Mitochondrial RNAs of myxomycetes terminate with non-encoded 3′ poly(U) tails

Abstract

We examined the 3′ ends of edited RNAs from the myxomycetes Stemonitis flavogenita and Physarum polycephalum using a modified anchor PCR approach. Surprisingly, we found that poly(A) tails are missing from the cytochrome c oxidase subunit 1 mRNA (coI) from both species and the cytochrome c oxidase subunit 3 mRNA (cox3) from P.polycephalum. Instead, non-encoded poly(U) tails of varying length were discovered at the 3′ ends of these transcripts. These are the first described examples of 3′ poly(U) tails on mature mRNAs in any system.

INTRODUCTION

RNA molecules are routinely processed by splicing, editing and 3′ polyadenylation, resulting in transcripts that contain information not encoded by the DNA genome. In eukaryotic nuclear genes, 3′ end processing consists of mRNA cleavage followed by the addition of a poly(A) tail by a group of factors, including poly(A) polymerase and the C-terminal domain of RNA polymerase II (1). Some prokaryotic, mitochondrial and chloroplast mRNAs also have poly(A) tails. While polyadenylation of eukaryotic nuclear genes enhances stability and translation initiation (2), in prokaryotes and chloroplasts polyadenylation provides a signal for rapid RNA degradation (3,4). In mitochondrial systems, polyadenylation has a variety of functions: creating stop codons on most human mitochondrial transcripts (5), possibly signaling for translation initiation in trypanosomes (6), and stimulating quick degradation of a plant mitochondrial mRNA (7) and some trypanosome RNAs (8).

Adenosine is not the only nucleotide found in unencoded 3′ tails. Kinetoplastids, unicellular eukaryotes with extensive uridine insertional/deletional editing in their mitochondria, have 3′ polyuridine tails on guide RNAs (gRNAs), a population of mitochondrial RNAs involved in editing (9). These non-encoded poly(U) tails may aid the editing complex in holding together the broken halves of the purine-rich mRNA during the cleavage stage of the editing process (10,11), or might denature secondary structure in the regions of the mRNAs being edited (12). The other class of non-messenger RNA encoded by the kinetoplastid mitochondrial genome, the ribosomal RNAs, also have non-encoded 3′ poly(U) tails (13). Poly(U) tails have even been detected on the ends of many unedited and partially edited pre-mRNAs in Trypanosoma brucei, and are thought to be added by a rampant terminal uridyl transferase activity operating on editing intermediates (14).

Physarum polycephalum is a myxomycete, or plasmodial slime mold, that is amenable to cellular study in the laboratory. The production of functional mitochondrial transcripts for almost all of P.polycephalum’s messenger and structural RNAs requires several types of RNA editing. Many single cytidine insertions, a small number of uridine and mixed dinucleotide insertions, and a few instances of cytidine to uridine base conversions modify the RNA sequences. For instance, the cytochrome c oxidase subunit 1 (coI) mRNA is edited by insertion of 59 Cs, a single U and three mixed dinucleotides. Four C to U conversions are also found in this transcript (15).

Here, we present our discovery of a new form of RNA processing that alters myxomycete mitochondrial transcripts. We show that P.polycephalum and Stemonitis flavogenita both have non-encoded 3′ poly(U) tails added to edited mitochondrial mRNAs. The unusual tails on mRNAs that have also undergone editing suggest a possible connection between these types of RNA sequence change.

MATERIALS AND METHODS

Cultures

Freeze-dried cultures of S.flavogenita (24714) were obtained from the American Type Culture Collection, and grown on half-strength cornmeal agar plates into full size plasmodia. Plates of P.polycephalum plasmodia were obtained from Carolina Biologicals.

Isolation of nucleic acids

RNA and DNA were extracted from the slime mold plasmodia by use of Trizol reagent from Life Technologies. RNA was treated with DNase (Promega); DNA was treated with RNase A (Sigma). Nucleic acids were then extracted with phenol/chloroform, ethanol precipitated, and resuspended in 10 mM Tris, pH 7.4, 0.1 mM EDTA.

Artificial RNA tailing, reverse transcription and PCR of cDNA

GTP tails were added to total RNA by yeast poly(A) polymerase as described (16) in the presence of 0.5 mM GTP. Reverse transcription was performed using SuperScript II reverse transcriptase from Life Technologies, and primer UXR′C12 on S.flavogenita RNA, and TXRC12 for P.polycephalum RNA (primer sequences listed below). The nested PCR of S.flavogenita coI cDNA was performed for 20 cycles with primer CUAUXR′ and primer coi551st, followed by 20 cycles with CUAUXR′ and primer coi561st. Physarum polycephalum coI cDNA was amplified in 37 cycles of PCR, with primers 3PPcoiF and TXR. Physarum polycephalum cox3 cDNA was amplified in 40 cycles of PCR with primers 3PPcox3F and TXR. Physarum polycephalum mitochondrial LSU cDNA was amplified in a nested PCR of 20 cycles with primers 3PLSUF1 and TXR, followed by 25 cycles with primers 3PPLSUF and TXR. Physarum polycephalum nuclear SSU cDNA was amplified in 37 cycles of PCR with primers 3PPSSUF and TXR.

