Introduction

The mitochondrial (mt) genomes of almost all bilaterian animals comprise a single circular molecule of 14–24 kb that contains the genes for 13 energy pathway proteins and the two RNA components of the mt-ribosome, as well as 22 transfer RNAs (tRNAs) that appear to be sufficient to decode the mt-protein genes (for references, see Barrell et al. 1979, 1980; Anderson et al. 1981; Wolstenholme 1992; Wolstenholme and Fauron 1995; Boore 1999; Gissi et al. 2008; Bernt et al. 2012). A similar mt-genome in regard to size, conformation, and protein gene and rRNA gene content was found to occur in the sea anemone Metridium senile (subclass Hexacorallia, Anthozoa, Cnidaria) (Beagley et al. 1996, 1998). However, a notable difference was that this genome encodes only two transfer RNAs (tRNAf-Met and tRNATrp). This has since been shown to be a common feature of the mt-genomes of hexacorallian mtDNAs, and that mt-genomes of the anthozoan subclass Octocorallia encode only tRNAf-Met (Beagley et al. 1995; Beaton et al. 1998; Lavrov 2007). Although among bilateria, such a large deficiency in mt-tRNA genes has only been recorded for a chaetognathan (Helfenbein et al. 2004), mt-genomes lacking various numbers of tRNA genes have been recorded for a number of organisms including the protozoans, Tetrahymena pyriformis, Leishmania tarentolae, and Trypanosoma brucei, some sponges and angiosperm plants, and in some cases evidence for importation of nuclear DNA-encoded tRNAs into mitochondria has been obtained (Marechal-Drouard et al. 1988, 1990; Simpson et al. 1989; Hancock and Hajduk 1990; Burger et al. 1995; Tarassov et al. 2007; Lavrov 2007, 2011; Adshya 2008; Alfonzo and Soll 2009; Duchene et al. 2009). However, experimental data regarding the involvement of mtDNA-encoded tRNAs and nuclear DNA-encoded tRNAs in cnidarian mt-protein synthesis is lacking. Therefore, we undertook the presently reported study to elucidate the likely involvement of tRNAs transcribed from the two mt-tRNA genes and from nuclear tRNA genes in M. senile mt-protein synthesis.

In addition to a reduction in tRNA genes, the mt-genomes of bilateria exhibit a number of unusual features that include extreme compaction of the component genes, genetic codes that have up to five changes in codon specificity relative to the standard (universal) genetic code, unorthodox translation initiation codons, and replication by a unique asymmetrical mode (for references, see Wolstenholme 1992; Beagley et al. 1998; Brown et al. 2005; Watanabe and Yokobori 2011; Bernt et al. 2012; Abascal et al. 2012). A further, notable general feature of bilaterian mt-genomes is the considerable diversity of structure of the encoded tRNAs relative to standard tRNAs (i.e., including tRNAs encoded in DNAs of prokaryotes, eukaryotic nuclei, plant chloroplasts and plant mitochondria; Sprinzl et al. 2005; Juhling et al. 2009). Most bilaterian mt-tRNAs have a reduced dihydrouridine loop that lacks the usually conserved G18–G19 pair and a TψC loop in which the usually conserved 5′-G53T54T55C56R57A58N59Y60C61 sequence is either reduced or totally lacking (Wolstenholme 1992; Juhling et al. 2009). In all bilaterian mt-tRNAs that recognize AGN (Serine) codons, and occasionally mt-tRNAs that recognize UCN (serine) codons, the dihydrouridine arm is replaced by a simple loop of between 2 and 13 nucleotides (De Bruijn et al. 1980; Wolstenholme 1992; Beagley et al. 1999). A more extreme situation is that in all tRNAs encoded in mtDNAs of the nematode worms Ascaris suum and Caenorhabditis elegans and other closely related nematodes, with the exception of those tRNAs recognizing serine-specifying codons, the TψC arm is replaced with a simple loop of between 6 and 12 nucleotides (Wolstenholme et al. 1987, 1994; Okimoto et al. 1990; Montiel et al. 2006). Further, sequences in mtDNA molecules of species of enoplean nematodes have been interpreted as encoding some tRNAs in which the dihydrouridine and TψC arm are each replaced with a simple loop of nucleotides (Juhling et al. 2009). Interestingly, primary and predicted secondary structures of the tRNAs encoded in cnidarian mtDNAs more closely resemble standard tRNAs rather than any of the sorts of reduced mt-tRNAs found in bilateria (Pont-Kingdon et al. 1994, 1998, 2000; Beagley et al. 1995, 1998).

All mature tRNAs examined, regardless of their genetic origin, posses a 3′-terminal CCA which is required for amino acid addition (Tamura et al. 1994; Liu and Horowitz 1994; Nagaike et al. 2001), for base pairing interaction of the tRNA with the ribosome (Tamura 1994; Samaha et al. 1995), and for deacetylated tRNAs to leave the ribosome (Lill et al. 1988). In many prokaryotes and some archaea, the 3′-terminal CCA is encoded in all or some tRNA genes (Marck and Grosjean 2002). In contrast, neither mt-tRNA genes nor nuclear DNA-encoded tRNA genes of metazoa have CCA at their 3′ ends; rather CCA is added post-transcriptionally by nucleotidyl transferases (Deutscher 1990; Martin 1995; Tomita and Weiner 2001).

Although the mt-genome of M. senile contains the same circularly arranged, compact gene complement as that of most bilateria, besides the deficiency of tRNA genes, this genome was found to have a further unusual feature (Beagley et al. 1996, 1998). Each of two protein genes, cox1 and nad5 contain a group 1 intron. The cox1 intron encodes a putative homing endonuclease, and the nad5 intron contains the genome’s only copy of two other protein genes, nad1 and nad3. An nad5 intron alone, or both a cox1 intron and an nad5 intron have since been located in other hexacorallian mtDNAs but not in the mtDNAs of other Cnidaria (Beagley et al. 1996; Lavrov 2007). Also, with the exception of rare TGA codons that appear to specify tryptophan (as in all other metazoan mt-genetic codes), the M. senile mt-protein genes utilize a standard genetic code.

That there is, in fact, a transcription–translation system in M. senile mitochondria that results in the generation of functional mtDNA-encoded proteins is fully supported by the processing of primary transcripts to eliminate introns and provide complete and uninterrupted transcripts of the nad5, nad1 and nad3 genes (Beagley et al. 1996), by uninterrupted open reading frames in all other mt-protein genes, and by the occurrence of genes for rRNAs that have the potential to fold into secondary structures consistent with other prokaryotic and metazoan mt-rRNAs (Pont-Kingdon et al. 1994; Beagley et al. 1996, 1998). Further support for this contention has been provided by bioinformatic evidence indicating the occurrence in a hexacoralian cnidarian of a nuclear DNA-encoded aminoacyl-tRNA synthetase specific to the mtDNA-encoded tRNATrp and nuclear DNA-encoded aminoacyl-tRNA synthetases that include mitochondrion-targeting sequences (Haen et al. 2010).

