Coevolution of paired receptors in Xenopus carcinoembryonic antigen-related cell adhesion molecule families suggests appropriation as pathogen receptors

Abstract

Background

In mammals, CEACAM1 and closely related members represent paired receptors with similar extracellular ligand-binding regions and cytoplasmic domains with opposing functions. Human CEACAM1 and CEACAM3 which have inhibitory ITIM/ITSM and activating ITAM-like motifs, respectively, in their cytoplasmic regions are such paired receptors. Various bacterial pathogens bind to CEACAM1 on epithelial and immune cells facilitating both entry into the host and down-regulation of the immune response whereas interaction with granulocyte-specific CEACAM3 leads to their uptake and destruction. It is unclear whether paired CEACAM receptors also exist in other vertebrate clades.

Results

We identified more than 80 ceacam genes in Xenopus tropicalis and X. laevis. They consist of two subgroups containing one or two putative paired receptor pairs each. Analysis of genomic sequences of paired receptors provide evidence that their highly similar ligand binding domains were adjusted by recent gene conversion events. In contrast, selection for diversification is observed among inhibitory receptor orthologs of the two frogs which split some 60 million years ago. The allotetraploid X. laevis arose later by hybridization of two closely related species. Interestingly, despite the conservation of the genomic landscape surrounding the homeologous ceacam loci only one locus resembles the one found in X. tropicalis. From the second X. laevis locus more than 80 % of the ceacam genes were lost including 5 of the 6 paired receptor genes. This suggests that once the gene for one of the paired receptors is lost the remaining gene cluster degrades rapidly probably due to lack of selection pressure exerted by pathogens.

Conclusions

The presence of paired receptors and selection for diversification suggests that also in amphibians CEACAM1-related inhibitory proteins are or were used as pathogen receptors.

Background

A number of families of cell surface receptor with very similar extracellular domains and inhibitory or activating intracellular signaling motifs have been identified in vertebrates. The best investigated families represent the KIR, Ly49, Nkpr, SIGLEC, SIRP and CEACAM families [1]. These so called paired receptors are commonly encoded in the same gene cluster and some are thought to play a role in homeostasis of the immune system by controlling activation and downregulation of immune reactions [2]. Many of the inhibitory members of paired receptors are expressed on natural killer (NK) cells where they sense major histocompatibility antigens (MHC) present on uninfected cells leading to tolerance. Loss of MHC expression frequently found in virus infected cells releases NK cell inhibition with concomitant destruction of the infected cells. Activating members of paired receptors seem to have evolved to counter common virus immune escape mechanisms in serving as decoy receptors on NK cells. They recognize virally encoded fake MHC self-molecules that are expressed by virus infected cells thus overcoming viral immune escape by NK cell activation [1].

Other paired receptors directly interact with viral or bacterial pathogens [1]. Among those are SIRPα and CEACAM1 and CEACAM3, members of the human carcinoembryonic antigen-related cell-cell adhesion molecule (CEACAM) family which have inhibitory ITIM/ITSM motifs and activating ITAM-like motifs in their cytoplasmic regions, respectively [3, 4]. A number of bacterial pathogens like pathogenic Neisseria (N. gonorrhoeae, N. meningitis) Haemophilus influenzae and Moraxella catarrhalis have been shown to bind to the N-terminal immunoglobulin (Ig) variable-like domain of CEACAM1 on epithelial and immune cells allowing both entry into the host by transcytosis and down-regulation of the host’s immune response by inhibiting adaptive and innate immune reactions [5–11]. Pathogens thus exploit the normal physiological function of CEACAM1 which acts as an immune inhibitory receptor on leukocytes upon homotypic or heterotypic interactions for example with other CEACAM members [7, 12]. In contrast, binding to granulocyte-specific CEACAM3 leads to uptake and destruction of these pathogens by triggering bactericidal processes [13–16]. Interestingly, phylogenetically unrelated adhesins such as opacity-associated (Opa) protein, outer membrane protein P5 and ubiquitous surface protein (UspA1) mediate interaction with the pathogen receptor CEACAM1 indicating convergent evolution [17–19].

