Characterization of 3′-Phosphate RNA Ligase Paralogs RtcB1, RtcB2, and RtcB3 from Myxococcus xanthus Highlights DNA and RNA 5′-Phosphate Capping Activity of RtcB3

ABSTRACT

Escherichia coli RtcB exemplifies a family of GTP-dependent RNA repair/splicing enzymes that join 3′-PO₄ ends to 5′-OH ends via stable RtcB-(histidinyl-N)-GMP and transient RNA_3′pp_5′G intermediates. E. coli RtcB also transfers GMP to a DNA 3′-PO₄ end to form a stable “capped” product, DNA_3′pp_5′G. RtcB homologs are found in a multitude of bacterial proteomes, and many bacteria have genes encoding two or more RtcB paralogs; an extreme example is Myxococcus xanthus, which has six RtcBs. In this study, we purified, characterized, and compared the biochemical activities of three M. xanthus RtcB paralogs. We found that M. xanthus RtcB1 resembles E. coli RtcB in its ability to perform intra- and intermolecular sealing of a _HORNAp substrate and capping of a DNA 3′-PO₄ end. M. xanthus RtcB2 can splice _HORNAp but has 5-fold-lower RNA ligase specific activity than RtcB1. In contrast, M. xanthus RtcB3 is distinctively feeble at ligating the _HORNAp substrate, although it readily caps a DNA 3′-PO₄ end. The novelty of M. xanthus RtcB3 is its capacity to cap DNA and RNA 5′-PO₄ ends to form GppDNA and GppRNA products, respectively. As such, RtcB3 joins a growing list of enzymes (including RNA 3′-phosphate cyclase RtcA and thermophilic ATP-dependent RNA ligases) that can cap either end of a polynucleotide substrate. GppDNA formed by RtcB3 can be decapped to pDNA by the DNA repair enzyme aprataxin.

IMPORTANCE RtcB enzymes comprise a widely distributed family of RNA 3′-PO₄ ligases distinguished by their formation of 3′-GMP-capped RNAppG and/or DNAppG polynucleotides. The mechanism and biochemical repertoire of E. coli RtcB are well studied, but it is unclear whether its properties apply to the many bacteria that have genes encoding multiple RtcB paralogs. A comparison of the biochemical activities of three M. xanthus paralogs, RtcB1, RtcB2, and RtcB3, shows that not all RtcBs are created equal. The standout findings concern RtcB3, which is (i) inactive as an RNA 3′-PO₄ ligase but adept at capping a DNA 3′-PO₄ end and (ii) able to cap DNA and RNA 5′-PO₄ ends to form GppDNA and GppRNA, respectively. The GppDNA and GppRNA capping reactions are novel nucleic acid modifications.

INTRODUCTION

Escherichia coli RtcB is a founding member of a recently discovered family of RNA repair/splicing enzymes that join RNA 2′,3′-cyclic-PO₄ or 3′-PO₄ ends to RNA 5′-OH ends (1–4). RtcB executes a four-step pathway that requires GTP as an energy source and Mn²⁺ as a cofactor (5–7). RtcB first reacts with GTP to form a covalent RtcB-(histidinyl-N)-GMP intermediate. It then hydrolyzes the RNA 2′,3′-cyclic-PO₄ end to a 3′-PO₄ end and transfers guanylate to the RNA 3′-PO₄ end to form an RNA_3′pp_5′G intermediate. Finally, RtcB catalyzes the attack of an RNA 5′-OH on the RNA_3′pp_5′G end to form the 3′-5′ phosphodiester splice junction and liberate GMP. The unique chemical mechanism of RtcB overturned a longstanding canon of nucleic acid enzymology, which held that the synthesis of polynucleotide 3′-5′ phosphodiesters proceeds via the attack of a 3′-OH on a high-energy 5′-phosphoanhydride, either a nucleoside 5′-triphosphate in the case of RNA/DNA polymerases or an adenylylated intermediate A_5′pp_5′N- in the case of classic RNA/DNA ligases.

