Our genome and transcriptomic surveys revealed a total of 198 TLRs in 25 species (Table 1, Figure 3). No TLRs were found in 20 species. Additionally, our analysis also revealed a large number of TLR-like proteins (TIR-only or LRR-only). However, only sequences containing a TIR domain, a transmembrane domain and, at least, one LRR domain were considered as criteria for TLRs.
TLRs are absent in the genomes and transcriptomes of xenacoelomorphs and in some spiralians
Our surveys revealed that TLRs are absent in the genomes and transcriptomes of all Xenacoelomorpha, Platyhelminthes, Cycliophora, Micrognathozoa and Gastrotricha species analyzed (Table 1). Furthermore, TLRs are also absent in the transcriptomes of all the rotifer species investigated, except for E. senta (Table 1, Figure 3). Moreover, although TLRs were present in the bryozoan M. membranacea, they were not found in the transcriptome of the bryozoan B. neritina. However, although TLRs were not detected, TLR-like proteins were present in all these animal groups (data not shown).
Table 1. TLR genome/transcriptome survey results and classification of TLRs included in the phylogenetic analysis.
|
Species
|
TLRs
|
V-type /scc
|
P-type /mcc
|
NC
|
Reference
|
|
Cnidaria
|
|
|
|
|
|
|
Nematostella vectensis
|
1
|
0
|
1
|
0
|
L: [27]
|
|
Acropora digitifera
|
4
|
1
|
3
|
0
|
L: [69]
|
|
Acropora millepora
|
1
|
0
|
1
|
0
|
L: [69]
|
|
Orbicella faveolata
|
1
|
0
|
1
|
0
|
L: [80]
|
|
Xenacoelomorpha
|
|
|
|
|
|
|
Xenoturbella profunda
|
0
|
0
|
0
|
0
|
G: Unpublished
|
|
Hofstenia miamia
|
0
|
0
|
0
|
0
|
G: GCA004352715
|
|
Praesagittifera naikaiensis
|
0
|
0
|
0
|
0
|
G: PRJDB7329
|
|
Isodiametra pulchra
|
0
|
0
|
0
|
0
|
G: Unpublished
|
|
Meara stichopi
|
0
|
0
|
0
|
0
|
G: Unpublished
|
|
Convolutriloba macropyga
|
0
|
0
|
0
|
0
|
T: [97]
|
|
Bryozoa
|
|
|
|
|
|
SPIRALIA
|
Membranipora membranacea
|
6
|
4
|
1
|
1
|
T: SRX1121923
|
Bugula neritina
|
0
|
0
|
0
|
0
|
T: [98]
|
Cycliophora
|
|
|
|
|
|
Symbion pandora
|
0
|
0
|
0
|
0
|
T: [99]
|
Annelida
|
|
|
|
|
|
Galathowenia oculata
|
39
|
18
|
12
|
9
|
T: Unpublished
|
Eisenia fetida
|
11
|
0
|
1
|
10
|
T: SRX3108745
|
Helobdella robusta
|
4
|
1
|
3
|
0
|
G: [100]
|
Phyllochaetopterus prolifica
|
3
|
1
|
0
|
2
|
L: [64]
|
Mollusca
|
|
|
|
|
|
Crassostrea gigas
|
12
|
10
|
2
|
0
|
G: [101]
|
Octopus bimaculoides
|
9
|
1
|
6
|
2
|
G: [102]
|
Cyclina sinensis
|
2
|
1
|
1
|
0
|
L: [85]
|
Leptochiton rugatus
|
1
|
0
|
0
|
1
|
L: [64]
|
Biomphalaria glabrata
|
27
|
16
|
10
|
1
|
G: [84]/NCBI
|
Brachiopoda
|
|
|
|
|
|
Terebratalia transversa
|
15
|
4
|
4
|
7
|
T: [97]
|
Hemithris psittacea
|
6
|
3
|
1
|
2
|
T: [64]
|
Lingula anatina
|
25
|
15
|
7
|
3
|
G: [103]