A host-pathogen arms race involving receptors and decoy receptors with very similar adhesin-binding domains should lead to selection of pathogens with preferential binding to the inhibitory receptor and reduced binding to its decoy counterpart. Indeed, clinical isolates of N. gonorrhoeae from male urethra and female genital tract often express Opa proteins which bind to CEACAM1 but not to CEACAM3 [20]. The capability of Neisseria to randomly switch on expression of variant Opas from a panel of Opa genes aides natural selection from a heterogenous Neisseria population. On the other hand, individuals with variant CEACAM1 receptors with low or no binding to pathogens should have an selective advantage. This will inevitably lead to poorly matched paired receptors and loss of decoy function. Intrachromosomal recombination or gene conversion between exons encoding ligand-binding domains of inhibitory and activation receptors within the CEACAM gene cluster could correct this deficit. Indeed, replacement of part of CEACAM3 exon 2 encoding the ligand-binding domain with sequences from the corresponding exon of CEACAM1 has happened in humans [3, 21].

CEACAM families differ greatly in gene number and domain composition of the encoded proteins between mammalian species. Most of the analyzed mammals also contain putative paired CEACAM receptors [3, 22]. Allelic variants of CEACAM1 in mice and cattle have been shown or are suspected to serve as coronavirus receptors [23, 24]. Therefore, the rapid divergence of CEACAM1 and corresponding activating receptors during mammalian evolution is thought to be pathogen-driven [1, 3, 22].

Also more distantly related CEACAM genes exist in mammals (CEACAM16, CEACAM18, CEACAM19 and CEACAM20) which do not represent paired receptors. They are clustered distally from the CEACAM1-related genes. They differ in domain organization and sequence among each other and exhibit specialized functions [25–27]. However, they are conserved between mammalian species which allows unequivocal assignment of orthologs [3].

CEACAM gene families seem to be restricted to vertebrates. CEACAM family members have been recently identified in reptiles, amphibians and in bony and cartilaginous fishes [28, 29]. However, the exact composition, the presence of paired receptors and the driving forces behind their evolution have not been investigated. Here we present comprehensive analyses of the ceacam families of two clawed frog species; the western clawed frog Xenopus tropicalis and the African clawed frog X. laevis the ancestors of which split some 60 million years ago [30]. We identified two distantly related ceacam families which both contain rapidly evolving paired receptors. Analysis of the ceacam family in X. laevis allowed us to follow the fate of a group of rapidly evolving genes after allotetraploidization.

Results

Identification of ceacam gene families in X. tropicalis and X. laevis

Based on their syntenic location between the flanking genes lipe and bcl3, and the presence of exons with conserved phasing encoding Ig variable (IgV)- and Ig constant (IgC)-like domains and ITIM and ITAM-like motifs most similar to mammalian CEACAM members (Fig. 1 and Additional file 1) 44 and 38 ceacam genes were identified on chromosomes 7 in X. tropicalis and X. laevis, respectively (Additional files 2 and 3). Interestingly, two ceacam gene loci exist in X. laevis on the homeologous chromosomes 7 L and 7S generated during speciation by hybridization of closely related species (Fig. 1; for nomenclature see [31]). Amino acid sequence comparison of the N-terminal IgV-like domains (N domains) revealed the presence of two distantly related subgroups group 1 and group 2 in both species (Fig. 2). N domains were chosen because they represent functionally important domains which have been shown in other species to be responsible for ligand binding [32]. Group 1 and group 2 genes are localized in clusters next to each other and, different from mammals, are not disrupted by non-ceacam genes (Fig. 1). Group 1 and group 2 Ceacam N domain amino acid sequences are most closely related to cartilaginous and bony fish and reptile and mammalian CEACAM N domain sequences, respectively (~35 % identity). Within subgroup 1 and 2, members exhibit between 40 and 93 % N exon amino acid sequence identity, while between subgroups only 20-30 % sequence identity is observed (data not shown). Similarly, transmembrane and cytoplasmic sequences also exhibit a higher degree of identity within groups than between groups (Additional file 1 and data not shown). Despite the low sequence identity group 1 and group 2 Ceacam IgV-like domains exhibit a very similar three-dimensional structure predicted by modeling using corresponding human and murine CEACAM1 sequences (Additional file 4).

Fig. 1

Chromosomal arrangement of ceacam genes in X. tropicalis and X. laevis. Arrowheads represent genes with their transcriptional orientation. Xenopus ceacam gene clusters are subdivided in group 1 (yellow) and group 2 members (blue). For X. laevis two homeologous ceacam clusters are found on chromosome 7 L and 7S which were generated by hybridization during speciation. Homeologs are indicated by L and S. Human CEACAM1-related genes are indicated in yellow, when predominantly expressed in trophoblast cells in red; the CEACAM genes for which orthologs can be identified in mammals are shown in blue and marker genes in black. Their syntenic relationship is shown by blue lines. Names of CEACAM1-like genes with ITIM-encoding exons are shown in red and with ITAM and ITAM-like motif-encoding exons in green and blue, respectively. Red double arrows symbolize potential recombination events between regions with ceacam genes containing ITIM or ITAM/ITAM-like motifs (genes shown in red boxes). Note that, in general, the genes in the same subcluster show the same, in different subclusters opposite transcriptional orientation. The nucleotide numbering of the chromosomes starts at the telomere of the short arms located to the left. The databases and the versions used are indicated below the species name. c, ceacam, C, CEACAM; chr, chromosome; P, pregnancy-specific glycoprotein (PSG) genes