The wide distribution of RtcB proteins in bacteria, archaea, and metazoa and the fact that crystal structures of thermophilic archaeal (8–10) and bacterial (Protein Data Bank ID 2EPG) RtcBs unveiled RtcB to have a novel fold and active site, with no similarity whatsoever to classic RNA/DNA ligases, highlighted RtcB as a harbinger of an alternative enzymology based on covalently activated 3′-PO₄ ends (6). After speculating that the chemistry of 3′-PO₄/5′-OH end joining by RtcB might be applicable to DNA transactions, Das et al. (11) showed that E. coli RtcB can indeed modify DNA 3′-PO₄ ends. Specifically, RtcB transfers GMP from a covalent RtcB-GMP intermediate to a DNA 3′-PO₄ to form a chemically stable “capped” 3′-end structure, DNA_3′pp_5′G.

The implications of RtcB-mediated DNA capping for nucleic acid break repair are potentially significant, given the many biological settings in which damage generates 3′-PO₄ ends that are refractory to the action of DNA polymerases and classic ligases. Biochemical studies of the impact of DNA 3′ capping on repair reactions highlighted two key points: (i) the cap protects DNA ends from 3′ exonucleases, and (ii) the cap guanosine 3′-OH can serve as a primer for templated DNA synthesis by exemplary members of five different DNA polymerase families (12, 13).

RtcB homologs are present in the proteomes of scores of bacterial taxa, even though bacteria have no evident need for protein-catalyzed tRNA splicing or mRNA splicing, which are the functions ascribed to metazoan RtcB homologs during the maturation of intron-containing tRNAs and during nonspliceosomal splicing of XBP1 mRNA during the unfolded protein response to endoplasmic reticulum (ER) stress (2, 14–17). In E. coli, RtcB is encoded on an operon with the RNA 3′-phosphate cyclase enzyme RtcA (18). The fact that the rtcBA operon in E. coli is regulated by the σ⁵⁴ coactivator RtcR (encoded by the rtcR gene, which is located immediately upstream of rtcBA and transcribed in the opposite orientation) suggested that the RNA repair functions are induced in response to cellular stress (18), conceivably as a means to recover from RNA damage inflicted by bacterial ribotoxins that are turned on in stress situations. This scenario remains speculative, because the stress signal that activates RtcR is unknown, and the deletion of rtcBA elicits no overt phenotype in E. coli (18).

Based on the biochemical activity of E. coli RtcB on RNA and DNA substrates, there is no basis to restrict our ideation about RtcB function to RNA-versus-DNA transactions. Indeed, the fact that some bacteria contain genes encoding multiple RtcB paralogs raises the prospect that bacterial RtcB-type ligases might acquire functional specialization for particular repair pathways or for the sealing of specific nucleic acid substrates. By analogy, many bacteria that have genes coding for multiple DNA ligase enzymes have evolved a division of labor whereby one ligase serves an essential replicative function (e.g., sealing of Okazaki fragments), while a different ligase is dedicated to double-strand break repair via nonhomologous end joining (19). The limitation to this line of thinking regarding RtcB is that the E. coli enzyme is the only bacterial RtcB that has been characterized to date.

The goal of the present study was to gauge the activities of paralogous RtcB proteins from a single bacterial species. We chose to focus on Myxococcus xanthus by virtue of the presence of six predicted RtcBs in its proteome. Here, we report the production, purification, and characterization of recombinant M. xanthus RtcB1, RtcB2, and RtcB3.

MATERIALS AND METHODS

Recombinant M. xanthus RtcB proteins.

