Expression and processing of an unusual tRNA gene cluster in the cyanobacterium Anabaena sp. PCC 7120


Correspondence: Agustín Vioque, Instituto de Bioquímica Vegetal y Fotosíntesis, Universidad de Sevilla and CSIC, 41092 Sevilla, Spain. Tel.: +34 954489519; fax: +34 954460065; e-mail:


Anabaena sp. PCC 7120 is a filamentous cyanobacterium that bears a cluster of 26 tRNA genes and pseudogenes in the delta plasmid. The sequences of these tRNAs suggest that they have been acquired by horizontal gene transfer from another organism. The cluster is transcribed as a single transcript that is quickly processed to individual tRNAs. RNase P and RNase Z, in vitro, are able to process precursors containing some of these tRNAs. Deletion of the cluster causes no obvious phenotype or effect on growth under diverse culture conditions, indicating that the tRNAs encoded in the cluster are not required for growth under laboratory conditions, although they are aminoacylated in vivo. We have studied a possible tRNASer [tRNASerGCU(2)] present in the cluster with a sequence that deviates from consensus. This tRNA is processed in vitro by RNase P at the expected position. In addition, this tRNASerGCU is specifically aminoacylated with serine by an Anabaena sp. PCC 7120 crude extract. These data indicate that tRNASerGCU(2) is fully functional, despite its unusual structure. Similar clusters are found in other three cyanobacteria whose genomes have been sequenced.


Anabaena sp. PCC 7120 (hereafter Anabaena 7120) has 48 tRNA genes in its chromosome, which should be theoretically enough to decode all amino acids for protein synthesis. In addition, a cluster of 26 tRNAs, seven of them pseudogenes, is encoded in one of the plasmids found in this organism (plasmid delta; Kaneko & Tabata, 1997; Fig. 1).

Figure 1.

The tRNA cluster of Anabaena 7120 delta plasmid. (a) tRNA genes encoding the 3′-CCA sequence are stripped. Pseudogenes are shown in grey. Small arrows indicate primers used for RT-PCR. Long double-headed arrows indicate the RT-PCR products shown in (b). (b) Overlapping RT-PCR of the delta plasmid tRNA cluster. M, size markers, labelled on the left; 1, RT-PCR with primers F1 + R1; 2, RT-PCR with primers F2 + R2; 3, RT-PCR with primers F3 + R3; 4, RT-PCR with primers R4 + F4; 5, RT-PCR with primers F5 + R5; 6, RT-PCR with primers F6 + R6; 7, RT-PCR with primers F7 + R7.(+) complete reaction; (−) control without reverse transcriptase. (c) Northern blot of total RNA from Anabaena 7120 with probes corresponding to several of the tRNA genes in the cluster or the intergenic region Int2. The probes were generated by PCR with primers specific for the indicated tRNAs or intergenic region.

Clusters of tRNA genes that are transcribed together are found in large DNA viruses and in bacterial genomes, but not in cyanobacteria, where tRNA genes are dispersed in the genome and transcribed as single precursors, except tRNATyr and tRNAThr that generally are transcribed together as a dimeric precursor (Tous et al., 2001).

Cyanobacterial tRNA genes mostly lack the 3′-end CCA sequence. In many species, none of the tRNA genes contain the 3′-CCA sequence. In most other cyanobacterial strains, only one, usually the initiator inline image, or two tRNA genes contain the 3′-CCA sequence. CCA-lacking precursors are processed at the 3′ side by RNase Z (Hartmann et al., 2009). Cyanobacterial RNase Z, like other studied RNase Z, cannot cleave CCA-containing precursors (Ceballos-Chávez & Vioque, 2005). CCA-containing precursor tRNA (pre-tRNAs) are processed exonucleolytically (Schurer et al., 2001). In cyanobacteria, the processing of CCA-containing pre-tRNAs has not been characterized. All tRNA precursors are processed at the 5′ side by RNase P.

We have studied the expression and processing of the tRNAs encoded in the delta plasmid of Anabaena 7120, and we have determined that they are correctly processed and aminoacylated. During the study of the tRNA cluster, we have identified a variant tRNASerGCU that was not annotated in the database. A structural analysis of this tRNA shows that it presents a tRNA-like structure, with a serine GCU codon, and other determinants of a tRNASer. We demonstrate that this newly identified tRNA is aminoacylated in vitro and in vivo.