Fig. 2

Phylogenetic relationship of Xenopus Ceacam proteins. Phylogenetic trees were constructed based on mature N domain amino acid sequences (signal peptides excluded) from X. tropicalis (a) and X. laevis Ceacams (b) using the Maximum Likelihood method (MEGA6 software). The tree with the highest log likelihood is shown. The percentage of trees in which the protein sequences clustered together is shown next to the branches. For Ceacams predicted to contain more than one N domain, only the most N-terminal N1 domains were included in the analysis. Two distantly related Ceacam groups of N domains termed group 1 and group 2 can be discriminated in both species. Some of the most closely related Ceacams in the allotetraploid X. laevis represent homeologs located on the small (S) or the large (L) version of chromosome 7. A closely related pair of paralogous proteins termed .1 and .2 is found on chromosome 7S but not on chromosome 7 L. The bar below the phylogenetic tree shows the scale for the number of substitutions per site. P, pseudogene; Xla, X. laevis; Xtr, X. tropicalis

Taken together, this indicates that two ceacam groups exist in Xenopus which were probably derived early in amphibian evolution possibly from two different ceacam ancestors and their origin predates the divergence of X. laevis and X. tropicalis.

Groups of paralogous Ceacams contain paired receptors

In mammals, CEACAM families consist of a group of orthologous members (CEACAM16, CEACAM18, CEACAM19, CEACAM20) where counterparts can be clearly assigned in different species and a group of paralogous members which are most closely related to CEACAM1 within the same species [3]. To identify orthologous Ceacam pairs as well as Ceacam paralogs, X. tropicalis and X. laevis group 1 and group 2 amino acid sequences from mature N domains (signal peptide sequence removed) were compared and their relationship displayed as dendrograms. In group 1 and group 2, two and seven pairs of orthologous Ceacams, respectively, could be identified based on their degree of sequence identity (Fig. 3). Their predicted domain organization is heterogenous. Three members consist of only one IgV-like domain and are either secreted or membrane-bound by a transmembrane domain or a GPI anchor while two transmembrane-bound orthologous pairs are composed of one IgV- and one IgC-like domain and an ITAM-containing cytoplasmic region (Fig. 4).

Fig. 3

Orthologous and paralogous members of the Ceacam family in X. tropicalis and X. laevis. Phylogenetic trees were constructed based on mature N domain amino acid sequences (signal peptides excluded) from X. tropicalis and X. laevis group 1 (a) and group 2 Ceacams (b) using the Maximum Likelihood method (MEGA6 software). The tree with the highest log likelihood is shown. The percentage of trees in which the protein sequences clustered together is shown next to the branches. For Ceacams predicted to contain more than one N domain, only the most N-terminal N1 domains were included in the analysis. Ceacams whose N domains exhibit the highest degree of identity within the same species (paralogs) are boxed with red and blue lines for X. tropicalis and X. laevis, respectively. Pairs consisting of most closely related proteins from different species (sometimes including recently generated paralogs) are highlighted in gray (orthologs). The names of ITIM-containing members are highlighted in red, proteins with ITAM and ITAM-like motifs are marked with green and blue background, respectively. P, pseudogene; Xla, X. laevis; Xtr, X. tropicalis

Fig. 4

Domain organization of Xenopus Ceacam proteins. The domain organization of Ceacam family members from X. tropicalis and X. laevis was predicted by gene analysis only (domains shown in light colors) or, where available, confirmed by EST and cDNA sequences (domains shown in intense colors). Only Ceacams whose complete domain organization could be delineated were included. Homeologous Ceacams in X. laevis arisen by hybridization during speciation are discriminated by L (encoded on long chromosome variant 7) and S (encoded on short chromosome variant 7). For comparison the domain organization of human CEACAM proteins is shown. The conserved human CEACAM family members are highlighted in green. IgV-like domains are shown as red, IgC-like domains are blue ovals. The predicted signaling motifs in the cytoplasmic domains are schematically shown as green (ITAM), blue (ITAM-like motif), red (ITIM) and yellow boxes (ITSM). Transmembrane domains and GPI anchors are indicated by black and green lines, respectively. Orthologous relationship as suggested by sequence relatedness and/or synteny is indicated by gray boxes. Note that Xtr_Ceacam301 and Xla_Ceacam326 appear to represent orthologs based on the degree of N domain sequence identity (see Fig. 3) despite their different domain organization. Paired receptors identified by their similar N domain sequences and the presence of ITIM/ITSM or ITAM/ITAM-like motifs are connected by gray lines. C, CEACAM or Ceacam; P, PSG

