Origin and evolution of TRIM proteins: new insights from the complete TRIM repertoire of zebrafish and pufferfish

Abstract

Tripartite motif proteins (TRIM) constitute a large family of proteins containing a RING-Bbox-Coiled Coil motif followed by different C-terminal domains. Involved in ubiquitination, TRIM proteins participate in many cellular processes including antiviral immunity. The TRIM family is ancient and has been greatly diversified in vertebrates and especially in fish. We analyzed the complete sets of trim genes of the large zebrafish genome and of the compact pufferfish genome. Both contain three large multigene subsets--adding the hsl5/trim35-like genes (hltr) to the ftr and the btr that we previously described--all containing a B30.2 domain that evolved under positive selection. These subsets are conserved among teleosts. By contrast, most human trim genes of the other classes have only one or two orthologues in fish. Loss or gain of C-terminal exons generated proteins with different domain organizations; either by the deletion of the ancestral domain or, remarkably, by the acquisition of a new C-terminal domain. Our survey of fish trim genes in fish identifies subsets with different evolutionary dynamics. trims encoding RBCC-B30.2 proteins show the same evolutionary trends in fish and tetrapods: they evolve fast, often under positive selection, and they duplicate to create multigenic families. We could identify new combinations of domains, which epitomize how new trim classes appear by domain insertion or exon shuffling. Notably, we found that a cyclophilin-A domain replaces the B30.2 domain of a zebrafish fintrim gene, as reported in the macaque and owl monkey antiretroviral TRIM5α. Finally, trim genes encoding RBCC-B30.2 proteins are preferentially located in the vicinity of MHC or MHC gene paralogues, which suggests that such trim genes may have been part of the ancestral MHC.

Figure 1. TRIM proteins from zebrafish and pufferfish.

Classification of fish TRIM proteins based on their C-terminal domain(s) and the categories proposed by Short & Cox in Ref. 3. Previously unreported TRIM proteins found in fish were tentatively numbered TRIM101-111. Conserved TRIM proteins are represented on the left panel (Classes I–III and V–IX). Other TRIM proteins are shown on the right panel (Class IV). Dotted lines delimit groups of closely related human TRIM (modified from Ref. 6) corresponding to a diversification that occurred during tetrapod evolution. Blue frames indicate multigenic families observed in teleost fish. RBCC: Ring-BBox-Coiled Coil; COS: C-terminal subgroup one signature; FN3: Fibronectin, type III; B30.2: PRY/SPRY domain; PHD: Plant Homeo Domain; NHL: NCL-1, HT2A and Lin-41 repeat; Filamin: named from the protein Filamin; Bromo: acetylated lysine binding domain; ARF/SAR: from ARF and SAR GTP binding proteins; Pyrin: a member of the six-helix bundle death domain-fold superfamily; TM: transmembrane; Math: meprin and TRAF homology domain; Chromo: CHRromatin Organization Modifier domain.

Figure 2. Expression profile of selected trim genes.

The expression of 6 zebrafish trim genes was measured by quantitative RT-PCR in pools of tissues from 10–12 animals. E1f-α was used as a housekeeping gene, and the relative expression levels of trim genes were normalized on the geometric mean of the values measured for “whole males” and “whole females”, to take both sexes into account in the normalization. The data are represented as a heat map, with expression level and standard deviation is indicated for each condition.

Figure 3. Three different ways to generate TRIM proteins with new domains.

For these three cases observed in the zebrafish genome, the new gene is shown on the bottom of panel, and is compared with a closely related, typical member of the ftr or btr family illustrated on top. Percentages of identity refer to DNA sequences. Rectangles represent exons, numbers refer to nucleotides of coding sequence, stop included. Introns not to scale. A: the ftr06 gene, contained within a large cluster of tandem ftr genes on chromosome 2, encodes a protein with a C-terminal chromodomain instead of a B30.2. This is due to the insertion of a single chromodomain-encoding exon just upstream of the usual exon 6. The previous B30.2 exon, shown in parenthesis, is still present downstream, nonmutated, but is not included in the chromodomain-encoding transcript. B: the ftr52 gene, isolated on chromosome 9, encodes for a TRIM protein with a C-terminal RanBD/cyclophilin A domain instead of a B30.2. In this case, the new C-terminal domain is encoded by multiple exons; no B30.2-encoding sequence can be detected in this genomic area. C. The btr31 gene, located on chromosome 19 tandemly to its close relative btr32, encodes for a protein with the typical N- and C- ends of bloodthirsty-like proteins, but the B-boxes and the coiled-coil regions in the middle have been replaced by a transmembrane domain.