Materials and methods

Strains and growth conditions

Anabaena 7120 (Rippka et al., 1979) was grown photoautotrophically at 30 °C under white light (65–100 μE m−2 s−1). Medium used for growth on plates was BG11 (NaNO3 as the nitrogen source) or BG110 (N2 as the nitrogen source; Rippka et al., 1979). Liquid cultures were grown in the same media supplemented with 10 mM NaHCO3 and bubbled with 1% CO2-enriched air.

Preparation of Anabaena 7120 crude extracts

Cells from cultures grown to 5 μg chlorophyll mL−1 were collected by filtration (filter type: 0.45 μm HA; Millipore HAWP05000) and washed in RNase-free TE buffer [10 mM Tris–HCl (pH 7.5), 1 mM EDTA]. Pelleted cells were reduced to dust after freezing in liquid nitrogen and resuspended in a buffer containing 50 mM HEPES-KOH (pH 7.5), 10 mM MgCl2, 5 mM CaCl2 and 20% glycerol, and the samples were centrifuged at 2500 g for 10 min at 4 °C. Protein was quantified by Lowry's method (Lowry et al., 1951).

RNA isolation

Cells pellets prepared as described above were resuspended in 100 μL of a lysozyme solution (3 mg mL−1) and subjected to three freeze–thaw cycles to facilitate cell lysis. RNA was isolated with 1 mL of Trizol reagent (Invitrogen), using manufacturer instructions. RNA was extracted with phenol and with chloroform/isoamyl alcohol (24 : 1), precipitated with absolute ethanol and washed with 70% ethanol. Finally, RNA was resuspended in 30 μL of RNase-free water.

To isolate RNA under acidic conditions, we used the method described by Varshney et al. (1991). Briefly, cells from a 25-mL culture were collected by filtration and resuspended in 300 μL of 0.3 M sodium acetate (pH 4.5) and 10 mM EDTA and subjected to two extractions with phenol equilibrated with the same buffer. The aqueous phase was then precipitated with absolute ethanol and resuspended in 60 μL of 0.3 M sodium acetate (pH 4.5) and 1 mM EDTA. The RNA was again precipitated with absolute ethanol and resuspended in the same buffer.


A total of 10 μg of total RNA was treated with 2 units of RQ1 DNase (Promega), in 20 μL, for 1 h at 37 °C. Reaction was stopped with 2 μL of the stop buffer provided and heated for 10 min at 70 °C. 5 μg of this treated RNA was used for reverse transcription (RT) using 3 μg of random hexamer primers and 100 units of Superscript-II Reverse Transcriptase (Invitrogen), following manufacturer instructions. A control reaction, in which the RT enzyme was omitted, was included to rule out the amplification of contaminant DNA. PCR was performed using 1 μL of the generated cDNA, and 30 PCR cycles were performed with primers F1 to F7 and R1 to R7 (Supporting information, Table S1). PCR products were visualized in a 2% agarose gel.

In vitro processing of pre-tRNAs by RNase P and RNase Z

32P-labelled pre-tRNA substrates for RNase P and RNase Z processing assays were generated with T7 RNA polymerase from plasmids pSer, pYSR and pNQQ (Fig. 2). These plasmids contain, in pUC19, the indicated pre-tRNA(s) obtained by PCR from genomic DNA with appropriate oligonucleotides (Table S1) cloned downstream of a T7 promoter. For run-off in vitro transcription, pSer and pNQQ were digested with HindIII, and pYSR was digested with NruI.

Figure 2.