In addition, sets of proteins exist whose closest relatives are found in the same species thus representing paralogous proteins: one and two in X. tropicalis and X. laevis group 1 Ceacams, respectively, and one in each species in group 2 (Fig. 3). Interestingly, these groups of closely related Ceacam members harbor one member with an ITIM and one or more with ITAM-like motifs (Fig. 3; Additional file 1). Pairs of cell surface proteins with similar extracellular domains which are able to interact with the same ligand, however, transmitting opposing i.e. inhibitory or activating signals represent so called paired receptors. Based on these definition, in X. tropicalis Ceacam301 and Ceacam350 with an ITIM and Ceacam303 and Ceacam351 or Ceacam368 with an ITAM-like motif and an ITAM, respectively, and correspondingly in X. laevis Ceacam325 and Ceacam326 (ITIM) and Ceacam327 or Ceacam328 and Ceacam332 (ITAM) and Ceacam389 (ITIM) and Ceacam387 (ITAM) could represent paired receptors (Figs. 3 and 4). These putative paired receptors share between 80 and 93 % of their IgV-like ligand binding domain amino acid sequences (data not shown).

In summary, orthologous and paralogous members exist in both Ceacam groups. Putative paired receptors could be identified among the paralogous members.

Similarity of receptor binding domains of paired receptors is maintained by recombination

In paired receptors, pathogen-binding regions have to stay similar thus allowing the host to counterbalance the immune suppressive signal elicited by engagement of the inhibitory receptor through the pathogen by providing an activating receptor as a mimic [33, 34]. In other paired receptor gene families this is often achieved by recombination or gene conversion between genes encoding inhibitory and activating receptors, restricted to the gene region encoding the ligand binding domains [33]. We, therefore, screened the potential Xenopus paired receptor genes for recombination/gene conversion events. Indeed, in X. tropicalis group 1 ceacam301 and ceacam303 and in group 2 ceacam350 and ceacam351/ceacam368 N exons have recently undergone gene conversion which is exactly restricted to the exon just including the splice consensus sequences. This is supported by the high conservation of N exon sequences with virtual absence of synonymous mutations (which in general do not encounter purifying selection) and very low sequence conservation in the flanking introns (Fig. 5a, b, c; Additional file 5). No other genomic regions seem to have been involved in the gene conversion event in ceacam301 and ceacam303 (Fig. 5b). Recombination events were also noticed in N domain exons of other putative paired receptor gene pairs. This was evident from the lack or low rate accumulation of synonymous mutations mostly restricted to certain regions of the N exons of receptor pairs like ceacam325 and ceacam328 and possibly of ceacam389 and ceacam387 in X. laevis and ceacam350 and ceacam368 in X. tropicalis (Fig. 5d-f). Without recurrent recombination/gene conversion events between paralogous genes one would expect a steady accumulation of synonymous nucleotide changes along the N exons as it is found between X. tropicalis and X. laevis ceacam orthologs, like ceacam362 and ceacam380 (Fig. 5g). Interestingly, possibly due to continued pressure from pathogen adhesin-receptor interaction, regions which represent putative interaction sites (the CC’C”FG face of the Ig fold [35]) appear to rapidly accumulate non-synonymous mutations occurring after a gene conversion event (Fig. 5c-f). For intrachromosomal recombination/gene conversion to take place, involved genes must exhibit opposite transcriptional orientation in order that homologous sequences can be aligned with looping out of the intervening sequences [3, 36]. Indeed, putative paired receptor genes in X. laevis and X. tropicalis from both ceacam subgroups exhibit opposite transcriptional orientation with group 1 and group 2 ITIM-encoding genes facing each other (Fig. 1).