The rtcB open reading frames (ORFs) encoding RtcBs 1 to 6 were amplified by PCR from M. xanthus genomic DNA with primers designed to introduce a BamHI site flanking the predicted ATG start codon and an NotI site immediately downstream of the stop codon. The PCR products were digested with BamHI and NotI and then inserted between the BamHI and NotI sites of pET28b-His₁₀Smt3. The pET28b-His₁₀Smt3-RtcB plasmids were transformed into E. coli BL21-CodonPlus(DE3). One-liter cultures derived from single transformants were grown at 37°C in LB medium containing 50 μg/ml kanamycin until the A₆₀₀ reached 0.6 to 0.8. The cultures were adjusted to 2% (vol/vol) ethanol, chilled on ice for 30 min, and then adjusted to 0.1 mM isopropyl-β-d-thiogalactoside. Incubation was continued at 17°C for 16 h with constant shaking. Cells were harvested by centrifugation and stored at −80°C. All subsequent procedures were performed at 4°C. The cell pellets were suspended in 50 ml of buffer A (50 mM Tris-HCl [pH 8.0], 350 mM NaCl, 10% sucrose, 10% glycerol). After the addition of lysozyme at 0.2 mg/ml and a protease inhibitor tablet (Roche), the suspensions were rocked for 1 h. The lysates were sonicated to reduce viscosity, and insoluble material was removed by centrifugation at 20,000 × g for 30 min. The soluble lysates were mixed for 1 h with 4 ml of His60 Ni Superflow resin (Clontech) that had been equilibrated in buffer A. The resins were recovered by centrifugation and resuspended in 25 ml of buffer B (50 mM Tris-HCl [pH 8.0], 350 mM NaCl, 10% glycerol) containing 25 mM imidazole. The cycle of centrifugation and resuspension of the resin in buffer B was repeated three times, after which the washed resins were poured into columns. The bound material was eluted stepwise with 8-ml aliquots of 50, 100, 200, 300, 400, and 500 mM imidazole in buffer B. The elution profiles were monitored by SDS-PAGE. The eluate fractions containing each His₁₀Smt3-RtcB complex were pooled and digested with 90 μg of the Smt3-specific protease Ulp1 during overnight dialysis against buffer B. The dialysates were mixed for 1 h with 1.5 ml of His60 Ni Superflow resin that had been equilibrated in buffer B and then poured into columns. After washing the columns with buffer B, the flowthrough and wash fractions containing each tag-free RtcB were pooled, adjusted to 25 mM EDTA, and then concentrated by centrifugal ultrafiltration (Amicon Ultra-15, 10-kDa cutoff; Millipore). The RtcB proteins were then gel filtered through a 120-ml 16/60 HiLoad Superdex 200-pg column (GE Healthcare) equilibrated with buffer C (10 mM Tris-HCl [pH 8.0], 350 mM NaCl, 1 mM dithiothreitol [DTT], 1 mM EDTA, 10% glycerol). The peak RtcB-containing fractions were pooled and concentrated by centrifugal ultrafiltration. Protein concentration was determined by using the Bio-Rad dye reagent with bovine serum albumin as the standard. The yields of purified RtcB1, RtcB2, and RtcB3 proteins from 1-liter cultures were 0.2 mg, 0.6 mg, and 0.6 mg, respectively.

RNA and DNA substrates.

A 20-mer _HORNA_3′p strand labeled with ³²P at the penultimate phosphate (see Fig. 2) was prepared by T4 Rnl1-mediated addition of 5′ [³²P]pCp to a 19-mer synthetic oligoribonucleotide, as described previously (5). 5′ ³²P-labeled pDNAp, pDNA_OH, and pRNA_OH strands were prepared by enzymatic phosphorylation of synthetic oligonucleotides using [γ-³²P]ATP and a phosphatase-dead mutant of T4 polynucleotide kinase (Pnkp-D167N). The radiolabeled strands were gel purified prior to use in enzyme assays.

FIG 2.

