Introduction

Pathogenic microorganisms including bacteria and virus are freely available in the environment and can cause acute infections in host organisms. Therefore, natural defense mechanisms have been established to prevent the deleterious effects of these pathogenic encounters. Innate immunity acts as a basic defense mechanism against the majority of microbial infections. This first-line defense system is one arm of the immune system in fish, whereas it is the exclusive immune system in invertebrates. Teleost fish possess innate immune parameters that are more active and diverse than those of their mammalian counterparts (Magnadottir 2006).

Pattern recognition receptors (PRRs) play a crucial role in innate immunity by contributing to the identification of different pathogen-associated molecular patterns (PAMPs) (Medzhitov 2007) and damage-associated molecular patterns (DAMPs) (Matznger 2002). Toll-like receptors (TLRs) are one of the predominant type of evolutionarily conserved PRRs that recognize microbial PAMPs (Rebl et al. 2010) such as lipopolysaccharides (LPS), lipoproteins, lipoteichoic acid, flagellin, and nucleic acids (Akira et al. 2006). TLRs are composed of a characteristic extracellular domain, which harbors up to 26 leucine-rich repeats (LRRs) flanked by N- and C-terminal cap motifs (Bell et al. 2003); a transmembrane domain; and an intracellular C-terminus, which includes a Toll/Interleukin (IL)-1 receptor domain (TIR) (Palti 2011).TLRs are diverse among fish species. In this regard, teleost fish show remarkable diversity, rendering them as the most diverse group of vertebrates which bear TLRs (Venkatesh 2003; Volff 2005). The rainbow trout was the first fish in which the IL/TLR super family was identified (Sangrador-Vegas et al. 2000). Typically, fish lineages lack TLR6 and TLR10, whereas TLR20, TLR21, TLR22, and TLR23 are virtually exclusive to fish species (Rebl et al. 2010; Takano et al. 2011; Priyathilaka et al. 2014).

TLR1 and TLR2 belong to the TLR1 family (Magnadottir 2006) and form heterodimers that detect lipoproteins on Gram-positive bacteria (Hajjar et al. 2001) and LPS on Gram-negative bacteria (Takeuchi et al. 2002).

Rock bream is currently a very popular sashimi in East Asia, especially in Korea and Japan, rendering its increasing economic value. Rock bream inhabit the coastal areas of the Pacific and Indian Oceans. In the recent years, the prevalence and virulence of bacterial and viral pathogenic infections have negatively affected this teleost fish species, enhancing its mortality and resulting in its shortage (Zenke and Kim 2008; Park 2009). Therefore, it is necessary to elucidate disease resistance mechanisms in rock bream to develop effective preventive and therapeutic strategies. Moreover, characterization of TLRs in fish like rock bream is a basic approach in comparative studies on fish and mammalian innate immune systems.

Investigation of TLRs in bony fish is a crucial undertaking. Because fish species serve as models of physiology, immune mechanisms identified in fish can be applied to other organisms as well. However, studies characterizing the complete genomic sequence of TLR1 in bony fish are limited. In this study we identified and characterized the complete rock bream TLR1 (RbTLR1) gene at the genomic level. To understand the relationship between immune stimulants and TLR1 mediated host defense in rock bream, we examined the regulation of RbTLR1 mRNA in the spleens of fish, stimulated with LPS, Edwardsiella tarda, Streptococcus iniae, rock bream iridovirus (RBIV), and polyinosinic:polycytidylic acid (poly I:C).

Materials and methods

Identification of partial cDNA sequences of RbTLR1

A rock bream sequence database was established using the Roche 454 genome sequencer FLX system (GS-FLX™), a next-generation DNA sequencing technology (DNA Link, Republic of Korea). Using the Basic Local Alignment Search Tool (BLAST) algorithm (http://www.ncbi.nlm.nih.gov/BLAST), we identified partial-length TLR1 cDNAs in the rock bream sequence database.