Fig. 5

Evidence for gene conversion between putative paired receptor ceacam genes. (a) Alignment of exons 1 and 2 encoding the leader and N domain, respectively, and flanking introns of ceacam301 and ceacam303. Note the strong sequence conservation restricted to the N domain exon. The start codons are marked in green, the splice acceptor and splice donor sequences are highlighted in yellow and blue, respectively. (b) The nucleotide sequence of ITIM/ITSM-encoding ceacam301 gene was compared with that of ITAM-like motif-encoding ceacam303 from X. tropicalis. For contiguous stretches of nucleotides conserved between the gene pairs using a sliding window, the degree of identity was calculated and displayed as horizontal lines. The location of ceacam301 exons is indicated by numbered boxes. The genomic region involved in gene conversion is marked with a red box, the affected N exon is shown in red. Different repeat sequences are indicated by differently shaped forms. (c-f) The accumulation of nonsynonymous (green curves) and synonymous substitutions (red curves) along the N exons of putative paired receptor genes were determined after manual removal of gaps in the compared nucleotide sequences. The type of encoded signaling motif is indicated by the color of the gene name: red, ITIM/ITSM; green, ITAM; blue, ITAM-like motif. Stretches of codons with no or minimal accumulation of synonymous substitutions which run parallel to the x-axis suggest recent gene conversion/recombination events. Note the rapid accumulation of nonsynonymous mutations in the CC’C”FG β-strand regions (black broken lines) which indicates selection for diversification. This contrasts with conserved regions indicated by red broken lines. For comparison, N exon sequences of an orthologous gene pair (Xtr ceacam362/Xla ceacam380) were analyzed (g). The location of CC’C”and FG β-strand regions determined by 3D modeling (Additional file 4) are indicated by gray boxes above the graphs. N, N domain exon; TM, transmembrane domain exon; Xla, X. laevis; Xtr, X. tropicalis

The non-random transcriptional orientation of ceacam genes encoding proteins with inhibitory or activating signaling motifs, gene conversion within exons encoding ligand binding domains as well as the conservation of these domains in ITIM and ITAM-containing Ceacams strongly argue that these Ceacams function as paired receptors.

Selection for diversification in paired receptor Ceacam groups

Pathogen receptors which allow entry into a host often exhibit selection for diversification of their amino acid sequences. This is evident from high ratios (>1) of their rate of non-synonymous over their rate of synonymous mutations (dN/dS) especially in regions relevant for pathogen binding [37]. When whole domains or proteins are analyzed the dN/dS ratios can drop below 1 despite the presence of regions with strong positive selection, due to the presence of regions with negative or neutral selection. To test whether ITIM/ITSM-bearing Ceacam receptors in Xenopus might represent pathogen receptors as found for human and mouse CEACAMs we analyzed dN/dS ratios of N domain exons of ceacam orthologs in X. tropicalis and X. laevis. Orthologous genes encoding receptors with ITIM exhibited the highest dN/dS ratios, i.e. 1.3 for Xtr ceacam301/Xla ceacam326 (group 1) and 1.0 for Xtr ceacam350/Xla ceacam389.L orthologs (group 2) (Fig. 3; Fig. 6a). In contrast, dN/dS ratios between 0.18 and 0.65 were found for the other ceacam orthologs. The lowest dN/dS ratios were observed for the flanking non-ceacam genes (dN/dS = 0.1–0.25) with the exception of the immune function gene cd79a (dN/dS = 0.35), which encodes an ITAM-bearing component of the B cell receptor. The orthologous ceacam pair with the lowest dN/dS ratio (dN/dS = 0.18) encodes glycosylphosphatidylinositol (GPI) membrane-anchored proteins (Fig. 6a). No large dN/dS ratio differences were found when X. tropicalis genes were compared with the X. laevis homeologs on chromosome 7S (Fig. 6a) which are paralogous genes that were formed by the hybridization event in X. laevis which probably occurred during speciation (see below). This suggests that there is no loss of function or gain of new function for one of the two homoeologs.

Fig. 6

Differential conservation of orthologous and homeologous Xenopus ceacam and flanking genes. Nucleotide sequences of N exons and the open reading frames from ceacam and flanking genes, respectively, from X. tropicalis were compared codon-wise with each of the two homeologous orthologs of X. laevis (a) or the X. laevis homeologs with each other (b) after manual removal of gaps and the ratio of the rate of nonsynonymous and synonymous mutations was calculated. The resulting dN/dS values were plotted using the X. tropicalis gene order on chromosome 7 (a) or the X. laevis gene order on chromosome 7S (b). In cases of recent gene duplication of one of the orthologs in X. tropicalis or X. laevis dN/dS ratios for all orthologous pairs were calculated and plotted as mean and deviation or standard deviation. The type of signaling motif present in the corresponding proteins can be inferred from the schematic representation of the protein domains below the graphs (see Fig. 4 for keys to domains and motifs). Chr, chromosome

N exon dN/dS ratios of ≥ 1 of Xenopus inhibitory receptor orthologs indicate selections for diversification typically observed in pathogen receptors.