|
Micrognathozoa
|
|
|
|
|
|
Limnogathia maerski
|
0
|
0
|
0
|
0
|
T: SRX1121929
|
Gastrotricha
|
|
|
|
|
|
Lepidodermella squamata
|
0
|
0
|
0
|
0
|
T: [104]
|
Macrodasys sp
|
0
|
0
|
0
|
0
|
T: [105]
|
Megadasys sp
|
0
|
0
|
0
|
0
|
T: [105]
|
Diuronotus aspetos
|
0
|
0
|
0
|
0
|
T: SRX1121926
|
Mesodasys laticaudatus
|
0
|
0
|
0
|
0
|
T: SRX872416
|
Nemertea
|
|
|
|
|
|
Lineus longissimus
|
10
|
7
|
2
|
1
|
T: [97]
|
Lineus ruber
|
6
|
2
|
3
|
1
|
T: Unpublished
|
Notospermus geniculatus
|
7
|
5
|
1
|
1
|
G: [86]
|
Paranemertes peregrina
|
2
|
1
|
0
|
1
|
L: [64]
|
Phoronida
|
|
|
|
|
|
Phoronopsis harmeri
|
2
|
0
|
1
|
1
|
T: SRX1121914
|
Phoronis australis
|
24
|
14
|
8
|
2
|
G: [86]
|
Phoronis psammophila
|
3
|
1
|
1
|
1
|
L: [64]
|
Phoronis vancouverensis
|
6
|
5
|
0
|
1
|
L: [64]
|
Platyhelminthes
|
|
|
|
|
|
Macrostomum lignano
|
0
|
0
|
0
|
0
|
G: [106]
|
Echinococcus multilocularis
|
0
|
0
|
0
|
0
|
G: [107]
|
Hymenolepis microstoma
|
0
|
0
|
0
|
0
|
G: [107]
|
Rotifera
|
|
|
|
|
|
Epiphanes senta
|
1
|
1
|
0
|
0
|
T: Unpublished
|
Rotaria tardigrada
|
0
|
0
|
0
|
0
|
T: [108]
|
Echinorhynchus gadi
|
0
|
0
|
0
|
0
|
T: SRX1121912
|
Macracanthorhynchus hirudinaceus
|
0
|
0
|
0
|
0
|
T: [105]
|
|
Priapulida
|
|
|
|
|
|
ECDYSOZOA
|
Priapulus caudatus
|
3
|
0
|
3
|
0
|
T: [97]
|
Halicryptus spinulosus
|
4
|
1
|
3
|
0
|
T: [97]
|
Tardigrada
|
|
|
|
|
|
Hypsibius exemplaris
|
1
|
0
|
1
|
0
|
G: [109]
|
Ramazzottius varieornatus
|
1
|
0
|
1
|
0
|
G: [110]
|
Onychophora
|
|
|
|
|
|
Peripatopsis capensis
|
1
|
0
|
0
|
1
|
T: [111]
|
Nematoda
|
|
|
|
|
|
Loa loa
|
1
|
0
|
1
|
0
|
G: [112]
|
Onchocerca volvulus
|
1
|
0
|
1
|
0
|
G: [113]
|
Caenorhabditis elegans
|
1
|
0
|
1
|
0
|
L: NCBI
|
Loricifera
|
|
|
|
|
|
Armorloricus elegans
|
2
|
1
|
1
|
0
|
T: SRX1120677
|
Arthropoda
|
|
|
|
|
|
Daphnia pulex
|
5
|
2
|
3
|
0
|
G: [114]
|
Drosophila melanogaster
|
9
|
1
|
8
|
0
|
L: NCBI
|
Ixodes scapularis
|
5
|
1
|
3
|
1
|
L: [115]
|
DEUTEROSTOMIA
|
Tunicata
|
|
|
|
|
|
Ciona intestinalis
|
2
|
1
|
1
|
0
|
L: [94]
|
Oikopleura dioica
|
1
|
1
|
0
|
0
|
L: [93]
|
Echinodermata
|
|
|
|
|
|
Strongylocentrotus purpuratus
|
8
|
7
|
1
|
0
|
L: [62]
|
Craniata
|
|
|
|
|
|
Homo sapiens
|
10
|
10
|
0
|
0
|
L: NCBI
|
NC column indicates the number of TLRs that could not be classified for each species. In the reference column it is indicated whether the survey was performed in a genome (G) or a transcriptome (T), followed by its reference or NCBI accession number. For protein sequences extracted from the literature (L), the paper source is also cited. NCBI indicates that sequences were collected individually from NCBI database. For further details, see Supplementary Table 1.