Because the M. senile mt-genome contains only two tRNA genes, we considered it worthwhile to seek direct evidence as to whether one or both are, in fact, transcribed and processed into tRNAs that could function in protein synthesis. We provide experimental data indicating that in M. senile mitochondria, there occur tRNAs corresponding in sequence to the mt-tRNAf-Met and mt-tRNATrp genes that at their 3′ ends have post-transcriptionally added nucleotides expected for them to be functional in protein synthesis. Further, we provide evidence that mt-tRNAf-Met is confined to mitochondria but that nuclear-encoded tRNAVal and tRNAi-Met are located in both the cytoplasm and mitochondria indicating them to be needed for both cytoplasmic and mt-protein synthesis.

Materials and methods

Preparation of RNA fractions

Preparation of mitochondria from M. senile and isolation of RNA from mitochondria (mt-RNA) followed previously described (Pont-Kingdon et al. 1994). Isolation of whole cell RNA (wc-RNA) was achieved using the REX™ total RNA extraction kit (Amersham Life Sciences). RNA was also isolated from a cytoplasmic fraction from which mitochondria had been removed as follows. A single (3 cm diameter) M. senile was homogenized in a Waring blender at high speed for 5 s in ice cold 10 % sucrose TE. The homogenate was immediately centrifuged for 10 min at 10,000g to sediment any undisrupted cells, and mitochondria. The supernatant was extracted three times with phenol–chloroform–isoamyl alcohol (PCI; 50:50:1) and the final aqueous phase was precipitated with 2.5 volumes of 95 % ethanol.

RNA preparations were incubated in 0.5 M Tris, pH 8.7 for 30 min at 37 °C to discharge amino acids from tRNAs (Norma Wells, personal communication). The RNA preparations were then extracted once with PCI and the aqueous phases were passed through a Centricon 100 microconcentrator (Amicon). RNA molecules in the flow-through were concentrated with a Centricon 30 microconcentrator and distribution of sizes of molecules in each retentate was determined by electrophoresis through a denaturing (15 % polyacrylamide, 8 M urea) gel (Fig. 1). RNA molecules in the microconcentrator-selected fraction ranged in size from approximately 40 to 160 nt, and the predominant size classes are similar to those of Escherichia coli tRNAs (Fig. 1).

Fig. 1
figure 1

Size profiles of RNA molecules isolated from a mitochondrial (mt) fraction and from whole cells (wc) of Metridium senile. An ethidium bromide-stained denaturing (15 % polyacrylamide, 8 M urea) gel following electrophoresis of RNA molecules is shown. Lanes mt and wc contain 1 μg each of RNA molecules isolated from a mt-fraction and from whole cells, respectively, which in both cases were selected by passage through a Centricon 100 microconcentrator (Amicon) and concentrated with a Centricon 30 microconcentrator. Lane E. coli uf contains 1 μg of unfractionated Escherichia coli tRNAs (Sigma) and Lane E. coli Tyr contains 100 ng of E. coli tRNATyr (Sigma) with an expected size of 85 nt (Sprinzl et al. 2005)

Full size image

Determination of mt-tRNA 3′-end sequences

3′ Rapid amplification of cDNA ends (RACE); (Frohman 1990) analyses were carried out to determine the 3′ ends of M. senile mtDNA-encoded tRNAf-Met and tRNATrp, following the procedures described previously (Pont-Kingdon et al. 1994; Beagley et al. 1996, 1998). RNA molecules in the 40–160 nt mt-RNA preparation were polyadenylated in a 30-μl reaction containing 1 μg RNA, 20 U RNasin (Promega), 0.25 mM ATP, 1,500 U cloned yeast Poly-A polymerase (Amersham Life Sciences) and the suppliers reaction buffer, at 30 °C for 30 min. Following a second PCI extraction, the polyadenylated RNA was desalted and concentrated with a Centricon 30 microconcentrator, then reverse transcribed in a 25-μl reaction containing 20 U AMV reverse transcriptase (Amersham Life Sciences), 20 U RNasin, 0.5 mM each of dATP, dGTP, dCTP and dTTP and 0.5 pmol primer TB17-1 [an oligonucleotide containing BamHI, BglII and PstI restriction sites (underlined):

5′-CCAGATCTGGATCCTGCAGTTTTTTTTTTTTTTTT (Beagley et al. 1996)] in the supplied reverse transcriptase reaction buffer for 5 min at 25 °C, and then for 10 min at 45 °C. The reaction mixture was then extracted with PCI, concentrated with a Centricon 30, and the retentate (designated the 3′-RACE pool) was brought to 100 μl with TE and stored at 4 °C.

The 3′-end region of each mt-tRNA sequence was amplified by polymerase chain reaction (PCR) in a 100-μl reaction containing 2 μl of the 3′-RACE pool, 0.2 mM each of dATP, dGTP, dCTP and dTTP, 2.5 U AmpliTaq™ DNA polymerase (Perkin Elmer) and 25 pmol each of a primer comprising the BamHI, BglII and PstI site-containing segment of TB17-1 (TB17-2: 5′-CCAGATCTGGATCCTGCAG) and a sense oligonucleotide primer: for the mt-tRNAf-Met, TMB-2 [5′-TGGGAATTCACC-AGACTCATGATCT (see Fig. 2)] which contains two mismatched nucleotides (underlined) to create an EcoRI restriction site is used for subsequent cloning of the double-stranded PCR product; for tRNATrp, TBW-2 (5′-AAGGAAGATCTTTAGCCTTCAAAGC) and TBW-3 (5′-AGCTAAAACTGCAGGTTCAAGT), each containing two mismatched nucleotides (underlined) to create a BglII restriction site and a PstI restriction site, respectively. Cycling was 45 s at 94 °C, 45 s at 60 °C, and 45 s at 72 °C. The double-stranded PCR products were sequenced by the dideoxynucleotide chain-terminating method (Sanger et al. 1977) using Sequenase™ (Amersham Life Sciences).