Identification of the complete RbTLR1 genomic sequence

A random-shear bacterial artificial chromosome (BAC) library of rock bream genomic DNA was constructed (Lucigen, USA). We screened this library to identify the BAC clone bearing RbTLR1 gene using a two-step polymerase chain reaction (PCR) based genomic library screening approach with gene specific-primers (RbTLR1-F and RbTLR1-R; Table 1). The putative RbTLR1 containing clone was then sequenced using the GS-FLX™ system (Macrogen, Korea) to obtain full-length genomic sequence. The open reading frame (ORF) of RbTLR1 was then identified from the obtained genomic DNA sequence by using a National Center for Biotechnology Information (NCBI)-BLAST algorithm (http://www.ncbi.nlm.nih.gov/BLAST). Annotated sequence information for RbTLR1 was deposited in the NCBI GenBank database (Accession Number JQ754308).

Table 1 Primers used in this study

In silico analysis of the rock bream TLR1 sequence

RbTLR1 orthologs were compared using NCBI-BLAST. Pairwise sequence alignments (http://www.ebi.ac.uk/Tools/psa/emboss_needle) and multiple sequence alignments (http://www.ebi.ac.uk/Tools/msa/clustalw2) were performed using the ClustalW2 program. The phylogenetic relationship of RbTLR1 was determined by the neighbor-joining method, supported by 1000 bootstrapping replicates, using Molecular Evolutionary Genetics Analysis (MEGA) software version 3 (Kumar et al. 2004). Protein domain predictions were carried out using the ExPASy-prosite database (http://prosite.expasy.org), SMART online server (http://smart.embl-heidelberg.de), and MotifScan scanning algorithm (http://myhits.isb-sib.ch/cgi-bin/motif_scan). The signal peptide was predicted using the signalP server (http://www.cbs.dtu.dk/service/signalP) and the physicochemical properties of the deduced protein were predicted using the ExPASy prot-param tool (http://web.expasy.org/protparam). The tertiary structure of RbTLR1 was modeled based on the homology modeling strategy using SWISS-MODEL online server (Schwede et al. 2003; Arnold et al. 2006). The three-dimensional (3D) image was generated using the RasMol software version 2.5.7.2 (Goodsell 2005).

Experimental fish and tissue collection

Healthy rock breams with an average body weight of 50 g were obtained from the Jeju Special Self-Governing Province Ocean and Fisheries Research Institute (Jeju, Republic of Korea). The fish were maintained in a controlled environment (salinity 34 ± 1 ‰; pH 7.6 ± 0.5) at 22–24 °C. Fish were acclimated for 2 weeks prior to the experimentation. Whole blood (1 mL per fish) was collected from the caudal fin by using a sterilized syringe. The samples were immediately centrifuged at 3000×g for 10 min at 4 °C to separate the blood cells from the plasma. The collected cells were snap-frozen in liquid nitrogen. The fish were sacrificed and the gill, liver, skin, spleen, head kidney, kidney, skin, muscle, and brain were excised, immediately snap-frozen in liquid nitrogen, and stored at −80 °C.