The number of TLRs detected in members of Ecdysozoa is low when compared to Spiralia and Deuterostomia
The TLR survey of the ecdysozoan genomes and transcriptomes revealed only one TLR for the tardigrade, nematode, and onychophoran species analyzed (Table 1, Figure 3). Furthermore, we detected up to 4 different TLRs in priapulids, 2 in loriciferans, and 5 in arthropods.
Multiple TLRs are detected in trochozoan species
TLRs were found in the genomes/transcriptomes of all trochozoan species analyzed (Table 1, Figure 3). Our results reveal that, in general, multiple TLRs are present in highly variable numbers in trochozoan species. The number of TLRs is not reflected by the phylogeny, meaning that species belonging to a same clade do not have a more similar number of TLRs than species belonging to another clade. For instance, the annelids G. oculata and H. robusta have 39 and 4 TLRs, respectively, while the Phoronid P. harmeri has 2. Thus, in this case, the number of TLRs is more similar between an annelid and a phoronid than between two annelids. This is explained by the multiple duplications and losses that have independently occurred in the Toll receptor family during trochozoan evolution.
P-type/mcc and V-type/scc are not specific for any planulozoan clade
Previous studies suggest that V(ertebrate)-type/scc and P(rotostome)-type/mcc TLRs are restricted to vertebrates and protostomes, respectively [62]. However, our results show that both, P-type/mcc and V-type/scc type TLRs, are present in cnidarians, spiralians, ecdysozoans, and deuterostomes (Table 1; Supplementary Table 2). V-type/scc TLRs are the most abundant TLR type in the spiralian species analyzed. However, many spiralians also have several P-type/mcc TLRs. P-type/mcc TLRs are the predominant TLR type in the ecdysozoan species included in this analysis. For nematodes, tardigrades and onychophorans, which only have one TLR, this TLR was always classified as P-type/mcc. Ecdysozoan species analyzed with more than one TLR have one or more P-type/mcc TLRs and only one V-type/scc. Although the vertebrate TLR complement seems to only contain V-type/scc TLRs [14, 65, 116, 117], P-type/mcc TLRs are also present in other deuterostomes, such as the tunicate C. intestinalis [94] and the echinoderm S. purpuratus [62] (Table 1, Supplementary Table 2). This suggests that P-type/mcc TLRs were lost in the lineage to the Craniata.
TLRs form three well-supported clades
Our phylogenetic analysis showed that TLRs group into three clades (Figure 4A), which we named clade α (89 TLRs), clade β (102 TLRs) and clade γ (79 TLRs). Although these three clades are well supported (>60), some of the internal nodes have low support values (<60). The phylogenetic analysis showed that clades β and γ are sister clades and together form the sister group to clade α. All three clades contain both P-type/mcc and V-type/scc TLRs, which makes it difficult to reconstruct whether P-type/mcc or V-type/scc show the ancestral state of TLRs. Furthermore, 2 deuterostome TLRs (from H. sapiens and C. intestinalis) and 11 spiralian TLRs (2 from species of mollusks and 9 from brachiopods) could not be assigned to any of the above clades. The 9 brachiopod TLRs form a clade with a high support value (>60), but do not group with either the mollusk or the deuterostome sequences. This TLR brachiopod clade is the sister clade to the three main clades (α, β and γ). For these sequences, the alignment showed brachiopod-specific deletions in the amino acid positions 150-220 that are not present in the TLRs belonging to the three main clades (Supplementary Figure 1). To investigate whether this insertion is causing the clustering of the TLRs into three clades, we performed a second phylogenetic analysis (Supplementary Figure 2) with the same parameters of the main analysis (Figure 4A) but excluding the 150-200 amino acid region. The second analysis (Supplementary Figure 2) is able to reconstruct clade α with high support value (>60). However, clade γ is nested within clade β and both of them have low support values (<60). In the second analysis (Supplementary Figure 2), as in the main analysis (Figure 4), the 9 brachiopod sequences cluster together and form the sister clade to the three main clades. However, in the analysis shown in Supplementary Figure 2, the mollusk and deuterostome sequences are included in the clade γ. In the main analysis (Figure 4A), no distinctive motifs were observed in the alignment that justify the exclusion of these sequences from the main clades.