Fig. 2
figure 2

Northern blot and 3′-RACE analysis of M. senile mt-tRNAf-Met. a The M. senile mt-tRNAf-Met gene shown in the presumed secondary structural form of the corresponding tRNA (Pont-Kingdon et al. 1994; Beagley et al. 1998) to which the three nucleotides CCA (each in parenthesis) sequentially at the 3′ end and 16 As representing polyadenylation by yeast poly-A polymerase are added. The two solid lines TMB-1 and TMB-2 on the outside of the sequence identify segments to which oligonucleotide primers were made, with the arrowhead indicating the 5′–3′ direction. The solid line TB17-1 is a poly-A primer with a 3′-EcoRI restriction site (RS)-containing 19 nt sequence. The short solid line TB17-2 is a primer complementary to the 3′ EcoRI site-containing sequence of TB17-1. The solid dots in primer TMB-2 indicate changes to the tRNAf-Met gene sequence made to create an EcoRI site. b Northern blot hybridizations. Lanes 1 and 2 contain RNA isolated from a mitochondrial fraction (mt), and whole, unfractionated cells (wc), and lanes 3 and 4 contain purified E. coli tRNATyr, electrophoresed through a denaturing (15 % polyacrylamide, 8 M urea) gel. Lanes 1 and 2 were probed with 32P-labeled oligonucleotide TMB-1 and lanes 3 and 4 were probed with a 32P-labeled oligonucleotide that is complementary to nucleotides 46–66 or E. coli tRNATyr (Sprinzl et al. 2005). c Autoradiograph showing the product of sequencing a clone of a cDNA of the 3′-end proximal region of mt-tRNAf-Met that includes the post-transcriptionally added CC (and presumably A), shown to the right of the sequence, and the synthetically added poly(A) tail

Full size image

RACE amplification of nuclear DNA-encoded tRNA sequences

Fractions of small (40–160 nt) wc-RNAs were 3′ polyadenylated, and oligo-dT primed cDNAs were generated using AMV reverse transcriptase as above. The cDNAs were extracted with PCI, concentrated with a Centricon 30 microconcentrator, and the retentate was brought to 100 μl with TE to provide a 3′-RACE pool that was stored at 4 °C, and then used in the following amplification procedures.

The first procedure was designed to amplify nt 63–73, comprising the 3′ strands of the TψC and aminoacyl stems of previously unidentified nuclear DNA-encoded M. senile tRNAs. An EcoRI site-containing degenerate primer (TBTψC-1: 5′-CCGAATTCGGTTCRANYCC) was made, based on the highly conserved nt 52–62 tRNA sequences of standard tRNAs (Sprinzl et al. 2005). The reverse primer used in PCR amplification was TB17-3 (5′-GGAGATCTTTTTTTTTTTTTTTTTGG) that contains a 5′ BglII restriction site (underlined), 16 Ts and two 3′-terminal Gs to recognize the expected post-transcriptionally added 3′-terminal CCA. PCR amplification was carried out in a 100-μl reaction that included 1 μl of 3′ poly-A tailed cDNA, 25 pmol each of TBTψC-1 and TB-17-3, 0.2 mM each of dATP, dGTP, dCTP and dTTP and 2.5 U AmpliTaq™ DNA polymerase. For cycling, touchdown PCR (Don et al. 1991) was used as follows: 30 s at 94 °C, 30 s annealing temperature and 30 s at 72 °C with the annealing temperatures being 2 cycles at 48 °C, 2 cycles 46 °C, 2 cycles 44 °C, 2 cycles 42 °C, 2 cycles 40 °C, 2 cycles 38 °C then 25 cycles 37 °C. The double-stranded PCR products were digested with BglII and EcoRI, inserted into M13mp18, and resulting clones were sequenced.

For the second procedure, 5′-RACE analysis used a series of primers based on the 11 nt sequences obtained in the first procedure to amplify tRNA sequences that are 5′ to nt 63. Each of these primers [designated TBAA (amino acid)-1 to TBAA-N, Fig. 3] comprised a PstI site-containing sequence followed by a sequence complementary to one of the 11 nt sequences (tRNA sequence 63–73). Poly-C was added to the 3′ ends of cDNAs, from the 3′-RACE pool, using terminal deoxynucleotidyl transferase (Tdt, Amersham Life Science) in a 50-μl reaction containing 10 μl cDNA synthesized from poly-A tailed RNAs, 0.4 mM dCTP, and 30 U Tdt in the supplied reaction buffer and incubated for 5 min at 37 °C. Following extraction with PCI and Centricon 30 concentration, the retentate (designate the 5′-RACE pool) was made up to 100 μl with TE and stored at 4 °C. Amplification of specific 5′-extended sequences was achieved in 100 μl reactions containing 2 μl of the 5′-RACE pool, 0.2 mM each dATP, dCTP, dTTP and dGTP, 2.5 U AmpliTaq™ DNA polymerase (Perkin Elmer) and 25 pmol each of the tRNA-specific primer (TBAA-1 to TBAA-n) and primer TB17-4 (5′-CTGCGAATTCGGATCCTGTGGGGGGGG) which contains EcoRI and BamHI sites (underlined) and recognizes the poly-C tail added with Tdt. Cycling used the touchdown PCR method described above to account for the uncertainty regarding the optimum annealing temperature of primer TB17-4. The double-stranded amplification products were cleaved with EcoRI and PstI and cloned into EcoRI–PstI-digested M13mp19. This insertion strategy placed the shorter (about 20 nt) poly-C tailed sequence closest to the M13 universal sequencing primer, which facilitated sequencing (the poly-A tails average about 50 nt). The resulting cloned cDNAs were sequenced using Sequenase™ (Amersham Life Science).

Fig. 3
figure 3

Northern blot and 3′-RACE analysis of M. senile mt-tRNATrp. a The M. senile mt-tRNATrp gene shown in the presumed secondary structural form of the corresponding tRNA (nucleotides 1–70, Beagley et al. 1998) to which are added sequentially to the 3′ end the three nucleotides CCA (each in parenthesis) and 16 As representing polyadenylation with yeast poly(A) polymerase. The three solid lines TBW-1, TBW-2 and TBW-3 on the outside of the tRNA sequence identify segments to which oligonucleotide primers were made, with the arrowheads indicating the 5′–3′ direction. The solid line TB17-1 is a poly(A) primer with a 3′ EcoRI site-containing 19 nt (RS) segment. The short solid line TB17-2 is a primer complementary to the 3′ EcoRI site-containing sequence of TB17-1. The solid dots in primer TBW-2 indicate changes to the mt-tRNATrp gene sequence made to create a BglII site and the solid dots in primer TBW-3 indicate changes relative to the mt-tRNATrp gene sequence made to create a Pst1 site. b Northern blot hybridizations. Lanes 1 and 2 contain RNA isolated from a mitochondrial fraction (mt) and from whole unfractionated cells (wc), and lane 3 contains purified E. coli tRNATrp electrophoresed through a denaturing (15 % polyacrylamide, 8 M urea) gel. Lanes 1 and 2 were probed with 32P-labeled oligonucleotide TBW-1 and lane 3 was probed with a 32P-labeled oligonucleotide that is complementary to nucleotides 46–66 of E. coli tRNATyr. c Autoradiograph showing the products of sequencing of two clones of cDNAs of the 3′-end proximal regions of two mt-tRNATrp molecules. The sequence to the right (II) includes post-transcriptionally added CC (and presumably A) and the synthetically added poly(A) tail. The sequence to the left (I) also includes the synthetically added poly(A) tail but does not include post-transcriptionally added CC nucleotides