Immune challenge experiment

To determine the immune responses of RbTLR1, pathogenic bacteria E. tarda and S. iniae, RBIV, poly I:C, and LPS were used as immune-stimulants in time-course experiments. Each rock bream was administered a single intraperitoneal (i.p.) injection of 100 μL LPS (1.25 μg/μL, Escherichia coli 055:B5; Sigma) or poly I:C (1.5 μg/μL; Sigma) suspended in phosphate-buffered saline (PBS). E. tarda and S. iniae used for the bacterial-challenge experiments were obtained from the Department of Aqualife Medicine, Chonnam National University, Korea. The bacteria were incubated at 25 °C for 12 h in brain–heart infusion broth (Eiken Chemical Co., Japan) supplemented with 1 % sodium chloride. The cultures were resuspended in sterile PBS and then diluted to the desired concentration. Each rock bream was i.p.-injected with 100 μL live E. tarda (5 × 103 CFU/μL) or S. iniae (1 × 105 CFU/μL) in PBS. For the virus-challenge experiment, kidney tissue specimens obtained from RBIV-infected moribund rock bream were homogenized in 20 volumes of PBS. Tissue homogenates were centrifuged at 3000×g for 10 min at 4 °C, and the supernatants were filtered through a 0.45-μm membrane. Each rock bream was infected using a single i.p. injection of 100 μL RBIV in PBS. A control group was injected with an equal volume (100 μL) of PBS. Rock bream spleen samples were collected at 3, 6, 12, 24, and 48 h post injection (p.i.) from LPS-, poly I:C-, E. tarda-, S. iniae- and RBIV-infected rock breams. PBS-injected samples (injection control) were also isolated at each time point. A group of three uninjected animals served as a negative control. Tissue samples from three rock breams were obtained at each time point and pooled. Total RNA was extracted from the pooled tissue samples, and cDNA was synthesized.

Total RNA extraction and cDNA synthesis

Using TRI Reagent™ (Sigma), total RNA was extracted from the blood, gill, liver, spleen, head kidney, kidney, skin, muscle, and brain of healthy rock breams as well as from spleen tissues of immune-challenged fish. cDNA was synthesized from total RNA samples as described previously (Whang et al. 2011).

RbTLR1 mRNA expression analysis by using qPCR

Quantitative real time polymerase chain reaction (qPCR) was used to measure the expression levels of RbTLR1 in the aforementioned tissues under ‘Experimental fish and tissue collection’. The temporal expression of RbTLR1 in the spleens of fish challenged by different PAMPs and microorganisms (Section ‘Immune challenge experiment’) was also examined using qPCR. qPCR was performed in a Dice™ thermocycler (Real-Time System TP800; TaKaRa, Japan). Reactions (20 µL total volume) contained 4 µL of diluted cDNA, 10 µL 2× TaKaRa Ex Taq™, SYBR premix, 0.5 µL of each primer (RbTLR1-F and RbTLR1-R; Table 1), and 5 µl of double-distilled H2O as per the essential MIQE guidelines (Bustin et al. 2009). qPCR was performed under the following conditions: 95 °C for 10 s; 35 cycles of 95 °C for 5 s, 58 °C for 10 s, and 72 °C for 20 s; and a final cycle of 95 °C for 15 s, 60 °C for 30 s, and 95 °C for 15 s. The baseline was set automatically by using the Dice™ Real-Time System software (version 2.00). RbTLR1 expression was determined using the Livak (2−ΔΔCT) method (Livak and Schmittgen 2001). The same qPCR cycle parameters were used for the internal reference gene, rock bream β-actin (GenBank ID: FJ975146). At least three replicates were performed for each sample. RbTLR1 mRNA levels, relative to the rock bream β-actin expression, are reported. In the immune challenge experiments, RbTLR1 expression levels were normalized to the corresponding PBS-injected controls at each time point. The expression level of the uninjected control at 0 h was set as the basal level and was compared with the relative mRNA expression of the injected groups at each time point. All the data are presented as mean ± standard deviation (SD). To determine statistical significance (P < 0.05), a two-tailed unpaired Student’s t- test was performed.

Results

Molecular characterization of RbTLR1

The complete ORF derived from the RbTLR21 gene comprised a 2406 bp sequence encoding 802 amino acids (GenBank ID: JQ754308). The predicted molecular mass of RbTLR1 was approximately 90.7 kDa, and the theoretical isoelectric point was 6.39.

Sequence analysis of RbTLR1 revealed that it possessed typical TLR domain organization. RbTLR1 contained eight LRRs (residues 44–66, 91–110, 375–396, 401–426, 447–466, 470–489, and 516–535), a C-terminal LRR (residues 527–587), transmembrane domain (residues 588–610), and TIR domain (residues 646–791) (Fig. 1). Using the SignalP server, we showed that RbTLR1 possess a signal peptide (residues 1–24, Fig. 1).