Figure 4. Genomic location of zebrafish trim genes.

trim genes are depicted in different colors according to each trim class: class I in red, class II in yellow and boxed, class IV in blue, class V in green, class VI in pink, class VII in black, class VIII in yellow and red, class IX in orange. The different trim subsets belonging to class IV are indicated in shades of blue. This representation is based on the Zv8 assembly. RBCC: Ring-Bbox-Coiled Coil; COS: C-terminal subgroup one signature; FN3: Fibronectin, type III; B30.2: PRY/SPRY domain; PHD: Plant Homeo Domain; NHL: NCL-1, HT2A and Lin-41 repeat; Filamin: named from the protein Filamin; Bromo: acetylated lysine binding domain; ARF/SAR: from ARF and SAR GTP binding proteins; Pyrin: a member of the six-helix bundle death domain-fold superfamily; TM: transmembrane; Math: meprin and TRAF homology domain; Chromo: CHRromatin Organization Modifier domain.

Figure 5. Group of conserved synteny around btr 01, 02 and 33.

Synteny groups were determined from Ensembl assemblies using the genomicus database and browser (http://www.dyogen.ens.fr/genomicus-59.01/cgi-bin/search.pl) . The figure is edited from the PhyloView taking btr 01 (A), 02 (B) and 33 (C) genes as references. The reference btr gene and its orthologues are shown in light green over a thin vertical line, and are indicated in bold.

Figure 6. Fish counterparts of trim35 constitute multigenic subsets.

(A) Distance tree produced by ClustalW (Neighbor joining; boostrap = 1000) for the zebrafish TRIM35/HLTR sequences and representative TRIM sequences from other species. Relevant boostrap values are indicated. Separate phylogenetic analyses of the RBB (B) and B30.2 (C) regions of TRIM35 and other representative TRIM using Clustalw (Neighbor joining; boostrap = 1000). The same analyses were performed with PHYML and led to consistent trees. Sequences integrated into the trees: DareBty: zebrafish bloodthirsty (NP_001018311); DareFtr: zebrafish fintrim (XM_692536); GaacFtr: stickleback fintrim; OrlaFtr: medaka fintrim (ENSORLP00000003320); OnmyFtr: rainbow trout fintrim (AM887799); DareTr25: zebrafish trim25 (NP_956469); SasaTr25: salmon trim25 (gene index TC35355 accessible at http://compbio.dfci.harvard.edu/); GagaTr25: chicken trim25 (XP_415653); XetrTr25: Xenopus tropicalis Trim25 (Ensembl Xenopus genome scaffold255: 821309_819660); HosaTr25: human Trim25 (Q14258); GagaTr35 : chicken trim35 (ENSGALP00000026735); AncaTr35: lizard trim35 (ENSACAP00000002320); HosaTr35: human trim35 (NP_741983.2); MumuTr35: mouse trim35 (ENSMUSP00000022623); GaacTr35: stickleback TRIM35 (ENSGACP00000004694); OrlaTr35: medaka Trim35; TeniTr35: pufferfish Trim35; dareTr35: zebrafish Trim35-8 (ENSDARP00000064945); HosaTr21: human Trim 21 (NP_003132); HosaTr11: human Trim 11 (NP_660215); HosaTr60: human Trim 60 (AAI00986). The IDs of the other TRIM35 sequences from zebrafish used in (A) are available in Figure S1.

Figure 7. Group of conserved synteny around trim35 genes and gene clusters.

Synteny groups were determined from Ensembl assemblies using genomicus database and browser (http://www.dyogen.ens.fr/genomicus-59.01/cgi-bin/search.pl) . The figure is edited from the PhyloView taking the human trim35 (ENSG000000104228) as a reference (A), or taking the stickleback trim35-01, trim35-02, trim35-04, trim35-10, trim35-14, trim35-24, trim35-26 and trim35-27 genes as references (B). The reference gene and its orthologues is shown in light green over a thin vertical line and is indicated with its Ensembl ID.

Figure 8. Positive selection in the B30.2 domain of BTR and TRIM35/HLTR.

Distribution of hypervariable and positively selected residues in a multiple alignment of B30.2 domains from representative zebrafish BTR and TRIM35/HLTR, compared with a typical FTR sequence (Dareftr13: [GenBank: XM_695031]), and with TRIM5α (HosaTRIM5a). Hypervariable sites (shannon entropy >2) are indicated in red. Hypervariable sites previously described are indicated in pink in the FTR13 sequence. The four hypervariable regions of the TRIM5α B30.2 domain are underlined. The variable loop-connecting strands of the domain are named VL1–VL6. ß-strands of the B30.2 domain are indicated by dark (PRY region) or light (SPRY region) blue arrows from . Segments 1, 2 and 3 determined by the recombination GARD analysis in the BTR multiple alignment are shown under the BTR01 sequence. Positively selected sites (among zebrafish TRIM35/HLTR & BTRs: this study; among FTRs: and among primate TRIM5α: are boxed in blue when detected under models 2a and 8. Sites positive under M8 but not under M2 are boxed in green. In TRIM35/HLTR, Q (ß-strand 1) was detected under M2 not M8. In BTR, R (ß-strand 1) has been detected in the BTR analysis only under M8 with complete domain, not in segment 1. In BTR, S (ß-strand 12) has been detected only under M2a and M8 of BTR segment 3, not in the analysis using the complete domain. The detailed PAML results for each position under positive selection are available in Additional data file 8.