Full size image

A third procedure designed to amplify previously unidentified tRNA 3′-end sequences, that may have been missed using oligonucleotide TBTψC-1, was also used. In this experiment, the tRNA-specific primers (TBAA-1 to TBAA-n) were replaced in the 5′-RACE reactions by the non-specific primer TB17-3 (above). Reaction conditions and cycling were otherwise identical and this is designated a double-RACE amplification as both ends of the cDNA are unknown (except that TB17-3 selects for tRNA sequences by recognizing a 3′-terminal CCA). The double-stranded amplification products were cleaved with BglII and EcoRI and cloned into BglII–EcoRI-digested M13mp19 and sequenced as before.

Degenerate PCR amplification of the 3′-end sequence of M. senile cytoplasmic initiator tRNAi-Met

To determine the nucleotide sequence of the cytoplasmic (nuclear DNA-encoded) initiator methionine tRNA (tRNAi-Met) of M. senile, we used a degenerate oligonucleotide primer (TBiM-1: 5′-GCGCACYGGATCCGTGCTGGG; Fig. 4) which is a consensus based on the highly conserved nucleotides 10–31 of tRNAi-Met genes of human, mouse, Xenopus laevis and Saccharomyces cerevisiae (Fig. 4; Sprinzl et al. 2005). The two underlined nucleotides introduced mismatches that create a BamHI restriction site for cloning. The PCR volume was 100 μl and included 1 μl poly-A-tailed cDNA (above), 25 pmol each of TBiM-1 and TB17-3, 0.2 mM each dATP, dCTP, dTTP and dGTP and 2.5 U AmpliTaq™ DNA polymerase. Cycling for this reaction was 30 s at 94 °C, 30 s at 60 °C and 30 s at 72 °C for 30 cycles. The resulting double-stranded PCR product was digested with BglII and BamHI and cloned into BamHI-digested M13mp19 for sequencing.

Fig. 4
figure 4

Reverse transcriptase-PCR experiments carried out to obtain sequences of M. senile nuclear DNA-encoded tRNAs. A–D A nucleotide numbered standard tRNA (Sprinzl et al. 2005) is shown that has been 3′ polyadenylated with yeast poly-A polymerase. Anticodon nucleotides (#34, #35, #36) are shown as reverse numbers. The expected post-transcriptionally added CCA nucleotides on each tRNA are bracketed. In each experiment, cDNA synthesis was primed with poly-dT. Lines outside the tRNAs represent cDNAs (labeled) and oligonucleotide primers, the 5′–3′ directions of which are shown by arrowheads. The partial, specific tRNA sequences obtained from these experiments are shown to the right of each schematic tRNA, either as a linear sequence (A) or as sequences folded into presumed secondary structural configurations (B–D). A A scheme used to obtain sequences of nucleotides 63–73 of unknown tRNAs. Oligonucleotide primers used, shown as solid lines outside the sequence, are: TB17-3 a 5′-restriction site (RS)-containing sequence added to 16 Ts and GG; TBTψC-1, a 5′-restriction site-containing sequence added to an 11-nt degenerate sequence (5′-GGTTCPuANPyCC, shown outside the solid line) that matches the consensus sequence of the distal two nucleotide pairs of the TψC stem and the TψC loop of the majority of standard tRNAs (Sprinzl et al. 2005). Four of the 29 unique, 11 nt sequences obtained, that were used in further experiments are shown to the right (a–d, and labeled to the right with the identification of each determined in later experiments). B A scheme employing information gained in A to obtain tRNA sequences 5′ to nucleotide 63. The solid dots on the cDNA indicate positions at which cDNA synthesis was most frequently terminated. Poly-C was added to the 3′ ends of cDNAs using terminal nucleotidyl transferase. Primers TBAA-1-N are sequences comprising a 5′-restriction site (RS) and the 4 different 11 nt sequences obtained in A. TB17-4 is a 5′-restriction site-containing sequence (RS) followed by eight Gs. Two partial tRNA sequences obtained with anticodons identifying them as tRNAAsp and tRNAGlu are shown to the right. c A scheme (double RACE) to obtain tRNA sequences irrespective of whether the TψC arm sequence corresponds to the degenerate sequence shown in A. A partial tRNA sequence obtained with an anticodon indicating it to be tRNAGln is shown to the right. The arrow indicates an A52–T62 pair that would not be recognized by primer TBTψC-1 (see A). D An experiment designed to obtain the sequence of tRNAi-Met. Primer TBiM-1 is a consensus sequence based on nucleotides 10–31 of nuclear tRNAi-Met genes of four eukaryotes. The solid dots indicate two consensus sequence nucleotides (#21 and #22) that have been changed to create a BamHI site. The partial M. senile tRNAi-Met sequence obtained is shown to the right. The reverse lettered five nucleotides originate from the primer. Asterisks identify nucleotides (#33, #54, #60, #70, #71 and #72) that are characteristic of all known eukaryotic tRNAi-Mets, and arrows (nt 39–41) indicate 3 Cs at the base of the anticodon stem that are characteristic of tRNA initiator methionine and some formyl-methionine specifying tRNAs (RajBhandary and Chow 1995). A primer, TBiM-2, complementary to nt 54–71 of the partial cDNA-tRNAi-Met sequence was used to obtain the nuclear DNA-encoded gene sequence, shown to the far right

Full size image

Amplification of nuclear DNA-encoded tRNA genes

High molecular weight DNA was isolated from M. senile using protocol II of the Stratagene DNA extraction kit: Extraction from whole tissue. A DNA-based genomic library was created by digesting to completion 1 μg DNA with Sau3AI and ligating the resulting fragments into BamHI-digested M13mp19 double-stranded DNA (100 ng). The chimeric DNA was extracted once with PCI and desalted using a Centricon 100 microconcentrator. The retentate was made up to 100 μl with TE and the library was stored at 4 °C. From this library, DNA sequences were amplified using tRNA-specific oligonucleotide primers TBAA-N (Fig. 4B) to anchor one end of the PCR and the universal M13 sequencing primer (New England Biolabs) to anchor the other end. PCR included 1 μl of the library DNA, 0.2 mM each dATP, dCTP, dTTP and dGTP, 25 pmol each oligonucleotide primer and 2.5 U Amplitaq™ DNA polymerase. The sequence of the oligonucleotide primer used to amplify the tRNAIle DNA sequence (TB-Ile) was 5′-GGTCTAGATGGTCCGTATGGGGC (underlined nucleotides comprise an added XbaI site) and the annealing temperature was 40 °C. The sequence of the oligonucleotide primer used to amplify the tRNAVal DNA sequence (TB-Val) was 5′-GGTCTAGATGTTTCCGCCCGGGC and the annealing temperature was 40 °C. The sequence of the oligonucleotide primer used to amplify the cytoplasmic initiator tRNAi-Met DNA sequence (TB1 M-2) was 5′-GGTCTAGATGCAGAGCGAGGTTTCGATC [that includes an XbaI site (underlined)] and the annealing temperature was 50 °C. Amplification products were cloned into M13mp18 by cleaving with XbaI and EcoRI (amplified from the M13mp19 poly-linker cloning site) and sequenced.