Fig. 1
figure 1figure 1

Multiple sequence alignment of vertebrate TLR1 orthologs. The sequence alignment was generated using the ClustalW method. Signal peptide sequence and LRR motifs are indicated by the underline and gray arrows, respectively. The C-terminal LRR is denoted by the pattern-filled arrow. The transmembrane and Toll/Interleukin (IL)1 receptor (TIR) domains are denoted by unfilled double-headed and solid black arrows, respectively. Three moderately conserved domains within the TIR are indicated by boxes, in which well conserved sequence identified in third domain was shaded in blue color and underlined. Identical residues are shaded in gray color

Pairwise sequence similarity and identity comparisons revealed the homology of RbTLR1 with its vertebrate counterparts (Table 2). RbTLR1 shared 16.2–77.2 % identity with TLR1 proteins of other species. The highest and lowest identity was shared with the orange-spotted grouper and Caenorhabditis elegans, respectively. The TIR domain of RbTLR1 exhibited percent identities within the range of 26–95.2 %. The same species, orange-spotted grouper and C. elegans, had the highest and the lowest TIR domain identity, with RbTLR1 TIR domain, respectively. As expected, RbTLR1 shared its prominent sequence compatibility with its teleost homologs except with its channel catfish counterpart. Notably, the TIR domain shared greater identity and similarity values than the full-length RbTLR1 sequence. In accordance with the TLR1 multiple sequence alignment, three moderately conserved domains, previously identified in the green spotted pufferfish TIR region (Wu et al. 2008), were also identified in rock bream with minor substitutions, in which third domain harbors highly conserved sequence (FWANL) (Fig. 1).

Table 2 Percent similarity and identity (with gaps) of RbTLR1 and its TIR domain to TLR1 orthologs of other species

Phylogenetic analysis

To determine the evolutionary position of RbTLR1, a phylogenetic analysis was carried out using deduced amino acid sequences of several vertebrate and one invertebrate TLR1 proteins. C. elegans was used as an out-group. The tree we obtained showed three main clades of vertebrate TLR1 proteins: fish, mammals, and birds (Fig. 2). RbTLR1 was positioned in the fish clade, where it sub-grouped with orange-spotted grouper with maximum bootstrap support (100). The tree suggests that RbTLR1 evolved from a common ancestral vertebrate gene and showed substantial homology with its fish counterparts.

Fig. 2
figure 2

Phylogenetic tree constructed based on the ClustalW alignment of deduced amino acid sequences of various TLR1 proteins. Trees were estimated using the neighbor-joining method in MEGA version 4.0. Bootstrap values are shown at the nodes of the tree and major taxonomic clades are indicated. The corresponding NCBI-GenBank accession numbers of each homolog was indicated in Table 2

RbTLR1 genomic organization

The putative genomic RbTLR1 sequence lacked introns. RbTLR1 is encoded by a single exon which was 2931 bp in length. According to the comparison of RbTLR1 genomic DNA (gDNA) sequence with known TLR1 gDNA sequences of several other vertebrate species obtained from the NCBI-GenBank sequence database (http://www.ncbi.nlm.nih.gov/gene) (Fig. 3), both non-teleostan and teleostan TLR1counterparts have collectively portrayed a distinct genomic arrangements albeit RbTLR1 had a similar genomic organization to TLR1 from green spotted pufferfish (Wu et al. 2008) and bovine (GenBank ID: NC007304), including the lengths of their sequences (Fig. 3). Bearing a single exon along their whole genomic gene sequence, these two teleostan similitudes exhibited a clear demarcation from rest of the fish counterparts considered in the comparison. Besides these two TLR1 genes, others showed multi-exonic gDNA structure. However, from those, coding region of some of the teleostan similitudes including zebrafish, Nile tilpaia and zebra mubna TLR1s were found to be split into multiple exons. With the exception of bovine TLR1, orthologs in non-fish vertebrates also exhibited multi-exonic gDNA architecture. In general, the multi-exonic genes showed higher variability in total gene length due to the sizes of their introns.

