Pervasive Mitochondrial tRNA Gene Loss in Clade B of Haplosclerid Sponges (Porifera, Demospongiae)

Abstract

Mitochondrial tRNA gene loss and cytosolic tRNA import are two common phenomena in mitochondrial biology, but their importance is often under-appreciated in animals. This is because the mitochondrial DNA (mtDNA) of most bilaterally symmetrical animals (Bilateria) encodes a complete set of tRNAs required for mitochondrial translation. By contrast, the mtDNA of nonbilaterian animals (phyla Cnidaria, Ctenophora, Porifera, and Placozoa) often contains a reduced set of tRNA genes, necessitating tRNA import from the cytosol. Interestingly, in many nonbilaterian lineages, tRNA gene content appears to be set early in evolution and remains conserved thereafter. Here, we report that Clade B of haplosclerid sponges (CBHS) represents an exception to this pattern, displaying considerable variation in tRNA gene content even among relatively closely related species. We determined mt-genome sequences for eight CBHS species and analyzed them in conjunction with six previously available sequences. Additionally, we sequenced mt-genomes for two species of haplosclerid sponges outside the CBHS and used them with eight previously available sequences as outgroups. We found that tRNA gene content varied widely within CBHS, ranging from three in an undescribed Haliclona species (Haliclona sp. TLT785) to 25 in Xestospongia muta and X. testudinaria. Furthermore, we found that all CBHS species outside the genus Xestospongia lacked the atp9 gene, with some also lacking atp8. Analysis of nuclear sequences from Niphates digitalis revealed that both atp8 and atp9 had transferred to the nuclear genome, while the absence of mt-tRNA genes indicated their genuine loss. We argue that CBHS can serve as a valuable system for studying mt-tRNA gene loss, mitochondrial import of cytosolic tRNAs, and the impact of these processes on mitochondrial evolution.

Keywords: Haplosclerida; comparative genomics; mitochondrial DNA; tRNA import; tRNA loss.

Fig. 1.

Mitochondrial gene content and gene arrangement in the CBHS. Species are grouped by clades to which they belong (B1 to B5). Protein and rRNA genes (larger boxes) are: atp6, 8-9—subunits 6, 8, and 9 of the F0 ATPase, cox1-3—cytochrome c oxidase subunits 1 to 3, cob—apocytochrome b, nad1 to 6, and nad4L—NADH dehydrogenase subunits 1 to 6 and 4L, rns and rnl—small and large subunit rRNAs. tRNA genes (smaller boxes) are abbreviated using the one-letter amino acid code. The two arginine, isoleucine, leucine, and serine tRNA genes are differentiated by numbers with trnR(ucg) marked as R4, trnR(ucu)—as R2, trnI(gau)—as I2, trnI(cau)—as I1, trnL(uag)—as L4, trnL(uaa) as L2, trnS(ucu)—as S2, and trnS(uga)—as S4. The second trnM(cau) in Xestospongia species is labeled as Me, while the unusual trnI(aau) in the B2 clade as I?. All genes are transcribed from left to right. Genes are not drawn to scale and intergenic regions are not shown. Red asterisks indicate tRNA genes inferred to evolve by gene remolding. Black asterisks indicate genes that were not identified by tRNAscan-SE2 using the Infernal search. They were included in the figure either for consistency with previous studies (trnN(guu) and trnF(gaa) in A. queenslandica or because of high sequence conservation with closely related species (see below).

Fig. 2.

Phylogenetic relationships of haplosclerid sponges with inferred tRNA gene loss and gain. Posterior majority-rule consensus tree obtained from the analysis of concatenated mitochondrial amino acid sequences (3,696 positions) under the CAT+GTR+Γ model with the PhyloBayes-MPI program. The number at each node represents the Bayesian posterior probability. The maximum-likelihood analysis with the LG+G8+F model selected an identical topology (supplementary Fig. S1, Supplementary Material online). Five major clades in CBHS are shown as B1 to B5. We used the Dollo parsimony principle, which assumes the irreversible loss of characters, to manually map gene loss on the phylogenetic tree (in red). The results were adjusted based on the analysis of tRNA gene phylogenetic relationships, which revealed changes in some anticodon identities (in blue). Names of tRNA genes are abbreviated using the one-letter amino acid code.

Fig. 3.

Distinct patterns of mt-tRNA gene evolution within the CBHS. a) Accumulation of mutations leading to losses of mt-tRNA genes within the B4 clade. Nucleotides in red differ from those in Haliclona simulans. b) Conservation of unusual tRNA structures within the B3 clade. tRNA sequences are shown for G. conulosa. Nucleotides in red showed variation among B3 species. Nucleotides in lowercase were not inferred to be part of tRNA by tRNAscan and are shown here as would be expected under the normal cloverleaf structure. c) tRNA remolding in the B2 clade. Remolded tRNA names are in red. Nucleotides in red show differences from corresponding positions in Haliclona loe. Number under tRNA names signify covariance score, except when marked with an asterisk. Numbers marked with asterisks signify COVE scores.

Fig. 4.

Distribution of tRNA scores in CBHS. Covariance model bit scores for inferred mt-tRNAs in CBHS species are drawn arranged by species (a) and by tRNA (b). The color of the points denotes species’ placement within individual clades of CBHS. The box is drawn from the 25th to the 75th percentile with a horizontal line denoting the median.

Fig. 5.

Phylogenetic tree of mt-tRNA genes within the CBHS. Neighbor-joining tree based on uncorrected (P) distances is shown. Clusters containing all and only tRNAs for the same codon family were cartooned. When tRNAs from individual species are shown, species names are abbreviated as following: aquee, Amphimedon queenslandica; gconu, Gelliodes conulosa; hcaer, Haliclona caerulea; hloe, H. loe; hpahu, H. pahua; hplak, H. plakophila; hsimu, H. simulans; nsigm, Neopetrosia sigmafera; tlt723, Haliclona sp. (TLT723); tlt785, Haliclona sp. (TLT785); xmuta, Xestospongia muta; xtest, X. testudinaria. Numbers above selected branches indicate bootstrap support percentage based on 1,000 bootstrap replicates. Circle numbers correspond to listed cases of unexpected phylogenetic results described in the text.

Fig. 6.

Within-group correspondence analysis of synonymous codon usage in mitochondrial coding sequences of CBHS. The first axis explains 19.8% of the variance and the second axis explains 8.6%. Both coding sequence’s (points) and codon’s (rectangles) positions are shown on the first factorial map. Points are colored based on the subclade to which a particular species belongs.