted cleavage site, L-E-x-D to L-E-L-N (positions 622 to 625 in the linker region of CrRelish), may explain the abolishment of the potential recognition and processing by caspase and, hence, the lack of cleavage products from the CrRelish.

Interestingly, in the horseshoe crab, the downregulation of CrRelish protein coincides with the rapid clearance of P. aeruginosa within 6 hpi (11). The observed positive correlation between CrRelish protein and the rate of removal of P. aeruginosa suggests the involvement of CrRelish in antibacterial defense. Furthermore, with the clearance of P. aeruginosa within 6 hpi (11), the protein expression of CrRelish decreased drastically, suggesting autoregulation of CrRelish to regain homeostasis. The protein loss occurred after the elimination of the injected dose of bacteria, suggesting the suppression of the infection signal soon after the acute phase of infection to prevent excessive inflammatory response in the host.

We also found that the secondary bacterial challenge induced a more rapid CrRelish protein expression, bringing forward the appearance/accumulation of the protein as early as 0.5 hpi (Fig. 5A). Between 12 and 24 h after primary infection, the rate of protein loss largely exceeded the rate of protein synthesis, resulting in low or no full-length CrRelish in the hemocytes. However, the secondary infection presents an extra load of bacteria which needs to be eliminated. The newly injected bacteria increase the protein synthesis rate of CrRelish, possibly through transcriptional upregulation (Fig. 5B). Therefore, the synthesis rate now exceeds the degradation rate and the full-length protein, which had previously subsided at 12 hpi, is again detectable within 0.5 h after secondary infection. Priming Drosophila with a sublethal dose of Streptococcus pneumoniae has been reported to protect against an otherwise-lethal second challenge of S. pneumoniae (13). We note that similar to this effect, a more rapid induction of CrRelish protein in the secondary challenge is probably attributable to immune priming.

In summary, we report the discovery of a primitive but functional NF-κB p100 homologue, CrRelish, in the immune defense of a “living fossil,” the horseshoe crab C. rotundicauda. CrRelish shares numerous signature motifs with vertebrate class I NF-κB and Drosophila Relish, and its expression is highly dynamic and rapidly responsive to P. aeruginosa infection, suggesting the conservation of function from horseshoe crab to human.