Fig. 3
figure 3

Genomic organization of TLR1 genes in different vertebrate species. The exons and introns are indicated by boxes and solid lines, respectively. The exon sizes are indicated above the boxes, and intron sizes are indicated below the lines. Regions larger than 200 bp are abbreviated (diagonal lines). The sequences are written 5′ → 3′). TLR1 GenBank Gene-IDs or accession numbers are as follows: chicken, 426274; bovine, 574090; human, 7096; mouse, 21897; rainbow trout, GQ502184; zebrafish, 403127; pufferfish, EF095150; Zebra mbuna, 101475856; Amazon molly, 103135488; Nile tilapia, 102076024

Modeled tertiary structure of the RbTLR1 TIR domain

The 3D molecular organization of the RbTLR1 TIR domain was examined in a model generated using the human TLR1 TIR domain (Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSC-PDB – ID:1FYV) as the template. The model holds valuable information pertaining to RbTLR1, because TIR domains are conserved cytoplasmic components in TLRs that are important for initiating Toll signaling (Anderson 2000). The model resembled the typical TIR domain structure of TLR1, including a five-strand parallel β-sheet surrounded by five α-helices on both sides (Fig. 4). As in human TLR1 TIR (RCSB-PDB, 1FYV; Xu et al. 2000), each α-helix (αA–αE) and β-strand (βA–βE) was labeled in alphabetical order from the N- to C-terminus. Three main loop structures, the BB, CD, and DD loops, connected αB–βB, αC–βD, and αD–βD, respectively. Moreover, we identified four amino acid residues in our model that were present in structural positions corresponding to the conserved surface patch of human TLR1-TIR. Two residues (Asn and Gly) were located in the BB loop, whereas the other two (Phe and Glu) were localized near the beginning of the βA strand and the end of the αA helix, respectively. Conserved (F/Y)DA and FW motifs which are prominent features in human TLR1 TIR (Xu et al. 2000), were identified in our model close to the amino- and carboxy-terminal ends, respectively.

Fig. 4
figure 4

Predicted three-dimensional structure of the RbTLR1 TIR domain. Blue and green spheres represent the amino acids of the (F/Y)DA motif and surface patch, respectively. Amino acid residues of FW motif are represented by two white spheres. Carboxyl and amino terminals are denoted by the letters C and N, respectively. β-strands and α-helices are depicted in yellow and pink, respectively. Each strand and helix is labeled in alphabetical order, from the N- to C-terminus

Tissue-specific expression profile of RbTLR1

To evaluate the variable expression levels of RbTLR1 mRNA with respect to different tissues in rock bream, qPCR was performed for various tissues using gene-specific primers designed according to the RbTLR1 coding sequence (Table 1). The relative expression of each tissue was obtained using rock bream β-actin as the non-variant internal control. Relative tissue-specific expression was determined by normalizing values to the expression level in muscle (Fig. 5). RbTLR1 was expressed in all of the tissues analyzed. Further investigation showed pronounced RbTLR1 transcript levels in the spleen, kidney, and liver. Moderate levels were detected in the blood, heart and head kidney (P < 0.05). RbTLR1 transcript was less abundant in the brain, skin, and muscle tissues. Muscle tissue showed the lowest expression level (P < 0.05).

Fig. 5
figure 5

The tissue-specific expression of RbTLR1 mRNA by using quantitative polymerase chain reaction (qPCR). Error bars represent the SD (n = 3). Data labeled with different letters are significantly different (P < 0.05)