Selective gene loss in ceacam locus on chromosome 7S created by allotetraploidization in X. laevis

In X. laevis, allotetraploidy was probably caused by hybridization during speciation which led to whole genome duplication including the ceacam gene locus on chromosome 7. Despite the overall conservation of the duplicated genomic region surrounding the ceacam loci only the locus on chromosome 7 L resembles that found in X. tropicalis. More than 80 % of ceacam genes were lost from the locus on chromosome 7S and only 1 and 3 genes are left from group 1 and group 2 genes, respectively (Fig. 1). Interestingly, no paired receptor pair is preserved at the 7S locus; only the group 2 inhibitory receptor gene (ceacam389.S) is retained. Notwithstanding the massive ceacam gene loss, a gene duplication event probably after X. laevis speciation led to the generation of two closely related ceacam376 genes (ceacam376.1.S and ceacacam376.2.S) next to the bcl3 gene (Fig. 1).

To determine whether selection occurs to maintain the function of both homeologous gene copies, dN/dS ratios for N domain exons or whole coding sequences were determined for ceacam and flanking genes, respectively. Only one ceacam gene which encodes a GPI-linked group 2 member reveals similar conservation with a dN/dS ratio of 0.2 as flanking genes which exhibit dN/dS ratios between 0.03 and 0.2 with the exception of cd79a exhibiting a dN/dS ratio of 0.6 (Fig. 6b). In addition, both copies of all homeologous genes exhibit similar dN/dS ratios when compared with the X. tropicalis orthologs (Fig. 6a). Therefore, both homeologous copies of ceacam390.L/390.S and the non-ceacam genes seem to be functional and under selective pressure. In contrast, much higher dN/dS ratios between 0.6 and 0.75 are observed for the remaining ceacam genes, indicating either lack of selection for conservation of the homeologous pairs or selection for diversity. The latter appears to be at work for the inhibitory receptor-encoding genes ceacam389.L/389.S which exhibit the highest dN/dS ratios (Fig. 6a, b). In addition, regions with increased accumulation of non-synonymous mutations which are similar to those found for orthologous or homeologous pairs of inhibitory receptors could be identified. They co-localize with the CC’C”FG β-sheet which represent putative pathogen binding sites (Fig. 7).

Fig. 7

Selection for sequence diversification in inhibitory receptor Ceacams. The accumulation of non-synonymous (green curves) and synonymous substitutions (red curves) along the N exons of orthologous (a-c) and homeologous inhibitory receptor gene pairs (d) were determined after manual removal of gaps in the compared nucleotide sequences. The names of the analyzed genes are indicated in the top left corner of the graph. Note the rapid accumulation of nonsynonymous mutations in the CC’C”FG β-strand regions (black broken lines) which indicates selection for diversification. This contrasts with conserved regions between CC’C” and FG β-strands indicated by red broken lines. The location of CC’C” and FG β-strand regions determined by 3D modeling (Additional file 4) are indicated by gray boxes above the graphs. Note the sequence identity of the 38 N terminal codons in the N exon of X. laevis homeologs ceacam389.L and ceacam389.S possibly caused by interchromosomal recombination. N, N domain exon; Xla, X. laevis; Xtr, X. tropicalis

Discussion

Amphibian ceacam gene family exhibits an ancestral genomic arrangement

In this comprehensive analysis ceacam genes could be identified in X. tropicalis and X. laevis based on synteny and structural homology despite a low degree of sequence identity which is due to rapid divergence during evolution. The ceacam gene family consists of two distantly related subgroups (group 1 and group 2) with 15–20 members each. Only few orthologous ceacam members (mostly in group 2) were found in the two frog species. The Xenopus ceacam genes are arranged in one cluster separated by subgroups uninterrupted by non-ceacam genes (Fig. 1). This is also observed for most of the CEACAM genes in the marsupial opossum (Monodelphis domestica) but not in eutherian mammals. Here, the CEACAM locus is interrupted by two large regions with non-CEACAM genes (Fig. 1; [3]). This indicates that a continuous ceacam gene cluster was also present in the last common ancestor of amphibians and mammals.

No CEACAM16, CEACAM18, CEACAM19 and CEACAM20 genes, which are well conserved in mammals and are clustered next to BCL3, were found in Xenopus [3]. However, CEACAM19 orthologs can be identified unequivocally in reptiles including turtles, snakes, alligators and gecko but not the other genes ([29]; Zimmermann, unpublished results). Interestingly, group 2 Xenopus ceacam genes which are also located next to bcl3 but not group 1 genes are most closely related to reptilian CEACAM19 (32–37 % identity) representing the most common hits outside of the anuran order when Xenopus Ceacam group 2 N exon amino acid sequences are used as query sequences. This might indicate that CEACAM19 in reptiles and mammals and group 2 ceacam genes share a common ancestor.