RNA 3′-PO₄/5′-OH ligase activity. (A) Recombinant RtcB1, RtcB2, and RtcB3 proteins. Aliquots (5 μg) of the indicated protein preparations were analyzed by SDS-PAGE. The Coomassie blue-stained gel is shown. The positions and sizes (in kilodaltons) of marker polypeptides are indicated on the left. (B) RNA ligase reaction mixtures (10 μl) containing 50 mM Tris-HCl (pH 8.0), 2 mM MnCl₂, 0.1 mM GTP, 1 pmol (0.1 μM) 3′-³²P-labeled 20-mer _HORNAp substrate (depicted at the bottom of the panel with the ³²P label denoted by ●), and 0.2, 0.6, 2, 6, or 20 pmol RtcB protein (0.02, 0.06, 0.2, 0.6, or 2 μM RtcB, proceeding from left to right in each titration series) were incubated for 10 min at 37°C. RtcB was omitted from a control reaction in the first lane (–). The reactions were quenched with an equal volume of 90% formamide–50 mM EDTA. The products were analyzed by electrophoresis through a 40-cm 20% polyacrylamide gel containing 7.5 M urea in 90 mM Tris-borate and 2 mM EDTA. An autoradiograph of the gel is shown. The positions and identities of the radiolabeled RNAp substrate and the various sealed products are indicated on the left. (C) The substrate and ligated products were quantified by scanning the gel with a Fujix BAS2500 imager. The extents (percentages) of ligation, calculated as [(circle + multimers)/(circle + multimers + RNAp)] × 100, are plotted as a function of input RtcB. Each datum is the average ± standard error of the mean (SEM) of the results from four separate titration experiments.

RESULTS

The M. xanthus genome encodes six RtcB paralogs.

M. xanthus is a deltaproteobacterium of the order Myxococcales that is distinguished by its social behavior and starvation-induced differentiation into a multicellular fruiting body. The 9.14-Mbp single-chromosome genome of M. xanthus strain DK1622 encodes 7,388 predicted proteins, 3,532 of which comprise 872 paralogous sets (defined as having at least two members) (20). It is suggested that selective gene duplications and the divergence of paralogs enabled evolution of the multicellular lifestyle of Myxococcus (20). In this vein, it is notable that the M. xanthus proteome includes six RtcB-like paralogs, which we named as follows: RtcB1 (411 amino acids [aa], GenBank accession no. ABF90081), RtcB2 (384 aa, ABF90287), RtcB3 (385 aa, ABF86333), RtcB4 (474 aa, ABF92494), RtcB5 (497 aa, ABF87339), and RtcB6 (450 aa, ABF85892). Another social Myxococcales species, Sorangium cellulosum, has a 13-Mbp genome that encodes 9,367 predicted proteins (21), including five RtcB paralogs. In contrast, the genomes of other deltaproteobacteria, such as Geobacter metallireducens, Bdellovibrio bacteriovorus, and Desulfovibrio vulgaris, encode a single RtcB homolog.

A constraint-based alignment tool (COBALT) comparison of the amino acid sequences of the M. xanthus RtcBs highlights 77 positions of side chain identity/similarity in all six proteins (Fig. 1). The conserved amino acids embrace virtually all of the components imputed to the RtcB active site based on the crystal structure of a Pyrococcus horikoshii RtcB-GTPαS-(Mn²⁺)₂ complex and mutational analyses of E. coli and P. horikoshii RtcBs (2, 6, 9, 10). These include the following: (i) the histidine nucleophile that forms the covalent RtcB-(histidinyl)-GMP intermediate, (ii) the Asp-Cys-His-His-His pentad that coordinates two manganese ions, (iii) six residues that contact the guanine nucleobase of GTP, (iv) three residues that contact the GTP ribose, and (v) two asparagines that coordinate the GTP phosphates (Fig. 1). A conserved arginine and a lysine correspond to residues in P. horikoshii RtcB that coordinate sulfate anions proposed to mimic RNA phosphates. Also conserved are multiple amino acids that play a structural role, either by coordinating active-site residues (e.g., an aspartate near the N terminus that orients the histidine nucleophile) or by tethering secondary structure elements of the RtcB fold (e.g., via salt bridges).