Transcriptional regulation of RbTLR1 in response to immune challenges

The temporal regulation of RbTLR1 mRNA expression was analyzed in the spleen tissue of rock bream exposed to LPS, poly I:C, E. tarda, S. iniae, or RBIV. LPS exposure significantly upregulated RbTLR1 mRNA levels (P < 0.05) in spleen tissue 12 and 24 h p.i. by approximately 1.5 and 2.5 fold, respectively (Fig. 6a). At 48 h p.i., transcript levels were downregulated. E. tarda challenge also induced the RbTLR1 transcript levels (P < 0.05) that peaked in the early phase (3 h p.i.) and persisted into the late phase (12 and 24 h) (Fig. 6a). In the spleen tissue of fish injected with S. iniae, mRNA levels were increased significantly (P < 0.05) at each time point tested (3–24 h p.i.), reaching a maximum 3.4-fold difference at 12 h p.i. relative to the unchallenged control (0 h) (Fig. 6a). In contrast to other immune challenges, spleen tissues from poly I:C-challenged fish showed a significant ~2.0-fold elevation in TLR1 mRNA at 6 h p.i only. Upon RBIV induction, TLR1 transcription was significantly (P < 0.05) upregulated at 3 and 12 h p.i, while significantly downregulated at 24 h p.i (Fig. 6b). After 48 h, transcript levels reached to its basal level.

Fig. 6
figure 6

Expression profile of RbTLR1 in spleen tissue upon stimulation with a LPS, Edwardsiella tarda, and Streptococcus iniae, and b polyinosinic:polycytidylic acid (poly I:C) and rock bream iridovirus (RBIV) determined using qPCR. The relative expression was calculated by the 2−ΔΔCT method using rock bream β-actin as a reference gene. Values were normalized to the corresponding PBS controls at each time point. Expression in uninjected controls at 0 h post injection was set as the baseline. Error bars represent SD (n = 3); *P < 0.05

Discussion

Innate immunity plays a crucial role in the first line host defense through identification of PAMPs of invading pathogenic organisms. TLRs are considered as fundamental PRRs involved in innate immunity. Here we characterized TLR1 homolog from rock bream, which is an economically important fish species that is vulnerable to diseases caused by pathogenic microorganisms.

We identified a novel TLR1 ortholog of rock bream (RbTLR1) from our BAC genomic library by using BLAST analysis. The deduced amino acid sequence shared numerous features with its other vertebrate orthologs. Resembling typical TLR architecture, RbTLR1 contained multiple extracellular LRR motifs, a single transmembrane domain, and an intracellular TIR domain (Fig. 1). Eight predicted LRR motifs were encoded by the RbTLR1 gene sequence. Previous evolutionary studies showed that TLR extracellular domains have evolved more rapidly than TIR domains (Johnson et al. 2003). This might account for the observed numerical variation of LRRs in extracellular domains of teleostan TLR1 similitudes. Pairwise sequence alignment showed that RbTLR1 and its TIR domain shared a high degree of identity and similarity to their respective vertebrate orthologs. The highest percent identity was with the full-length sequence and the TIR domain of the orange-spotted grouper (77.2 and 95.2 %, respectively). This observation validates the primary structure of the novel RbTLR1 protein we identified. In pairwise sequence comparisons, intracellular TIR domains exhibited greater conservation than the full-length sequence (Table 2), suggesting that there was greater structural similarity between TLR1 intracellular domains than the extracellular domains of different vertebrate species. Since intracellular domains of TLRs are responsible for initiating a signaling cascade and extra-cellular domains are involved in identifying compatible PAMPs, this observation further convinces the activation of similar TLR signaling pathways to different PAMPs (Slack et al. 2000; Bell et al. 2003). As expected, amino acid number (146) of TIR in RbTLR1 lies within the residue range (140–165), conserved among the vertebrate species (Table 2). As detected by our multiple sequence alignment study, among the third domain of three semi-conserved domains identified, ‘FWANL’ sequence was found to be well conserved in RbTLR1 (Fig. 1) complying with the characteristic FW motif of TLR family proteins, which is known to be important in cellular localization of the receptor molecule (Gangloff et al. 2005).

The phylogenetic tree generated using the TLR1 protein sequences from different vertebrates was in accordance with generally accepted phylogenetic relationships, separating the orthologs into three main clades (Fig. 2). Fish TLR1 members were grouped closely and independently, placing RbTLR1 with TLR1 of the orange-spotted grouper in the fish clade, confirming the common ancestral origin of these species.

