Translational machinery of the chaetognath Spadella cephaloptera: a transcriptomic approach to the analysis of cytosolic ribosomal protein genes and their expression

Abstract

Background: Chaetognaths, or arrow worms, are small marine, bilaterally symmetrical metazoans. The objective of this study was to analyse ribosomal protein (RP) coding sequences from a published collection of expressed sequence tags (ESTs) from a chaetognath (Spadella cephaloptera) and to use them in phylogenetic studies.

Results: This analysis has allowed us to determine the complete primary structures of 23 out of 32 RPs from the small ribosomal subunit (SSU) and 32 out of 47 RPs from the large ribosomal subunit (LSU). Ten proteins are partially determined and 14 proteins are missing. Phylogenetic analyses of concatenated RPs from six animals (chaetognath, echinoderm, mammalian, insect, mollusc and sponge) and one fungal taxa do not resolve the chaetognath phylogenetic position, although each mega-sequence comprises approximately 5,000 amino acid residues. This is probably due to the extremely biased base composition and to the high evolutionary rates in chaetognaths. However, the analysis of chaetognath RP genes revealed three unique features in the animal Kingdom. First, whereas generally in animals one RP appeared to have a single type of mRNA, two or more genes are generally transcribed for one RP type in chaetognath. Second, cDNAs with complete 5'-ends encoding a given protein sequence can be divided in two sub-groups according to a short region in their 5'-ends: two novel and highly conserved elements have been identified (5'-TAATTGAGTAGTTT-3' and 5'-TATTAAGTACTAC-3') which could correspond to different transcription factor binding sites on paralog RP genes. And, third, the overall number of deduced paralogous RPs is very high compared to those published for other animals.

Conclusion: These results suggest that in chaetognaths the deleterious effects of the presence of paralogous RPs, such as apoptosis or cancer are avoided, and also that in each protein family, some of the members could have tissue-specific and extra-ribosomal functions. These results are congruent with the hypotheses of an allopolyploid origin of this phylum and of a ribosome heterogeneity.

Figure 1

Frequency of clones sequenced for each cDNA type. A: SSU RP cDNAs, B: LSU RP cDNAs. It is also indicated when isoforms (is.) have been found and when all the sequences contain frameshift(s) (FS). The characteristics of the 5'-end have been indicated for each cDNA type, TTT potential binding site(s) in black, TAT potential binding site(s) in white and partial sequences which do not contain the 5'-end in grey.

Figure 2

Alignment of the two consensus 5'-ends of the S. cephaloptera ribosomal protein genes. The 28 nucleotides region named n°2 has been found in 249 ESTs putatively encoding for 28 different SSU RPs, and in 319 ESTs putatively encoding for 37 different LSU RPs. The 28 nucleotides region of the sequence n°1 has been found in 75 ESTs putatively encoding for 25 different SSU RPs and in 96 ESTs putatively encoding for 29 different LSU RPs. The stars (*) indicate nucleotides which are conserved between these two sequences. The nucleotide regions which differ between these two sequences have been underlined and have named respectively TAC consensus site and TTT consensus site. The nucleotides which are conserved between these two consensus sites and Tinman/Nkx2.5 binding consensus sites are indicated in bold letters; K representing T or G.

Figure 3

Phylogenetic trees of the selected ribosomal proteins sequences (see Table 2 and Methods). The trees A, B and C were obtained using respectively Neighbor Joining (NJ), Fitch, Maximum Parsimony (MP), and Maximum Likehood (ML) methods on an amino acid dataset. The trees constructed using Fitch and MP methods have a similar topology. In D, the ML tree using the first two codon positions and the model selected by MrAIC, GTRIG, ML estimated base frequency, a gamma (2) distribution for site substitution rates, and an estimated proportion of invariant sites. Similar topologies were obtained with ML using codon models and with a non homogeneous non stationary ML method allowing G+C equilibrium frequency to vary (see text). Trees E and F were obtained using respectively the GTR model with a MCMC bayesian method and the CAT mixture model on an amino acid dataset. Numbers indicate bootstrap values or branch support; in tree B, MP and Fitch values are respectively at the left and at the right, in tree D, after the slash, the aLRT (actually the minimum of the CHI2-based parametric and non parametric aLRT estimated value). Abbreviations: D.m., D. melanogaster; Echino., Echinoderm; R.n., R. norvegicus; S.c., S. cephaloptera; S.d., S. domuncula; Yeast, S. cerevisiae.