FIG 1.

The M. xanthus genome encodes six RtcB paralogs. (A) A COBALT-based alignment of the amino acid sequences of the six M. xanthus RtcB paralogs is shown. The positions of amino acid side chain identity/similarity in all six polypeptides are denoted by dots. The histidine nucleophile that becomes covalently linked to GMP is denoted by |. The conserved residues that in the crystal structure of the Pyrococcus RtcB-GTPαS-(Mn²⁺)₂ complex contact the two manganese ions, the guanine nucleobase, ribose, and phosphate moieties of GTP, and sulfate ions (putative mimetics of RNA phosphate groups) are highlighted in color-coded boxes, as explained in the key on the lower right. Conserved structural residues in gray shading are those that either coordinate active-site residues or tether secondary structure elements of the RtcB fold, e.g., via salt bridges.

The six RtcB paralogs are dispersed in the M. xanthus chromosome in six entirely distinct genetic neighborhoods. M. xanthus rtcB1, like E. coli rtcB (18), is located in a ⇐rtcR·rtcB–rtcA⇒ RNA repair operon in which (i) the ORF encoding the RNA ligase RtcB is oriented tail to head with the ORF encoding the RNA 3′-phosphate cyclase enzyme RtcA and (ii) a σ⁵⁴ coactivator, RtcR, is encoded by the rtcR gene, which is located immediately upstream of rtcBA and transcribed in the opposite orientation. M. xanthus rtcB5 is adjacent and distal to a cooriented ORF encoding a protein called archease. Genetic clustering of RtcB or RtcA with archease is common in archaea and is also seen sporadically in bacteria. P. horikoshii and mammalian archeases have been shown to stimulate the ligase activities of P. horikoshii and mammalian RtcBs, respectively (22, 23). The M. xanthus rtcB2, rtcB3, rtcB4, and rtcB6 ORFs have distinctive flanking gene sets.

Recombinant M. xanthus RtcB proteins.

The primary structures of the six M. xanthus RtcBs suggest that they might all be enzymatically active. The questions are what activities are inherent to each paralog and how might they differ? To address these issues, we expressed the six RtcB proteins in E. coli as His₁₀Smt3-RtcB fusions. The RtcB1, RtcB2, and RtcB3 proteins were isolated from soluble bacterial extracts by Ni affinity chromatography. The His₁₀Smt3 tag was removed with the Smt3-specific protease Ulp1, and the native RtcB1, RtcB2, and RtcB3 proteins were separated from the tag by a second round of Ni affinity chromatography. The tag-free RtcBs were further purified by gel filtration. SDS-PAGE of the purified RtcB2 and RtcB3 proteins revealed predominant ∼44-kDa polypeptides, consistent with their similar lengths of 384 and 385 aa, respectively (Fig. 2A). The RtcB1 preparations contained a major ∼48-kDa polypeptide, as expected for the 411-aa RtcB protein, and an ∼73-kDa contaminant (Fig. 2A). The RtcB4, RtcB5, and RtcB6 proteins were produced in E. coli but were recovered exclusively in the insoluble pellet fraction, thereby impeding their purification and characterization.

Test of RNA ligase activity.

We reacted the recombinant M. xanthus RtcBs with a 20-mer _HORNAp substrate that was 3′ ³²P-labeled at the penultimate phosphate (Fig. 2B); this was the same substrate used previously to characterize the RNA sealing reaction of E. coli RtcB (6). When the products were analyzed by denaturing PAGE, we found that M. xanthus RtcB1 efficiently converted the linear substrate to a more rapidly migrating circular RNA species as a consequence of intramolecular ligation; more slowly migrating multimers (products of intermolecular ligation) were also generated (Fig. 2B). The extent of ligation by RtcB1 was proportional to the input enzyme, attaining 98% at saturation (Fig. 2C). RtcB2 catalyzed intra- and intermolecular RNA sealing (Fig. 2B) to an extent of 90% of the input substrate, albeit 5-fold less effectively than did RtcB1 on a per-enzyme basis in the linear range of the titration (Fig. 2C). In contrast, RtcB3 was a notably feeble RNA ligase (Fig. 2B), sealing only 7% of the substrate at the highest level of input enzyme and displaying a specific activity 130-fold lower than that of RtcB1 (Fig. 2C). Thus, the three RtcB paralogs are not equivalent with respect to their 3′-PO₄/5′-OH RNA sealing activities.