Preparation of mechanically damaged mitochondria

M. senile mitochondria were purified by isopycnic sucrose (1.0–2.0 M) centrifugation (Rickwood et al. 1987). Half of the sucrose gradient-purified mitochondria were passed three times through an 18-gauge needle and these mitochondria and the remaining half of the sucrose gradient-purified mitochondria were separately sedimented through 20–60 % percoll gradients. The mitochondria that had not been passed through a needle formed a single band and are designated undamaged. In contrast, the mitochondria which were passed through a needle formed two distinct bands of approximately equal intensity, one at the same position as undamaged mitochondria, and the other (presumably damaged) at a higher density. This change in density is characteristic of disruption of the outer mitochondrial membrane (Rickwood et al. 1987). From these two mitochondrial fractions, 40–160 nt fraction of RNA was prepared as before.

Northern analysis of tRNA cellular distribution

Cytoplasmic, mitochondrial and damaged mitochondrial RNAs (1 μg each) were electrophoresed in a denaturing (15 % polyacrylamide, 8 M urea) gel, electroblotted to nylon membranes and probed with 32P-labeled oligonucleotides as has been previously described (Okimoto et al. 1990). The sequences of the oligonucleotides used to probe for mt-tRNAf-Met (TMB-1) and mt-tRNATrp (TBW-1) were 5′-TGTAGAGGCAGGAATTGAA and 5′-CGCACTTTTAGCTTTGAAGGC, respectively. The sequences of the oligonucleotide probes for nuclear-encoded tRNAVal and tRNAi-Met were 5′-GTGATAACCACTACACTACG and 5′-GCAGAGCGAGGTTTCGATC, respectively.

Results

M. senile mt-tRNAs

We carried out experiments to examine whether M. senile mitochondria contain functional transcripts of the mt-tRNAf-Met and mt-tRNATrp genes. Northern blot analyses were performed on mt-RNA and wc-RNA (Fig. 1) using mt-tRNAf-Met-specific and mt-tRNATrp-specific oligonucleotide primers (Figs. 2a, 3a). In each of the resulting autoradiographs, a single band of the approximate sizes expected for mt-tRNAf-Met and mt-tRNATrp was observed (Figs. 2b, 3b). For each tRNA probe, the band was darker in the mt-tRNA-containing lane than in the wc-RNA-containing lane. These data indicate that RNA molecules of the sequences and approximate sizes expected for tRNAs encoded by the mt-tRNAf-Met and mt-tRNATrp genes are present in M. senile mitochondria.

To determine whether CCA is post-transcriptionally added to the 3′ ends of each of the two mtDNA-encoded tRNAs, as would be expected if they are functional, we did the following: M. senile mt-RNAs were polyadenylated, then used in an RT-PCR containing an mt-tRNAf-Met-specific sense primer (Fig. 2a). Sixteen clones of the PCR amplification product were obtained and sequenced. All were mt-tRNAf-Met sequences. Eight of these sequences contained all of the expected tRNA sequence 3′ to the gene-specific primer, that is, nucleotides 41–71, followed by CC and a poly-A sequence (Fig. 2c). This indicates that to at least half of the mt-tRNAf-Met molecules, CC nucleotides, and possibly an A nucleotide, but not any of the three other nucleotides, are added following transcription: a post-transcriptionally added A would not be distinguishable from the synthetically added A. In the remaining eight cloned sequences, a poly-A sequence followed various nucleotides of the 3′-end proximal region of the encoded tRNA sequence. It seems more likely that these sequences resulted from degradation of the tRNA 3′-end during the experimental procedures than from incomplete transcription.

A similar experiment was carried out for mt-tRNATrp. In this case, the RT-PCR contained an mt-tRNATrp-specific sense primer (Fig. 3a). Twelve clones from the amplified product were sequenced. All 12 were tRNATrp sequences, but all lacked a 3′-terminal CC(A). In three and nine, a poly-A sequence followed the 3′-terminal nucleotide and the 3′-sub-terminal nucleotide predicted by the mt-tRNA gene, respectively. Because the gene-specific primer (TBW-2) used in PCR amplifications recognizes sequences 3′ to the nucleotide following the anticodon [nt 35 in Fig. 3a, nt 37 in standard tRNAs (Sprinzl et al. 2005)] which is a modified A in many mt-tRNAs as well as cytoplasmic tRNAs (Randerath et al. 1981; Roe et al. 1982; Dirheimer et al. 1995; Helm et al. 1998), it seems plausible that the TBW-2 primer had selected tRNAs in which A35 modification, and perhaps CCA addition, had yet to occur. To test this possibility, we repeated the PCR experiment using a primer (TBW-3) that matches nucleotides 36–57 of the mt-tRNATrp gene (Fig. 3a). Of 15 mt-tRNATrp sequences recovered, one had a 3′-terminal CC followed by poly-A (Fig. 3c, II). In the other 14 sequences, poly-A followed the 3′-terminal nucleotide (Fig. 3c, I), the sub-terminal nucleotide and an internal nucleotide in 9, 4, and 1, respectively. These data suggest that of the tRNATrp molecules present in M. senile mitochondria, only a small fraction contain a post-transcriptionally added CC(A). At the same time, it remains tenable that nucleotide modifications specific to mature mt-tRNATrp molecules resulted in selection against fully processed tRNAs by the mt-tRNATrp gene-specific primers used (TBW-2 and TBW-3, Fig. 3a). Data from both of the above experiments again suggest degeneration at the tRNA 3′ end during processing.