Two paired receptor systems exist in the Xenopus Ceacam family which exhibit signs of pathogen-mediated selection

Each subgroup contains one or two paired receptors with oppositely signaling ITIM and ITAM/ITAM-like motifs and highly similar ligand binding domain (N domain) amino acid sequences (Figs. 3 and 4). This is different from mammalian CEACAMs which typically have only one set of paired receptors or none as found for mouse and rat [3].

What is the evidence that the paired Ceacam receptors are being or have been used as pathogen receptors in Xenopus? Both diversification of pathogen receptors to avoid binding of the pathogen (indicated by high dN/dS ratios) and maintenance of similarity of the pathogen adhesin-interacting domain in paired receptors functioning as pathogen or decoy receptors will be selected for in a pathogen/host arms race [1, 34, 37]. Indeed this is observed in both Xenopus Ceacam groups. The exons encoding the ligand-binding domains of X. tropicalis and X. laevis ITIM-containing orthologs exhibit the highest dN/dS ratios which is indicative of positive selection (Fig. 6a). Positively selected amino acid positions of a protein domain are expected to reside at the site of contact between the pathogen adhesin and its receptor [37]. This seems to be the case for putative Ceacam pathogen receptor orthologs in X. tropicalis and X. laevis which show selective accumulation of non-synonymous mutations in the CC’C” FG face of the IgV-like domain (Fig. 7a-c) responsible for pathogen interaction [35]. In contrast, the Ig β-sheet on the opposite side of the Ig fold is highly conserved. This recurrent host escape and pathogen adaptation (“Red Queen” scenario [38]) can lead to an imbalance with pathogen binding to the entry/inhibitory receptor being maintained while the decoy receptor-pathogen interaction is abolished [20]. This problem can be resolved through recombination/gene conversion whereby the decoy receptor pathogen binding domain is replaced by that of the pathogen receptor. Absence of synonymous mutations in large regions of the N exons revealed by comparison of the putative pathogen receptor and decoy receptor sequences suggest intra- or interchromosomal recombination or gene conversion events as cause of the similarity of paired receptor ligand binding domains (Fig. 5c-e). Original sequence identity is rapidly lost due to the on-going struggle between host and pathogens again noted by the selective accumulation of non-synonymous mutations in the presumed adhesin binding regions (Fig. 5c-e).

Comparative analyses of mammalian CEACAM loci revealed inversion of regions with non-CEACAM genes between oppositely oriented CEACAM1 and presumed decoy receptors genes in some species. This suggested that recombination involved an intrachromosomal loop formation mechanism that allows alignment of the exon sequences encoding the ligand binding domains [3]. The recombination mechanism in Xenopus is still unclear. However inverted transcriptional orientation of ITIM- and ITAM-bearing genes indicates a similar intrachromosomal recombination mechanism.

Loss of paired receptors from one of the two homeologous X. laevis ceacam loci

At the time of hybridization of two ancestral X. laevis species some 40 million years ago [30] which led to speciation, two probably functionally distinguishable sets of ceacam genes existed. The set on chromosome 7 L was better functioning presumably with respect to pathogen resistance and was consequently retained [39]. Loss of the ceacam locus on chromosome 7S is not complete. Interestingly, in both subgroups no paired receptor system persisted. Only one ITIM- and one ITAM-bearing Ceacam is found in group 2 and group 1, respectively (Fig. 6b). This indicates that once the genes for inhibitory or activating members have been lost the remaining gene cluster degrades rapidly probably due to lack of selection pressure exerted by pathogens which helps to maintain paired receptors.

Conclusions and Perspectives

Although we do not know which specific pathogens bind to to Ceacam receptors in Xenopus the presence of closely related paired receptors as well as selection for diversification suggests that also in amphibians CEACAM1-related inhibitory proteins still are or have been exploited in the past as pathogen receptors and similar defense strategies have been developed in amphibians and mammals by convergent evolution. Thus the CEACAM family is a prototype gene family which offers a unique opportunity to study “arms races” caused by host/pathogen interactions.