Processing of pre-tRNAs by RNase P and RNase Z. (a) Diagram of the pre-tRNAs used. tRNAs are labelled as in Fig. 1. Arrows indicate the expected cleavage sites for RNase P (P) or RNase Z (Z). The 3′ side of trnN-GUU(3) and trnQ-CUG is not expected to be substrates of RNase Z because these tRNA genes encode CCA. (b) Autoradiography of an 8% polyacrylamide sequencing gel. M, DNA sequencing ladder used as size marker; 1, pre-tRNA incubated in RNase Z assay buffer without enzyme; 2, pre-tRNA incubated with purified RNase Z from Synechocystis sp. PCC6803; 3, pre-tRNA incubated in RNase P ribozyme assay buffer without enzyme; 4, pre-tRNA incubated with P RNA from Anabaena 7120; 5, pre-tRNA incubated with reconstituted RNase P holoenzyme from Anabaena 7120; 6, pre-tRNA incubated with purified P protein from Anabaena 7120; 7, pre-tRNA incubated in RNase P holoenzyme assay buffer without enzyme. Black dots indicate the pre-tRNAs; white dots, the RNase P or RNase Z products; arrowheads, partially processed pre-tRNAs. See text for details.

RNA subunit of RNase P (P RNA) from Anabaena 7120 was obtained by in vitro transcription as described (Pascual & Vioque, 1999). The gene encoding Anabaena 7120 protein subunit of RNase P (P protein) (alr3413) was amplified by PCR with oligonucleotides all3413F1 and all3413R1 (Table S1) from Anabaena 7120 genomic DNA and cloned into pET28 (Novagen) in phase with a hexahistidine tag at the amino end. The protein was overexpressed in BL21(DE3) cells and purified by chromatography on a HiTrap chelating column followed by HiTrap CM-Sepharose column (GE Healthcare). RNase P holoenzyme was reconstituted as described for the Synechocystis enzyme (Pascual & Vioque, 1996). RNase P assays were performed under single turnover conditions essentially as described (Pascual & Vioque, 1999). Synechocystis RNase Z was purified and assayed as described (Ceballos-Chávez & Vioque, 2005).

Determination of the tRNA aminoacylation state

To identify aminoacylated tRNAs, we used the OXOPAP assay (Gaston et al., 2008). Briefly, RNA was isolated from Anabaena 7120 under acidic conditions as described previously, and 5 μg of total RNA was treated with sodium m-periodate to oxidize the 3′ ends of free tRNAs. Subsequently, the samples were deacylated, resulting in a population in which only aminoacylated tRNAs carry a 3′-OH suitable for polyadenylation. The samples were then polyadenylated and analysed by RT-PCR with an oligoT anchor (OXOPAPRTR) and oligos specific for each of the tRNAs being analysed (Table S1). A control reaction was always included, in which aminoacyl-tRNAs were deacylated before oxidation and therefore were not suitable for polyadenylation and RT-PCR.

Preparation of Anabaena 7120 tRNASer by in vitro transcription

The trnS-GCU(1) and trnS-GCU(2) genes were amplified by PCR from Anabaena 7120 and cloned in pGEM-T and pUC19 vectors, respectively. Both vectors were digested with MvaI (Takara) to produce a linear template for transcription with T7 RNA polymerase that generates the mature full-length tRNA containing the 3′ CCA sequence. Transcription was carried out in 50 μL as described (Pascual & Vioque, 1999). After transcription, tRNA transcripts were separated on 8% polyacrylamide and 7 M urea gel, and the corresponding bands were visualized by UV shadowing. Bands were excised from the gel, and the RNAs were eluted overnight in 10 mM Tris–HCl (pH 7.5), 0.01% SDS, 1 mM EDTA (pH 8.0) and 100 mM NaCl. Eluted RNAs were ethanol precipitated and resuspended in RNase-free water. Before using, RNAs were allowed to refold at 37 °C (10 min) after denaturation at 65 °C (10 min).

In-line probing

Approximately, 30–40 pmol of RNA prepared by in vitro transcription were dephosphorylated with alkaline phosphatase (Roche) and radiolabelled with [γ-32P]-ATP using T4 polynucleotide kinase (Roche), following protocols supplied by manufacturers. In-line probing reactions were assembled as previously described (Soukup & Breaker, 1999). Briefly, 5000 cpm of radiolabelled RNA were incubated at room temperature for 40 h in a buffer containing 50 mM Tris–HCl (pH 8.3), 100 mM KCl and 20 mM MgCl2. Samples were loaded on a high-resolution 8% polyacrylamide and 7 M urea gel and imaged using a Cyclone Storage Phosphor System (Packard).