Similar to the green spotted pufferfish, the RbTLR1 gene consists of a single exon (Fig. 3). Deviating from the typical genomic organization of mammalian TLR1, bovine TLR1 also harbors only a single-exon similar to RbTLR1. In contrast, all the other teleostan counterparts were found to be multi-exonic, possessing two to three exons. However, in the similitudes of zebrafish, Nile tilpaia and zebra mbuna coding regions were shared among multiple exons, suggesting the potential existence of spliced isoforms of those TLR1s. Moreover, we can further infer that the integration of introns in majority of the teleostan and non-teleostan TLR1s may have a function in substantial regulation of their expression, since some of the intronic sequences are known to harbor regulatory elements which involve in regulation of gene expression (Rose 2008). However, these suggestions merit further investigations.

TIR domains of TLRs are known to be important in the initiation of the intracellular signal transduction process through mediating protein–protein interactions (Kopp and Medzhitov 1999). Through binding different adaptor proteins such as myeloid differentiation primary response factor 88 (MyD88), TIRs play a crucial role in dictating which downstream signaling pathway is activated (Wesche et al. 1997; Medzhitov et al. 1998). Our model of the RbTLR1 TIR domain validated that we correctly identified the primary peptide sequence of RbTLR1. The model exhibited structural features that resembled the human TLR1 TIR domain (Fig. 4).

In our analysis of tissue-specific expression, the spleen exhibited the highest RbTLR1 transcript levels compared to the rest of the tissues analyzed (Fig. 5). We attributed this to the key role of the spleen in regulating the immune system (Tarantino et al. 2011). A similar tissue-specific expression profile is observed with TLR1 orthologs in other vertebrates including the orange-spotted grouper (Wei et al. 2011), rainbow trout (Palti et al. 2010), and chicken (Yilmaz et al. 2005). In contrast, TLR1 mRNA is scarcely expressed in every tissue examined in healthy green spotted pufferfish, including the spleen (Wu et al. 2008). TLR1 expression in torafugu is restricted to the kidney, heart and genital gland, with the strongest expression in kidney (Oshiumi et al. 2003). With the exception of torafugu and green spotted pufferfish, constitutive TLR1 expression was detected in the tissues examined from all the above mentioned species, suggesting its necessity for their survival.

To examine the regulation of RbTLR1 transcription levels in response to exposure to different pathogenic organisms and molecular patterns, we performed challenge experiments by using E. tarda, a Gram-negative bacteria; LPS, a component of Gram-negative bacterial cell walls; S. iniae, a Gram-positive bacteria; poly I:C, a PAMP that acts as dsRNA viral mimic; and RBIV, a virulent viral pathogen of rock bream.

As shown in Fig. 6a, RbTLR1 appears as a candidate immune responsive gene under Gram-negative bacterial invasion, because its expression was significantly elevated (P < 0.05) after stimulation with LPS and E. tarda. Notably, E. tarda could trigger an early phase (3 h p.i.) relatively higher (~2.9-fold) expressional induction compared to LPS, which may attribute to the difference of the strength of two stimulants. E. tarda is a live pathogen, whereas LPS is a non-living single chemical component of the bacterial wall. Hence, live pathogens like E. tarda can proliferate in host cells and increase their levels of PAMPs to mount more potent response. It is a documented fact that TLRs can lead to induce proinflammatory cytokines including IL-1β (Kawai and Akira 2010). Interestingly, IL-1β identified from rock bream was also found to upregulate at the same time points, 12 and 24 h after LPS injection and 3 h after E. tarda injection (unpublished data), reinforcing the potential activation of TLR1 mediated signaling cascade in rock bream spleen cells upon Gram negative bacterial invasion. Inductive response of RbTLR1 to the aforementioned stimulants is comparable to the modulation of TLR1 mRNA levels in the spleen of orange-spotted grouper (Oshiumi et al. 2003). Therein, increased expression levels of TLR1 were detected at 24 and 48 h after stimulation with Vibrio alginolyticus and 12 h after LPS induction. A similar positive regulation of TLR1 expression was observed in green spotted pufferfish after 12 h post challenge with LPS (Wu et al. 2008). As shown in Fig. 6a, the RbTLR1 transcription profile upon S. iniae challenge showed a nearly parabolic-shaped distribution, exhibiting a significant increase at every time point except 48 h p.i. These data suggest that RbTLR1 is a candidate PRR for sensing invaded Gram-positive bacteria, further convincing its role in activating immune signaling based on our detected upregulated expression of IL-1β in response to S. iniae exposure (unpublished data). Consistently, in zebrafish, Gram-positive Mycobacterium infection resulted an upregulation of TLR1 expression, even after 8 weeks of infection (Meijer et al. 2004). In contrast, bacterial TLR agonists such as lipoproteins and flagellins could not significantly upregulate the transcript level of TLR1 relative to the basal level in rainbow trout (Palti et al. 2010).