M. senile nuclear-encoded tRNA sequences

The primary nucleotide sequence of the TψC loop (nucleotides 54–60) of eukaryotic nuclear DNA-encoded tRNA genes is highly conserved (5′-TTCRANY) and the loop is almost always preceded by two G–C base pairs, G52–C62 and G53–C61 (Sprinzl et al. 2005; Juhling et al. 2009). This information was used to design a degenerate oligonucleotide primer that could be used to PCR amplify the 3′ terminal of 11 nucleotide-encoded sequences of M. senile nuclear DNA-encoded tRNAs (Fig. 4A). M. senile mt-RNAs were polyadenylated, then used in an RT-PCR containing an antisense primer that recognizes the 3′-CCA of proposed tRNAs (Fig. 4A). The double-stranded PCR products were cloned and sequenced. Of the cloned sequences derived from mt-RNAs and wc-RNAs, 22 (of 23 total) and 18 (of 21 total) were of the size [11 nt + CC(A)] expected for tRNA sequences between the 3′ end of the degenerate primer and the 3′ end of the tRNA (nucleotides 63–73 + CCA, Fig. 4A). Among the 22 mt-RNA-derived, 11 nt sequences, there were 16 different kinds (that is a small number of sequences were represented multiple times). Of the 18 wc-RNA-derived, 11 nt sequences, there were 17 different kinds but 4 of these were also recovered from mt-RNA. Therefore, a total of 29 different sequences were obtained from mt-RNA and wc-RNA.

To determine the sequences that extend 5′ to 4 of the above-mentioned 11 nt sequences (Fig. 4A; a–d), 5′-RACE experiments were carried out (Fig. 4B). CDNAs synthesized from 3′ polyadenylated wc-RNAs were 3′ poly-C tailed, then used in PCR amplification methods with a primer that includes a restriction site-containing sequence followed by 8 Gs, and with one of four primers (TBAA1-N, Fig. 4B) each comprising a restriction site-containing sequence followed by GG, and a sequence complementary to one of the four 11 nt sequences (a–d) shown in Fig. 4A. The PCR products were cloned and sequenced. The four sequences obtained were 52, 32, 63 and 36 nt and each could be folded into a partial tRNA. The variable, less than complete tRNA sequence lengths may have resulted from inhibition of reverse transcriptase by different modified nucleotides (Limbach et al. 1994) in the four tRNA sequences recognized by the primers. The 32 and 36 nt sequences stopped short of the anticodon. However, the 63 and 52 nt sequences included the anticodons 5′-TTC and 5′-GTC, expected for tRNAAsp and tRNAGlu (Fig. 4B). Each of the four 11 nt presumptive tRNA-specific primers used in this experiment were 32P-end labeled and used to probe Northern blots of mt-RNA and wc-RNA. In each lane, primers hybridized to a single band of the expected size (about 70 nt) for tRNAs, and in each case, the signal was stronger for wc-RNA than for mt-RNA (data not shown).

To determine the nucleotide sequences of M. senile nuclear DNA-encoded tRNAs with TψC arm sequences that might not have been recognized by the degenerate TBTψC-1 oligonucleotide, we carried out PCR amplification using double-tailed cDNAs [designated a double-RACE experiment (Fig. 4C)]. The double-stranded PCR products were cloned and sequenced. Twenty different sequences were obtained. Ten of these were extensions of those identified in experiments using the degenerate TBTψC-1 primer (Fig. 4A) and ten were new sequences. However, 19 of the 20 sequences were in the range of 15–17 nt. Again, it seems likely that inhibition of reverse transcriptase by modified nucleotides contributed to the short length of these sequences. The one exception was a new sequence of 45 nt that could be folded into a partial tRNAGln (anticodon 5′-CTG, Fig. 4C). The A52–U62 base pair present in this sequence may have impaired pairing of the TBTψC-1 primer in the experiment represented in Fig. 4A. An oligonucleotide recognizing the partial putative tRNAGln sequence hybridized to a single band of tRNA size in a Northern blot of wc-RNA (not shown).

M. senile cytoplasmic initiator tRNAi-Met

One of the most highly conserved tRNAs in eukaryotic cells is the nuclear DNA-encoded initiator tRNAi-Met (RajBhandary and Chow 1995). To determine the nucleotide sequence of the M. senile tRNAi-Met, we performed a 3′-RACE amplification using a degenerate oligonucleotide (primer TBiM-1, Fig. 4D) based on the consensus sequence of nucleotides 10–31 of 4 eukaryotic cytoplasmic initiator tRNAs (S. cerevisiae, X. laevis, mouse and human). The PCR product was cloned and sequenced, and found to comprise a 42-nt sequence that could be folded into a partial tRNA with a 5′-CAT anticodon, 3 Cs (nucleotides 39, 40, 41) at the base of the anticodon stem and 6 nucleotides (C33, A54, A60, G70, C71, and T72) that are characteristic of a eukaryotic tRNAi-Met (Fig. 4D) (RajBhandary and Chow 1995).

Cloning of M. senile nuclear DNA-encoded tRNA genes

To determine the identity of the tRNAs represented by 2 of the 11 nt sequences (c and d) shown in Fig. 4A, and to obtain the remainder of the tRNAi-Met sequence, a PCR-based genomic cloning strategy was used to locate the genes for these tRNAs. High molecular weight M. senile DNA was digested to completion with Sau3AI and the resulting restriction fragments were ligated into BamHI-cleaved bacteriophage M13 DNA. From this nuclear DNA library, the gene sequence for each tRNA was amplified in three different PCR using, as primers, restriction site-containing oligonucleotides complementary to each of the two 11 nt sequences (c and d, Fig. 4A) and to the known 3′ portion of tRNAi-Met (primer TBiM-2, Fig. 4D), and the universal M13 sequencing primer. The complete tRNAi-Met gene sequence and a short upstream segment were obtained (Fig. 4D). This gene has 72 nt and has an identity to the mtDNA-encoded tRNAf-Met gene (Fig. 2a) of only 55.4 %. Also obtained were sequences that could be folded into tRNA genes with 5′-AAT (isoleucine-specifying) and 5′-AAC (valine-specifying) anticodons (Fig. 5a, c). Additionally, 88 nt upstream of the 5′ end of the tRNAIle gene was located a sequence whose RNA product could be folded into the secondary structure of a tRNAThr (Fig. 5a).