Aminoacylation of tRNAs

Aminoacylation of in vitro-transcribed tRNAs was carried at 30 °C as described (Schulze et al., 2006). 1.3 μM tRNA, 0.5 μg μL−1 Anabaena crude extract and 25 μM of radioactive amino acid ([14C]-serine or [14C]-glutamate) were mixed in a buffer containing 50 mM HEPES (pH 7.5), 25 mM KCl, 15 mM MgCl2 and 5 mM DTT. Reactions were started by addition of 5 mM ATP. Samples were taken at different times and precipitated with 100 μL of 20% (w/v) trichloroacetic acid at 4 °C for 10 min and then were spotted on a nitrocellulose filter (0.45 μm HAWP; Millipore). The filters were washed sequentially with 10 % (w/v) trichloroacetic acid, 5% (w/v) trichloroacetic and 100% ethanol and were left to dry. Radioactivity in the filters was quantified by liquid scintillation.

Results and discussion

The tRNA cluster in the delta plasmid of Anabaena 7120

The delta plasmid of Anabaena 7120 contains a cluster of 26 tRNA genes or pseudogenes (Fig. 1). Twenty-two of them are annotated in the Cyanobase between coordinates 49 998 and 51 899 of the 55 414-bp delta plasmid. We found several additional tRNA genes and pseudogenes in the cluster by searching with tRNAscan-SE with the COVE only option (Schattner et al., 2005). The tRNAs encoded in the cluster are redundant with chromosomal tRNAs, except for tRNAGlnCUG and tRNAGluCUC, which are not present in the chromosome. tRNAGlnUUG and tRNAGluUUC normally have the position U34 modified, allowing decoding of both glutamine codons (CAA and CAG) or glutamate codons (GAA and GAG), respectively (Agris et al., 2007). Therefore, tRNAGlnCUG and tRNAGluCUC are not required for protein synthesis. In fact, most cyanobacteria have only the tRNAGlnUUG and tRNAGluUUC genes and lack tRNAGlnCUG and tRNAGluCUC. Eight of the tRNA genes present in the cluster encode the 3′-end CCA sequence, which is also unusual as very few cyanobacterial tRNA genes encode CCA. We were thus interested in analysing the function of the tRNAs in this cluster. In particular, we have analysed whether these RNAs were processed correctly and aminoacylated. We have studied one of them, trnS-GCU(2), in more detail.

Transcription and processing of tRNAs in the delta plasmid

Expression of the tRNAs in the delta plasmid was analysed by RT-PCR with a set of primers designed to generate overlapping fragments encompassing the whole tRNA cluster (Fig. 1a and b). RT-PCR products were detected for all primer pairs used, indicating that the cluster is transcribed as a single RNA. However, no full-length RNA could be detected with primers F7 and R1, suggesting quick processing of the primary transcript. The presence of a single active promoter upstream of the tRNA cluster has been confirmed recently by RNASeq (Mitschke et al., 2011).

Individual tRNAs were also detected by Northern blot (Fig. 1c). The sizes of the bands were the expected for correctly processed tRNAs. Correct 5′ ends were confirmed by primer extension for tRNASerGCU(2) and tRNAGlnCUG (not shown). In addition to tRNAs, an RNA corresponding to an intergenic region (Int2) was also detected by Northern blot, indicating stable accumulation of this RNA, generated after processing of the flanking tRNAs.

To study tRNA processing within the cluster, we prepared three different pre-tRNAs by in vitro transcription (Fig. 2a). These precursors were incubated with purified RNase Z from Synechocystis (Ceballos-Chávez & Vioque, 2005), which would cleave at the 3′ side of CCA-lacking pre-tRNAs. In the three cases, the expected processing products were detected. No products corresponding to cleavage at the 3′ ends of the CCA-encoding tRNAAsnGUU(3) and tRNAGlnCUG by RNase Z were observed (Fig. 2b), confirming the previously described inhibition of cyanobacterial RNase Z activity by the presence of CCA at the 3′-end of tRNAs (Ceballos-Chávez & Vioque, 2005). The pre-tRNAs were also incubated with the RNA subunit of the Anabaena 7120 RNase P in a high-salt buffer, reaction conditions appropriate for catalytic activity of the RNase P RNA in the absence of the protein cofactor (Pascual & Vioque, 1999), as well as with the complete RNase P holoenzyme in low-salt buffer. In both cases, the expected products were detected for all three pre-tRNAs (Fig. 2c). These results indicate that there is no specific cleavage order for RNase P and RNase Z, because both RNases can generate the expected final products.