CEACAM1-related pathogen receptors serve a dual role: They allow both entry into the host and inhibition of inflammatory responses to pathogen infections. Therefore, CEACAM1-like receptors are expected to be expressed on epithelial surfaces as well as on leukocytes involved in innate and adaptive immunity. Decoy receptors should be expressed on cells of the innate immune system which allow uptake and destruction of pathogens. The identification and characterization of individual members of the Xenopus ceacam gene families will now allow to use next generation sequencing data of RNA from multiple organs and cell types of Xenopus species as well as genomic sequence data of additional frog species like that of Nanorana parkeri a member of the species-rich Neobatrachia, which contains the vast majority of amphibian taxa (Sun et al., 2015) to support the suggested pathogen defence function of anuran ceacam families.

Methods

Identification and nomenclature of genes

Sequence similarity searches were performed using the NCBI BLAST tools (http://www.ncbi.nlm.nih.gov/BLAST) and the Ensembl BLAST/BLAT (http://www.ensembl.org/Xenopus_tropicalis/Tools/Blast?db=core) and Xenbase BLAST (http://www.xenbase.org/genomes/blast.do) search programs. For identification of ceacam genes regions syntenic to mammalian CEACAM loci were analyzed for the presence of Ig domain-encoding genes. The following databases were used for ceacam gene identification and loci analyses: Xenbase X. laevis J-Strain 9.1 and X. tropicalis Nigerian 9.0 (http://www.xenbase.org/entry/) and Ensemble X. tropicalis JGI 4.2 (http://www.ensembl.org/Xenopus_tropicalis/Info/Index). Xenopus genomes were reprobed with exon sequences from newly discovered ceacam genes. For estimation of the number of ceacam genes present in a given species, distinct ceacam N domain exons with a sequence divergence > 1 % were counted. Multiple N exons with no annotated non-N exon in between were considered to belong to the same gene. Genes that contained stop codons within their N domain exons or lacked appropriate splice acceptor and donor sites were considered to represent pseudogenes. Genes were assigned to their respective ceacam subgroups 1 and 2 based on phylogenetic analyses. Due to their non-orthologous relationship with mammalian CEACAM genes the new ceacam genes were numbered independently as follows: X. tropicalis, group 1: ceacam301-ceacam321; X. tropicalis, group 2: ceacam350-ceacam375; X. laevis, group 1; ceacam325-ceacam342; X. laevis, group 2: ceacam376-ceacam390. Nucleotide sequences from the N domain exons can be used as gene identifier (Additional files 2 and 3). Gene Nomenclature Guidelines recommended by Xenopus Gene Nomenclature Committee (2013) was followed (http://www.xenbase.org/gene/static/geneNomenclature.jsp).

Sequence motif identification and 3D modeling

The presence of ITAM, ITAM-like and ITIM/ITSM motifs were confirmed using the amino acid sequence pattern search program ELM (http://elm.eu.org/). Transmembrane regions, glycosylphosphatidylinositol (GPI) signal domains and leader peptide sequences were identified using the TMHMM (http://www.cbs.dtu.dk/services/TMHMM-2.0/), the big-PI predictor (http://mendel.imp.ac.at/sat/gpi/gpi_server.html), GPI-SOM (http://gpi.unibe.ch/) and the SignalP 4.1 programs (http://www.cbs.dtu.dk/services/SignalP/), respectively [40]. For three-dimensional modeling the geno3D software was used (https://geno3d-prabi.ibcp.fr/cgi-bin/geno3d_automat.pl?page=/GENOHLP/genohlp_help2.html). Images were constructed with the Swiss-PdbViewer software 4.1.

Phylogenetic analyses and determination positive selection

Phylogenetic analyses based on amino acid sequences were performed with the MEGA6 package [41]. The applied Maximum Likelihood method is based on the JTT matrix-based model [42]. The trees with the highest log likelihood are depicted. The percentage of trees in which the protein sequences clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. All positions containing gaps and missing data were eliminated. In order to determine the selective pressure on the maintenance of the nucleotide sequences, the number of nonsynonymous nucleotide substitution per nonsynonymous site (dN) and the number of synonymous substitutions per synonymous site (dS) were determined for N domain exons. The dN/dS ratios as well as the cumulative synonymous and nonsynonymous substitutions along coding regions of N domain exons from paralogous and orthologous genes were calculated after manual editing of sequence gaps or insertions guided by the amino acid sequences using the SNAP program (Synonymous Nonsynonymous Analysis Program; http://www.hiv.lanl.gov/content/sequence/SNAP/SNAP.html). The program PipMaker (http://bio.cse.psu.edu/) was used to identify conserved contiguous stretches of nucleotides between gene pairs and to calculate the degree of identity which is summarized as a ‘percent identity plot’ [43]. Multiple amino acid and nucleotide sequence alignments were performed with ClustalW programs (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_clustalw.html).