Amplification of DNA

Walking PCR of S.flavogenita DNA was performed as described (17). The single strand amplification was 40 cycles with primer coi551b. The second PCR was 22 cycles with primer coi561st and UXR′C12. The third PCR was 25 cycles with primer coi581st and CUAUXR′. A single clone was obtained by this method, the plasmid isolated and sequenced. Primer stem-ptR was designed at the 3′ most extreme of this clone sequence, then used in a 40 cycle PCR with primer coi551st. The PCR product was precipitated and directly sequenced with primer coi551st and stem-ptR.

Cloning, purification and sequencing

PCR products were cloned with the TOPO TA cloning kit from Invitrogen. Plasmids were purified with the High Pure Plasmid Isolation Kit from Boehringer-Mannheim/Roche. Both strands of all plasmids were sequenced at the Princeton University SynSeq facility.

Primer sequences

(Note that D designates a 1:1:1 mixture of A, G, and T.)

UXR′C12 (5′-CUACUACUACUACTCGAGAATTCCCCCCCCCCCCD-3′)

TXRC12 (5′-CATCATCATCATCTCGAGAATTCCCCCCCCCCCCD-3′)

CUAUXR′ (5′-CUACUACUACUACTCGAGAATT-3′)

TXR (5′-CATCATCATCATCTCGAGAATT-3′)

coi551st (5′-TTGTTAGCAAATGATTATCG-3′)

coi551b (5′-biotin-TTGTTAGCAAATGATTATCG-3′)

coi561st (5′-TACATTTCCTTTAACTGTTGC-3′)

3PPcoiF (5′-CGCCGTATTCCAGATTATCCTGATGC-3′)

3PPcox3F (5′-CATGCTCCTTTCTCTATTTCTGATGG-3′)

3PLSUF1 (5′-TCTGTCTAGTACGAAAGGACTGG-3′)

3PPLSUF (5′-TGAGCTGTTTGCGCACGCTCATTCGC-3′)

3PPSSUF (5′-GTAAAACGAGTGCTTGAACAAGGCGTCC-3′)

stem-ptR (5′-TAAGTAAATGCAGTAACATTTG-3′)

RESULTS

While investigating the distribution and types of RNA editing in myxomycetes, we attempted to recover the 3′ end of the coI mRNA of S.flavogenita by anchor PCR, a technique that relies on the presence of a 3′ poly(A) tail (18). Though sequences obtained by this method extended to near the 3′ end of the predicted coding region, the sequences lacked a stop codon. RT and PCR products were not full length, due to annealing of our modified poly(T) primer to the A-rich sequences still within the coding region. To circumvent the problems associated with traditional anchor RT and PCR of a transcript of such a high A/T content (70%), we used poly(A) polymerase to add an artificial G tail to the RNA (16), then reverse-transcribed and amplified from the introduced tail into the coding region.

Sequence analysis of six cloned cDNA fragments revealed that the 3′ end of S.flavogenita coI mRNA has a conventional termination codon (UAA), followed by a 24 nt untranslated region and a homopolymeric tail of 20–31 nt (Fig. 1). Surprisingly, the tail is not composed of poly(A), as expected, but consists primarily of uridines. We recovered the DNA sequence in this region by a walking PCR approach (17), and found that the poly(U) tail is not encoded in the genomic sequence (Fig. 1). Didymium nigripes coI cDNA clones also terminate with similar poly(U) tails (data not shown), although we did not determine the corresponding DNA sequence.

Figure 1.

Poly(U) tails on S.flavogenita coI cDNA. The 3′ untranslated region of S.flavogenita coI cDNA clones are aligned with directly sequenced DNA PCR product (GenBank accession no. AF239222). The alignment begins with nucleotide 1805 (T) in the GenBank record, and the inferred stop codon is underlined. Dotted regions within clone sequences indicate identity with the DNA sequence. The regions with white letters inside a black box indicate poly(U) tails. The gray shaded regions designate the primer used for reverse transcription. Slight variation in the length of poly(G) regions of reverse transcription primers may be due to minor imperfections in the primer pool, or PCR slippage through the homopolymeric region. The final base of the primer is a non-C anchor base, intended to direct annealing location of this primer to the most 5′ end of a potentially long poly(G) run during reverse transcription. Non-primer-derived nucleotides beyond the poly(U) tails (black letters on a white background within the clone sequences) are probably artifacts of the tailing proceedure, as explained in the Results.