Clade α includes TLRs from all cnidarian, spiralian and ecdysozoan species analyzed, except for the onychophoran TLR (Figure 4). Because all cnidarian TLRs cluster together, it is likely that only one TLR was present in the last common ancestor of Cnidaria. Clade β is formed by TLRs belonging to deuterostomes, spiralians and three ecdysozoans (two arthropods and the onychophoran TLR) (Figure 4). This suggests that at least the ancestral TLR of Clade β/γ was already present in the last common ancestor of Nephrozoa (Protostomia + Deuterostomia). Furthermore, lineage-specific expansions of clade β TLRs are detected in spiralians and deuterostomes. Clade γ TLRs are present in all trochozoan groups except for the nemertean species analyzed (Figure 4). Clade γ contains TLRs that radiated independently in several lineages. Our alignment shows that 159/181 TLRs belonging to the clades β and γ contain an insertion of 6 amino acids in the positions 349-354 (Supplementary Figure 1). In Clade α, this insertion is only present in Pcau-TLRα1, the sister TLR to all the remaining TLRs belonging to this clade. To exclude that this insertion causes the clustering in three distinct clades, we performed a third phylogenetic analysis (Supplementary Figure 3), in which we applied the same parameters as in the main analysis -shown in Figure 4A- but eliminated the 6 amino acid insertion regions. In the third analysis (Supplementary Figure 3), the three clades could be reconstructed with good support values (>60). However, due to low support values (<60), the relationship between the clades could not be resolved. Moreover, the clustering of the TLRs into the three clades (α, β, γ) was maintained with respect to the main analysis (Supplementary Figure 3, Figure 4A), except for eight phoronid and one human sequences. In the main analysis (Figure 4A), the phoronid sequences cluster together within clade γ, with high support values (>60). This clade of phoronid TLRs is the sister clade to all remaining TLRs in clade γ. Nevertheless, in the third analysis (Supplementary Figure 3), these phoronid TLR sequences constitute a well-supported (>60) clade within clade β, but it is not the sister clade to the remaining TLRs in this clade. In the main analysis (Figure 4A), the human sequence is not included in any of the three main clades, but in the third analysis (Supplementary Figure 3) it does cluster in clade α.
TLRs are expressed during development in the ecdysozoans P. caudatus and H. exemplaris and in the spiralians C. gigas and T. transversa
In order to study the temporal expression of TLRs during ontogeny, we analyzed stage-specific transcriptomes of the priapulid P. caudatus [118], the tardigrade H. exemplaris [119], the mollusk C. gigas [101] and the brachiopod T. transversa [120]. All the analyses were performed using both RSEM [121] and kallisto [122] methods.
The expression of the only TLR present in H. exemplaris was analyzed in stage-specific transcriptomes of 19 stages (one biological replicate) (Figure 5A; Supplementary Table 3) [119]. Expression of TLRα was detected (TMM ≥ 0.15) in time windows during development (zygote, morula, gastrula, elongation, segmentation and differentiation).
Three TLRs were identified in P. caudatus transcriptomic survey (Table 1). The expression of these TLRs was analyzed in five embryonic stages (two biological replicates) (Supplementary Table 4) [118]. Our results indicate that all three TLRs found in the transcriptomic survey are expressed during embryonic development (TMM ≥ 0.15). Pca-TLRα1 and Pca-TLRα2 are expressed in all developmental stages analyzed, whereas Pca-TLRα3 is expressed only in the later embryonic stages (Figure 5B; Supplementary Table 4).
The expression of the 12 C. gigas TLRs (Table 1) was analyzed in stage-specific transcriptomes of 19 stages (one biological replicate) (Supplementary Table 5) [101]. Our results show that at 11 of the 12 TLRs are expressed during development (Figure 5C; Supplementary Table 5). Some TLRs are expressed throughout development (Cgi-TLRα1, Cgi-TLRα4, Cgi-TLRβ4, Cgi-TLRδ1, Cgi-TLRδ2), while others (Cgi-TLRα2, Cgi-TLRα3, Cgi-TLRβ1, Cgi-TLRβ2, Cgi-TLRγ1, Cgi-TLRγ2) are only expressed at certain developmental stages. Cgi-TLRβ3 expression was not detected at any of the stages analyzed.