Aminoacylation of tRNAs

The results described previously indicate that the tRNAs encoded in the cluster are expressed and processed to mature tRNAs. We next analysed whether they were aminoacylated in vivo. For this purpose, we used the OXOPAP method (Gaston et al., 2008). We could detect aminoacylation for most tRNAs encoded in the cluster (Fig. 3), including several classified as pseudogenes by tRNAscan-SE: tRNASerGCU(2), inline image and tRNAArgUCU(2). Also, tRNAs whose genes contain the CCA sequence were aminoacylated [tRNAGlnCUG, tRNALeuCAA(2), tRNALysUUU(2), inline image, tRNAValUAC(2)]. This confirms that CCA-containing pre-tRNAs are processed correctly at the 3′ side in vivo to generate mature functional tRNAs, despite the inability of RNase Z to carry out the reaction in our in vitro assay. Possibly, a 3′ to 5′ exonuclease could generate the mature 3′ end of these tRNAs, similarly to what happens in Escherichia coli. In Anabaena 7120, there are homologues of RNase PH and RNase D that could be involved in 3′ maturation of CCA-containing tRNAs. The presence of these CCA-encoding tRNA genes in Anabaena 7120, which are correctly processed in vivo, provides a tool to investigate the function of these exonucleases, so far uncharacterized in cyanobacteria, in tRNA processing.

Figure 3.

Aminoacylation of tRNAs encoded in the delta plasmid. RT-PCR products obtained following the OXOPAP procedure with primers corresponding to the indicated tRNAs. Total RNA extracted under acidic conditions was subjected to periodate oxidation before (1) or after deacylation (2). After polyadenylation, samples were subjected to RT with oligonucleotide OXOPAPRTR, and PCR with the same oligonucleotide and a forward primer specific of each tRNA (Table S1). +, complete RT-PCR reaction; −, reaction performed in the absence of reverse transcriptase. The arrows indicate the specific RT-PCR product obtained from aminoacylated tRNA, which were confirmed by sequencing in each case. M, size markers (bp).

Characterization of tRNASerGCU(2)

tRNASerGCU(2) has a structure that deviates from consensus (Fig. 4) and is classified by tRNAscan-SE as a pseudogene. The T-stem has a U–U mismatch; position 9 is a U instead of the conserved purine, and the D-loop is smaller than usual. However, tRNASerGCU(2), as shown previously, is correctly processed and is aminoacylated in vivo, indicating that its overall shape must be tRNA-like to be recognized by processing endonucleases and aminoacyl-tRNA synthetases. We have compared the structure of tRNASerGCU(2) with the chromosomally encoded tRNASerGCU(1) by in-line probing (Soukup & Breaker, 1999). Positions more susceptible to spontaneous hydrolysis are mainly in the anticodon and in the variable stem–loop, as expected according to the tridimensional L-shaped structure of tRNAs. tRNASerGCU(2) has also hydrolysis susceptibility in the T-stem, indicating that the T-stem is less stable than in tRNASerGCU(1), as expected by the presence of a U–U mismatch. In addition, there are hydrolysis susceptibility sites in the T-loop, indicating that the interaction between the T-loop and D-loop that stabilizes the L-shape of the tRNA is weaker in tRNASerGCU(2).

Figure 4.

In-line probing of tRNASerGCU(1) and tRNASerGCU(2). Left, autoradiography of a 8% polyacrylamide sequencing gel where 5′-labelled tRNASerGCU(1) (1) or tRNASerGCU(2) (2) was loaded without previous incubation (−), or after incubation for 40 h at pH 8.3 (+). T1, RNase T1 ladder generated from tRNASerGCU(1). OH−, Alkali ladder generated from tRNASerGCU(1). Right, diagram of the predicted structures of tRNASerGCU(1) (top) and tRNASerGCU(2) (bottom). Arrowheads indicate positions more sensitive to spontaneous hydrolysis.