To expand our survey of the distribution of this type of poly(U) tailing in a myxomycete with a greater number of published mitochondrial gene sequences, we amplified the 3′ region of several P.polycephalum RNAs by the same modified anchor PCR technique. We examined the termini of two edited mitochondrial mRNAs, an edited mitochondrial structural RNA, and a non-edited nuclear structural RNA. We found that the non-encoded poly(U) tail is a common feature of the edited mitochondrial mRNAs in both species.

Although both species share the presence of a 3′ poly(U) tail on the coI transcripts, P.polycephalum’s coI tails are shorter than those of S.flavogenita. Physarum polycephalum’s coI tails are only 9–25 nt long, and vary in their start site on the RNA over a 33 nt region (Fig. 2). In seven clones analyzed, only one tail contained a single cytidine residue, as compared to two out of six S.flavogenita clones, which contain one and three cytidines apiece. Another P.polycephalum clone contained a guanidine residue amidst the uridine run. When the P.polycephalum tail sequences are aligned with their corresponding regions on the mitochondrial genomic DNA sequence (J.M.Gott, personal communication), the Us are not encoded in the genomic copy. The cDNA sequences agree with nuclease protection experiments that imply that the end of the P.polycephalum coI mRNA is ~50 bases downstream of the stop codon (L.M.Visomirski-Robic and J.M.Gott, personal communication).

Figure 2.

Poly(U) tails on P.polycephalum coI cDNA. The 3′ untranslated region of P.polycephalum coI cDNA clones are aligned with DNA sequence from Jonatha Gott (personal communication). Numerical notation is continuous with GenBank accession no. L14779. The inferred stop codon is underlined. Annotation as in Figure 1.

RNA editing adds 32 Cs and a single UC dinucleotide to the cytochrome c oxidase subunit 3 (cox3) mRNA in P.polycephalum (19). Five out of six clones of the 3′ end of cox3 terminate with poly(U) tails (Fig. 3). The tails range from 12 to 37 bases in length, with the starting point of the tails spanning a 21 base region. The tails in these clones were composed uniformly of U residues. We note that for all of the P.polycephalum cDNA transcripts, some clones contain a non-encoded A/G-rich sequence just upstream of the 3′ primer. These A/G stretches are probably caused by a few contaminating adenosines in the poly(A) polymerase-catalyzed G tailing reaction, since yeast poly(A) polymerase catalyzes the addition of adenosine twice as efficiently as guanosine (16). If the artificial tail were contaminated with a few As, then the 3′ anchor primer (TXRC12) would anneal slightly downstream of the real beginning of the artificial tail, resulting in the presence of these A residues interspersed with poly(G) tails. The fact that these A/G regions are found in common on all P.polycephalum sequences, even those without poly(U) tails (see below), supports this explanation, and confirms a real difference between poly(U)-tailed and non-poly(U)-tailed transcripts.

Figure 3.

Poly(U) tails on P.polycephalum cox3 cDNA. The 3′ untranslated region of P.polycephalum cox3 cDNA clones are aligned with corresponding DNA sequence (GenBank accession no. AF084526). The first nucleotide of the alignment corresponds to nucleotide 3368 (T) in the GenBank record, and the inferred stop codon is underlined. Annotation as in Figure 1.

We also amplified the terminal region of a structural RNA, the mitochondrial large subunit rRNA (23S), which is edited by 52 C insertions and five dinucleotide insertions (19). Analysis of six clones corresponding to the 3′ end of the mitochondrial LSU transcript revealed only a few U residues: two transcripts terminated with one U each, and one ended with three Us (Fig. 4A). These ‘tails’ were added over a 7 bp region, though one of the clones that lacked any apparent U tail ended 30 bases downstream of the end of the earliest-terminating clone. Some clones had small poly(A) tails, but these could be artifacts of the artificial tailing process, as described above.

Figure 4.

Physarum polycephalum mitochondrial large subunit ribosomal RNA (mit LSU) cDNA and nuclear small subunit ribosomal RNA (SSU) cDNA. (A) Physarum polycephalum mitochondrial large subunit ribosomal RNA (mit LSU) cDNA clones are aligned with the DNA sequence (GenBank accession no. AF080602). The first position of the alignment corresponds to nucleotide 2783 (T) in the GenBank record. (B) Physarum polycephalum nuclear small subunit ribosomal RNA (SSU) cDNA clones are aligned with the DNA sequence (GenBank accession no. X13160). The alignment begins with nucleotide 1950 (G) in the GenBank record. Annotation as in Figure 1. (The extreme ends of the lower clones’ reverse transcription primer sequences have been omitted for clarity.)