Fig. 5
figure 5

Nuclear DNA-encoded tRNA genes. M. senile nuclear DNA sequences containing genes for tRNAThr and tRNAIle (a) and for tRNAVal (c) shown in their presumed secondary structure configuration. The 3′-terminal 11 nt sequences for tRNAIle and tRNAVal were obtained from the results of a 3′-RACE experiment, using primer TBTψC-1 (Fig. 4A). Primers complementary to these 11 nt sequences were used with the M13 universal sequencing primer to obtain by PCR the remainder of the sequences shown (including the entire tRNAThr gene) from an M13 nuclear DNA library of Sau3AI fragments. b Northern blot. In the autoradiograph, the three lanes contain M. senile mt-RNA (mt), whole cell RNA (wc) and E. coli tRNATyr (85 nt), separated by electrophoreses through a denaturing (15 % polyacrylamide, 8 M urea) gel. Lanes mt and wc were probed with a 32P-labeled oligonucleotide that are complementary to the 11 nt sequence for tRNAIle shown in Fig. 4A. Lane E.coli Tyr was probed with a 32P-labeled oligonucleotide complementary to nt 46–66 of E. coli tRNATyr

Full size image

Distribution of nuclear DNA-encoded tRNAs in the cytoplasm and in mitochondria

Northern blot experiments were carried out to examine the cellular distribution of M. senile mtDNA-encoded tRNAf-Met and the nuclear DNA-encoded tRNAVal and tRNAi-Met (Fig. 6). Transfer RNA-specific oligonucleotides representing the mt-tRNAf-Met gene and the nuclear tRNAVal and tRNAi-Met genes were hybridized to blots of small RNAs prepared from a cytoplasmic fraction from which mitochondria had been removed by differential centrifugation, from a percoll gradient-purified mitochondrial fraction, and from a percoll gradient-purified fraction of mitochondria that had been mechanically damaged to the point of altering their relative density in a percoll gradient. As such a change in density is characteristic of mitochondria from which the outer membrane has been removed (Rickwood et al. 1987), it was reasoned that cytoplasmic tRNAs attached to the outer mitochondrial membrane would be eliminated. The results of this experiment are shown in Fig. 6. All three tRNAs were detected in each of the three RNA preparations. In the lanes containing cytoplasmic RNA, each of the three probes hybridized to a single band of the size expected for each mature tRNA. However, relative to the tRNAVal and tRNAi-Met bands (Fig. 6b, c), the mt-tRNAf-Met was much lighter (Fig. 6a). In each of the two mitochondrial fractions, each tRNA-specific oligonucleotide recognized two distinct bands, and in all cases, the upper is darker and of the size expected for the mature (CCA-containing) tRNAs. In each case, the lower band is at a position expected for molecules approximately three nucleotides smaller than those in the upper band, consistent with this band comprising tRNAs that lack the 3′-terminal CCA. It is of note here that 3′-RACE analysis of the M. senile mt-tRNAf-Met molecules showed that half of these tRNAs lacked the 3′-terminal CCA. It was surprising to find that the M. senile cytoplasmic initiator methionine tRNA occurs in high amounts in the mt-RNA fractions, as it is expected that this tRNA would not be used to initiate mitochondrial translation.

Fig. 6
figure 6

Northern blot analyses to examine the cellular distribution of M. senile mtDNA-encoded tRNAf-Met and nuclear DNA-encoded tRNAVal and tRNAi-Met. Small RNAs prepared from a cytoplasmic fraction from which mitochondria had been removed (cyt), percoll gradient-purified undamaged mitochondria (mt), and damaged mitochondria (dmt) (1 μg each) were electrophoresed in a denaturing (15 % polyacrylamide, 8 M urea) gel, electroblotted to nylon membranes and probed with 32P-end labeled oligonucleotides specific for the three tRNAs as shown

Full size image

Discussion

The aim of the present study was to obtain experimental evidence regarding the possible participation in mt-protein synthesis of the sea anemone, M. senile of the only two tRNAs (tRNAf-Met and tRNATrp) encoded in the mt-genome of this organism, and of tRNAs transcribed from the M. senile nuclear genome. The results we obtained clearly establish that RNAs of the size and sequence predicted by the mt-tRNAf-Met and mt-tRNATrp genes occur in mitochondria. Further, the finding of both mt-tRNAf-Met and mt-tRNATrp with CC (most probably followed by an A, but never by another nucleotide) at the 3′ terminus, the post-transcriptional modification expected for them to be charged with the respective amino acid, greatly strengthens the likelihood that these tRNAs are functional in mt-protein synthesis. Also, direct evidence was obtained indicating that a fraction of at least two nuclear DNA-encoded tRNAs (tRNAi-Met and tRNAVal) are imported into mitochondria, supporting the contention that these tRNAs are essential elements in the synthesis of mt-proteins.

The very low frequency of mt-tRNATrp molecules with a CC(A) detected could have resulted from interference with the PCR of post-transcriptionally modified nucleotides in CC(A)-containing tRNAs. However, there may, in fact, be a low frequency of post-transcriptional addition of CC(A) to mt-tRNATrp molecules which would correlate with the much lower use of TGA codons (14) than TGG codons (88) in the 14 M. senile mt-protein genes. The ratio of TGA to TGG codons (1:6.3) is inversely correlated to the overall ratio of other codons ending in A and G (1.7:1) in the M. senile mt-protein genes (Beagley et al. 1998). A similar finding was reported for the mtDNA of the octocoral Sarcophyton glaucum: The ratio of TGA to TGG codons is 1:14.0, and the ratio of codons ending in A and G is 2.6:1 (Pont-Kingdon et al. 1998). In contrast, in bilaterian mt-protein gene sets, the ratio of TGA to TGG codons is directly correlated with the ratio of other codons ending in A and G. The ratios of TGA:TGG codons and of A:G endings of other codons in some bilateria are: mouse, 13.9:1/14.6:1 (Bibb et al. 1981); Drosophila yakuba, 16.0:1/15.7:1 (Clary and Wolstenholme 1985); C. elegans, 6.1:1/4.1:1; A. suum, 0.6:1/0.5:1 (Okimoto et al. 1992).

Data from Northern blot analyses using an mt-tRNAf-Met-specific primer indicated that this tRNA is essentially confined to mitochondria, consistent with the expectation that it is utilized specifically to initiate the synthesis of mt-proteins. In each of the two mt-RNA fractions, the mt-tRNAf-Met-specific primer recognized two size classes of molecule, one at the position expected for a mature tRNA and one of lesser frequency at a position expected for a molecule three nucleotides shorter. This observation is consistent with the direct evidence obtained by sequencing mt-tRNAf-Met that a fraction of these molecules has a CC(A) post-transcriptionally added to their 3′ ends.

The presence of nuclear DNA-encoded tRNAVal and tRNAi-Met in both the mitochondria and the cytoplasm is clearly indicated in the Northern blots, suggesting that at least these two tRNAs function in both cytoplasmic and mt-protein synthesis. However, as was observed for the mt-tRNAf-Met primer, both the tRNAVal primer and the tRNAi-Met primer recognized two bands in each of the mitochondrion-derived RNAs, consistent with the presence of tRNAVal and tRNAi-Met molecules with and without post-transcriptionally added CCA. Again, the frequencies of the smaller molecules appeared to be less than those of the larger molecules. In contrast, each of the tRNAVal and the tRNAi-Met primers recognized only a single band in the cytoplasmically derived RNAs, at the same position as that for the larger bands in RNAs from mitochondria. Taken together, these observations indicate that although it is expected that CCA is added to nuclear DNA-encoded tRNAs immediately following transcription (Schneider 2011), tRNAVal and tRNAi-Met molecules destined for importation have CCA added only after they have entered mitochondria. However, the substantial fraction of tRNAVal and tRNAi-Met molecules that in mitochondria appear to lack CCA may indicate a lag in transportation of these tRNAs across the mitochondrial membrane, or that CCA addition is much less efficient in mitochondria than in the cytoplasm. Alternatively, these data are consistent with the interpretation that tRNAs with and without CCA may be transported into mitochondria.