15 TLRs were found in our transcriptome survey of T. transversa (Table 1). Expression of these TLRs was analyzed in stage-specific transcriptomes of 12 developmental stages (with two biological replicates) [120]. Our results suggest that at least 12 of the 15 TLRs are expressed at certain stages during T. transversa development (Figure 5D; Supplementary Table 6). Ttr-TLRα2, Ttr-TLRα5, Ttr-TLRβ1, Ttr-TLRβ4, Ttr-TLRβ5, and Ttr-TLRδ expression is detected in time windows during embryonic and larval stages. All these genes, except Ttr-TLRβ5, are expressed in juveniles. For some genes (Ttr-TLRα4, Ttr-TLRβ2, Ttr-TLRβ3, and Ttr-TLRγ4), expression was detected throughout development. Moreover, expression was not detected at the embryonic and larval stages analyzed for Ttr-TLRα1, Ttr-TLRγ1, Ttr-TLRγ2 and Ttr-TLRγ3. Similarly, Ttr-TLRα3 expression was only detected in the competent larvae and in the juveniles.
Our analyses show that TLRs are expressed during the development of the spiralians T. transversa and C. gigas and the ecdysozoans P. caudatus and H. exemplaris, which suggests that these genes could be involved both in development and immunity during ontogeny. Furthermore, these analyses show that the TLRs expressed during development are not restricted to one TLR clade in the tree shown above, but they are found in all three main clades (e.g. Ttr-TLRα4, Ttr-TLRβ3, Cgi-TLRγ1).
Furthermore, in order to validate our stage specific transcriptome results, we performed whole mount in situ hybridization (WMISH) for the T. transversa mRNAs of TLRα2, TLRα3, TLRα4, TLRα5, TLRβ3, TLRγ4 and TLRδ (Figure 6). Consistently with our stage specific transcriptomic analysis, our WMISH results show that Ttr-TLRα2 is not expressed at the late gastrula stage (Figure 6A), but the expression is present in the mesoderm and in two pairs of lateral domains in early larvae (Figure 6B). This gene is not expressed in late larvae (Figure 6C). In agreement with our stage specific transcriptomic analysis, we did not detect Ttr-TLRα3 either in late gastrulae or in the two larval stages analyzed (Figure 6D-F). Ttr-TLRα4 has a dynamic expression pattern during T. transversa development. This gene is expressed in the mesoderm at the early gastrula stage, but, consistent with the stage specific transcriptome analysis, it is not detected in late gastrulae (Figure 6G-H). In early larvae, Ttr-TLRα4 is expressed in the inner lobe epithelium and in a medial V-shaped mesodermal domain (Figure 6I). In late larvae, this gene is expressed in the brain and in the pedicle (Figure 6J). Consistently with the stage specific transcriptomic analyses, mRNA of Ttr-TLRα5 is detected in a uniform salt and pepper distribution at the late gastrula stage and the two larval stages for which WMISH was performed (Figure 6K-M). Congruently with the stage specific transcriptomic analyses, Ttr-TLRβ3 is expressed in the anterior region of the animal in late gastrulae (Figure 6N). However, although Ttr-TLRβ3 expression was detected in early larvae in the stage specific transcriptome analysis, expression was not detected by WMISH (Figure 6M). Furthermore, Ttr-TLRβ3 is not expressed in the late larvae (Figure 6P). The expression of Ttr-TLRγ4 and Ttr-TLRδ have a uniformly salt and pepper distribution at the late gastrula and early larvae stages (Figure 6 Q-R and T-U). This salt and pepper transcript distribution is similar in late larvae, although it is absent from the pedicle lobes (Figure 6 S and V). These results conflict with the stage specific transcriptome analyses, as, in this analysis, neither Ttr-TLRγ4 expression was detected in the early larvae nor Ttr-TLRδ in any of the two larval stages tested. Differences between the results of both analyses could be explained by differences and variation of the developmental stages of the specimens used for the stage-specific transcriptome and the WMISH.