To ascertain whether the tails were unique to mitochondrially encoded transcripts, which had undergone RNA editing, we also examined the 3′ end of the P.polycephalum 18S nuclear-encoded small subunit (SSU) rRNA. No nuclear transcripts in P.polycephalum are known to exhibit editing. Six clones of the SSU transcript displayed no evidence of poly(U) tails on the SSU rRNA (Fig. 4B). These sequences also serve as a negative control for the poly(G) tailing and reverse transcription, as both treatments were performed singly for the total P.polycephalum RNA sample. Detection of different RNA transcripts both with and without 3′ poly(U) tails by this method strengthens our assertion that the poly(U) tails are naturally present on the mitochondrial RNA transcripts.

DISCUSSION

Although poly(A) tails are present on nuclear mRNAs of P.polycephalum (20), and common on mitochondrial mRNAs in some organisms, they are not known to be universally present, and have not been directly detected on mitochondrial RNAs of myxomycetes. We have found that at least one S.flavogenita mitochondrial mRNA and two P.polycephalum mitochondrial mRNAs are not polyadenylated, but rather polyuridylated. With the exception of the incompletely edited pre-mRNAs of kinetoplastid mitochondria (14), these are the first described poly(U) tails on mRNA transcripts in any organism. Interestingly, both kinetoplastids and myxomycetes share RNA editing of mitochondrial transcripts, albeit by very dissimilar and presumably independently evolved mechanisms. However, in contrast to the poly(U)-tailed kinetoplastid pre-mRNAs, which appear to be transient intermediates in the editing process, the myxomycete poly(U)-tailed transcripts appear to be mature mRNAs, as they show complete editing in the 3′ region that was amplified and sequenced (21). Because insertional editing in myxomycetes is believed to be cotranscriptional, proceeding from the 5′ to the 3′ end of the transcript (22,23), the mRNAs are probably completely translatable.

The absence of clearly defined poly(U) tails on mitochondrial large subunit rRNA may mean that the tails perform translation-related functions for mRNAs analogous to the roles of poly(A) tails in other organisms. The tails may be actively targeted exclusively to mRNAs. In eukaryotic nuclear systems, where poly(A) tails are only found on mRNAs, their absence on rRNAs is explained by transcription of messenger and structural RNA by different polymerases, where pol II is actively involved in the tail addition (1). In mitochondrial systems, a single polymerase produces all transcripts. It is unclear how addition of poly(A) tails is restricted to mRNAs in mitochondria; in fact, a few mitochondrial RNAs with 3′ polyadenylation have been detected in Plasmodium falciparum, mosquitoes and mammals (24). It is possible that some fundamental difference in sequence or secondary structure of the P.polycephalum mitochondrial large subunit rRNA does not promote its termination by poly(U), or that another process specifically removes tails after addition.

The poly(U) tail sequences in this study vary somewhat in length and overall composition; a few tails include cytidine and guanine residues. A previous study of edited trypanosome cDNAs concluded that though PCR-induced mutation occurred at a low frequency overall, the PCR error within long T homonucleotide stretches was much higher than the other regions of the template. In fact, in nearly 3 kb of analyzed sequence, 11 of 14 total PCR-induced mutations occurred in homo(T) regions, with the majority of these in the longest T stretch. Most mutations consisted of a single T insertion or deletion per poly(T) run, probably due to polymerase slippage, and there was also one T to C transition (25). However, even if most of the variation we observed within the mononucleotide runs of myxomycete poly(U) tails is real, similar or greater mixed base composition has been noted in some poly(A) sequences in organelles. The poly(A) rich 3′ tail sequences of spinach chloroplast genes contain ~25% guanosines and a combined 5% uridines and cytosines (3). Trypanosome poly(A) tails also contain occasional U insertions, which may be added either by the RNA polymerase or by the terminal uridyl transferase (TUTase) activity present in these organisms (26).

This study demonstrates that RNA processing of mitochondrial mRNA transcripts in myxomycetes includes not only RNA editing, but also 3′ polyuridylation. All of the mRNA transcripts in which we found 3′ poly(U) tails were also processed by editing (21). To discern whether these two unique forms of RNA processing are related, one might examine the 3′ end of a mitochondrial mRNA uninvolved in RNA editing. However, all currently known P.polycephalum mitochondrial mRNAs require editing (19), and the editing process is extremely efficient (27). Thus, while mysteries abound in the organelles of myxomycetes, further conclusions await progress in molecular techniques, such as isolation of non-tailing or non-editing mutants and the facile transformation of mitochondria.