Since poly I:C is a well-known TLR agonist, it was used as an immune stimulant in our experiments to analyze the effect of immune challenge on RbTLR1 transcription. This viral PAMP enhanced RbTLR1 expression significantly (P < 0.05) only in the early phase (6 h p.i.). Similarly, poly I:C elevates the mRNA expression of TLR1 in the head kidney and spleen of the orange-spotted grouper, albeit in the late phase (24 h p.i.) (Wei et al. 2011). Intriguingly, our unpublished data showed a positive transcriptional response of IL-1β upon poly I:C exposure. These data further suggest that RbTLR1 may induce proinflammatory cytokines in rock bream under viral infection.

For the first time in a teleost, we identified a transcriptional inductive response of the TLR1 gene upon stimulation with a DNA virus, RBIV (Fig. 6b), revealing that TLR1 is potentially involved in recognizing PAMPs on DNA viruses. These data can be fairly supported by the previous evidence which claims that the glycoprotein PAMPs on human cytomegalovirus (a DNA virus), interacts with the TLR1/2 heterodimeric complex (Boehme et al. 2006). The pattern of IL-1β expression induced upon RBIV infection was found to comply with the transcriptional response of RbTLR1 (unpublished data). These data indicate the existence of a putative cytokine-inducible immune response pathway that can be triggered by the sensitivity of TLR1 to viral pathogens in rock bream. We attribute the downregulation of transcript at 24 h p.i to the host immune evasion mechanisms exerted by the virus during the late phase of the infection, since virus can orchestrate different protective mechanisms against host immunity, ensuring the survival of the virus in the host organisms (Iannello et al. 2006).

In summary, the complete coding sequence of RbTLR1 was identified from our custom constructed BAC genomic DNA library. We structurally characterized RbTLR1 and analyzed its transcriptional profile in healthy and immune-challenged animals. Phylogenetic analysis revealed the evolutionary proximity of RbTLR1 with orthologs of other vertebrates. The predicted genomic structure of RbTLR1 showed distinct arrangement to most of other fish species, considered. The immune response of RbTLR1 gene expression upon exposure to different PAMPs, bacterial and viral challenges, was similar to that of the proinflammatory cytokine IL-1β, providing evidence for the putative involvement of RbTLR1 in host immune defense. In addition, our data speculate that the PAMPs used in these experiments may be the potential TLR1 agonists in rock bream. RBIV induction increased the levels of RbTLR1 transcription in this study. However, future research should focus on elucidating the mechanisms of TLR1/2 heterodimeric complex-mediated sensing of DNA virus in teleost species. To fully evaluate the function of TLR1 in rock bream innate immunity, a comparative analysis between rock bream TLR1 and 2 is required to be performed since both TLR1 and 2 receptors together form a heterodimer for subsequent PAMP identification and initiation of the signaling cascade to mount an immune response as evidenced by previous studies.