Importation of nuclear DNA-encoded tRNAs into mitochondria has been extensively studied in a variety of organisms (reviewed in Schneider 2011), providing evidence for two general mechanisms. In one, examined in greatest detail for a tRNALys encoded in nuclear DNA of the yeast S. cerevisiae, and defined cytosolic factors are involved, the tRNA is linked to a lysosyl-tRNA synthetase and the complex is co-imported along the mitochondrial import pathway. In the second mechanism, found to operate in plants and protozoa, tRNAs are directly imported into mitochondria with much less or no cytosolic factors involvement. At this time, there is no direct evidence regarding the mechanism of importation of nuclear DNA-encoded tRNAs into cnidarian mitochondria. However, the detection in nuclear DNA of the hexacoral Nematostella vectensis of genes for aminoacyl-tRNA synthetases (Haen et al. 2010) that include mitochondrion-targeting sequences leaves open the possibility that a co-importation mechanism operates in Cnidaria.

From the bioinformatic analysis of DNA of N. vectensis, Haen et al. (2010) postulated the occurrence of a methionyl-tRNA synthetase expected to have a bifunctional role in cytoplasmic protein synthesis and mt-protein synthesis. However, a methionyl-tRNA synthetase expected to be specific to mtDNA-encoded methionine was not detected. In view of our finding of M. senile mt-tRNAf-Met molecules with post-transcriptionally added CC(A), it seems plausible that the single imported methionyl-tRNA synthetase is responsible for the charging of formylated tRNAMet molecules involved in the initiation of mt-protein synthesis as well as tRNAMet molecules specifying internal methionines. It remains unclear what role the imported nuclear DNA-encoded tRNAi-Met molecules we detected in M. senile mitochondria might play in mt-protein synthesis. One possibility is that this tRNA participates in decoding internal AUG codons in transcripts of mt-protein genes. However, this would not be consistent with the situation in either eukaryotic cytoplasmic protein synthesis or prokaryotic protein synthesis, where in each case a tRNAMet (elongator tRNAMet) with a considerably different primary structure (sequence) than the corresponding tRNAi-Met is used to decode internal AUG codons (RajBhandary and Chow 1995; Beagley et al. 1995).

As noted previously (Beagley et al. 1998), the retention of genes for tRNAf-Met and tRNATrp in the M. senile mt-genome is likely related to functions specific to these tRNAs: the initiation of metazoan mt-protein synthesis by formyl-methionine rather than methionine (Chomyn et al. 1981), and the need for a tRNA that recognizes TGA as tryptophan-specifying. Although the 5′-UCA anticodon of the mt-tRNATrp would be expected to recognize both 5′-UGA and 5′-UGG codons (Beagley et al. 1998), it is clearly worthwhile to consider whether in translation of M. senile mt-protein genes, 5′-UGG codons might be decoded by a tRNA transcribed from nuclear DNA that is also utilized to recognize 5′-UGG codons in transcripts of nuclear protein genes. This hypothesis is supported by the bioinformatic data of Haen et al. (2010) that N. vectensis nuclear DNA encodes two tryptophyl-tRNA synthetases, one specific for the mtDNA-encoded tRNATrp and one that is specific for a nuclear-encoded tRNATrp.

A paucity of tRNA genes appears to be a common feature of Cnidarian mtDNAs. Genes for only tRNAf-Met and tRNATrp occur in mtDNAs of species of the subclass Hexacorallia of Anthozoa, Hydrozoa, and Scyphozoa so far examined (Beagley et al. 1995, 1998; Pont-Kingdon et al. 2000; van Oppen et al. 2002; Shao et al. 2006; Kayal and Lavrov 2008). However, mtDNAs of species of the subclass Octocoralia of Anthozoa contain only tRNAf-Met. (Beagley et al. 1995; Beaton et al. 1998; Medina et al. 2006; Fukami and Knowlton 2005). Clearly, for protein synthesis to occur in M. senile mitochondria, all but two tRNAs must be imported from the cytoplasm. The data presented in this paper suggest that at least some of the M. senile nuclear DNA-encoded tRNAs are utilized in both cytoplasmic and mt-protein synthesis. This is in line with previous findings regarding angiosperm plants for which there is evidence for the common use of some nuclear DNA-encoded tRNAs in both the cytoplasm and the mitochondria (Marechal-Drouard et al. 1988), and for the protozoans L. tarentolae and T. brucei in which all tRNAs used in mt-protein synthesis are nuclear DNA-encoded, and are also used in cytoplasmic protein synthesis (Simpson et al. 1989; Hancock and Hajduk 1990). However, it has been shown that in T. brucei, the nuclear DNA-encoded tRNAi-Met is excluded from importation into mitochondria (Esseiva et al. 2004) which contrasts with our conclusion regarding this tRNA in M. senile.

Both M. senile mt-tRNAf-Met and mt-tRNATrp have primary and secondary structural properties characteristic of M. senile nuclear DNA-encoded tRNAs (and other standard tRNAs) rather than the reduced structural characteristics of bilaterian mt-tRNAs. This is in line with all of the tRNAs within a mitochondrion having to interact with a single ribosome type. Interestingly in this regard, it has been found that the predicted secondary structures of M. senile mt-s-rRNA and mt-l-rRNA both more closely resemble the corresponding rRNAs of prokaryotes than the mt-rRNA of any bilaterian species (Pont-Kingdon et al. 1994, 2000; Beagley et al. 1998).

The data reported from the present study provide new information regarding the mechanism of mt-protein synthesis in a cnidarian. Clearly, an analysis of the amino acids associated with isolated tRNAs is needed to confirm the suggested involvement of the two mtDNA-encoded tRNAs and of the nuclear DNA-encoded tRNAs, particularly tRNAi-Met. A continuation of the presently reported study can be expected to fully define the identity of all nuclear DNA-encoded tRNAs that are imported into mitochondria, and together with the bioinformatic information regarding importation of aminoacyl-tRNA synthetases (Haen et al. 2010) would greatly extend our understanding of how mt-proteins are synthesized in Cnidaria.