We have also compared the aminoacylation of tRNASerGCU(1) and tRNASerGCU(2) by an Anabaena 7120 crude extract in vitro (Fig. 5). Both tRNAs are aminoacylated with similar efficiency with serine (Fig. 5a) and are not aminoacylated with a noncognate amino acid such as glutamate (Fig. 5b).

Figure 5.

In vitro aminoacylation of tRNAs. (a) Aminoacylation assay of tRNASerGCU(1) (circles), tRNASerGCU(2) (squares) and Yfr-1 RNA (triangles) with [14C]-serine. (b) Aminoacylation assay of the same RNAs with [14C]-glutamate. Yfr-1 is an abundant noncoding RNA of 65 nucleotides present in Anabaena 7120 (Voss et al., 2007). Values shown are the mean of three independent experiments, with error bars representing the standard deviation.

Origin, evolution and function of the tRNA cluster

Diverse functions have been ascribed to the organization of tRNA genes in clusters, such as to coordinate transcription and processing, coordinate the amount of tRNA with translation rates, etc. (Rudner et al., 1993). In DNA viruses, they apparently help adjust translation rate during infection (Dreher, 2010). In yeast, tRNA genes are spatially clustered in the nucleolus, even though they are dispersed in the linear genome (Thompson et al., 2003), also an indication that clustering could be advantageous and therefore selected for in some circumstances.

To inquire about the function of the tRNA cluster, we have generated a mutant strain in which the tRNA cluster was completely replaced by an antibiotic resistance marker. The mutant could be fully segregated and showed no apparent phenotypic differences with wild type under standard growth conditions in media with nitrate or in media lacking combined nitrogen, confirming that the tRNAs encoded in the cluster are not required under normal conditions. The possibility remains that these tRNAs are required or confer some advantage under some unknown environmental conditions. Therefore, their function, if any, remains to be elucidated.

To inquire about the possible origin of the tRNA cluster present in the delta plasmid of Anabaena 7120, we have searched the sequenced genomes of cyanobacteria for similar clusters. We have identified tRNA clusters similar to the one in the delta plasmid of Anabaena 7120 in the chromosomes of Nostoc punctiforme PCC73102, Acaryochloris marina MBIC11017 and Oscillatoria sp. PCC6506 (Fig. 6). However, a similar cluster was not present in Anabaena variabilis ATCC 29413, a strain very closely related to Anabaena 7120. The four clusters are clearly related and have a common origin, with the same order of the tRNA genes. The differences between the four clusters can be explained by differential losses of individual tRNA genes, although some cases of tRNA identity change cannot be excluded. In addition, in A. marina and Oscillatoria sp. PCC6506, there are insertions that interrupt the clusters. These insertions contain ORFs that are unrelated between the two strains, and no homologues are detected by blast, except in the one closer to the 3′ side, between trnT and trnG, which contains the same gene in both strains, encoding an AraC family regulator that is more closely related to similar proteins in other bacteria than to any cyanobacterial protein. Sequence analysis of the tRNAs from the clusters strongly supports their specific relationship. There are four or five tRNALeu genes in each of the four clusters. They all have an unusually short variable region (Fig. S1) that is found only in some tRNALeu genes from actinobacteria but never in cyanobacteria (Juhling et al., 2009). In addition, phylogenetic analysis of the tRNALeu genes groups together with high confidence the tRNAs from the clusters to the exclusion of the other tRNALeu genes in the genomes of the four cyanobacteria (Fig. S2). Taken together, these results support the hypothesis that the tRNA cluster was acquired by horizontal transfer from another organism either at the common ancestor of these four strains, with subsequent differential losses, or as independent events.

Figure 6.

Synteny of the tRNA clusters present in four cyanobacteria. tRNA genes (white) or pseudogenes (grey) identified by tRNAscan-SE are labelled by their amino acid identity. The black triangles in Acaryochloris marina and Oscillatoria sp. clusters indicate the insertion of sequences whose size is shown on top of the triangles.


This work was supported by Ministerio de Ciencia e Innovación and the European Regional Fund (BFU2007-60651) and Plan Andaluz de Investigación (BIO215). L.P.-G. was supported by a predoctoral fellowship from Ministerio de Ciencia e Innovación. We are grateful to Alicia M. Muro-Pastor for critical reading.