Tests of DNA ligase and DNA 3′ capping activities.

A prior study (11) showed that E. coli RtcB could splice DNA 3′-PO₄ and 5′-OH ends in the context of the broken DNA stem-loop structure depicted in Fig. 3. DNA splicing proceeds via a chemically stable DNAppG intermediate formed by the transfer of GMP from RtcB-GMP to the DNAp substrate strand (6, 11). The DNAppG species migrates more slowly than DNAp when analyzed by denaturing PAGE (11). Here, when we tested the M. xanthus RtcB proteins for DNA sealing, we noted that whereas they were all capable of guanylylating the 5′ ³²P-labeled 12-mer pDNAp strand of the stem-loop to form a more slowly migrating capped pDNAppG strand, only RtcB1 and RtcB2 generated detectable amounts of a ligated DNA product (Fig. 3), which comprised 8% and 18% of the radiolabeled DNA at saturating levels of RtcB1 and RtcB2, respectively. RtcB1, RtcB2, and RtcB3 were also adept at capping the 3′-PO₄ end of the pDNAp single-strand substrate (Fig. 4). The RtcB3 reactions with the DNA stem-loop and pDNAp single-strand substrates were distinguished by the formation of a minor radiolabeled species (Fig. 3 and 4, asterisk) that migrated above the pDNAppG strand.

FIG 3.

Test of DNA 3′-PO₄/5′-OH ligase activity. Reaction mixtures (10 μl) containing 50 mM Tris-HCl (pH 8.0), 2 mM MnCl₂, 0.1 mM GTP, 1 pmol (0.1 μM) 5′ ³²P-labeled broken DNA stem-loop substrate (depicted at the bottom with the ³²P label denoted by ●), and 0.2, 0.6, 2, 6, or 20 pmol RtcB protein (0.02, 0.06, 0.2, 0.6, or 2 μM RtcB, respectively, proceeding from left to right in each titration series) were incubated for 10 min at 37°C. RtcB was omitted from a control reaction in the first lane (–). The products were analyzed by urea-PAGE. An autoradiograph of the gel is shown. The positions and identities of the radiolabeled pDNAp substrate strand, the 3′ capped pDNAppG strand, and the ligated DNA stem-loop are indicated on the right. An RtcB3 product migrating above pDNAppG is indicated by an asterisk.

FIG 4.

DNA 3′-PO₄ capping activity. Reaction mixtures (10 μl) containing 50 mM Tris-HCl (pH 8.0), 2 mM MnCl₂, 0.1 mM GTP, 1 pmol (0.1 μM) 5′ ³²P-labeled 12-mer pDNAp substrate (depicted at the bottom with the ³²P label denoted by ●), and 0.2, 0.6, 2, 6, or 20 pmol RtcB protein (0.02, 0.06, 0.2, 0.6, or 2 μM RtcB, respectively, proceeding from left to right in each titration series) were incubated for 10 min at 37°C. RtcB was omitted from a control reaction in the first lane (–) of each panel. The products were analyzed by urea-PAGE. An autoradiograph of the gel is shown. The positions and identities of the radiolabeled pDNAp substrate strand and the 3′ capped pDNAppG strand are indicated on the right. An RtcB3 product that migrated above pDNAppG is indicated by an asterisk.

RtcB3 can cap DNA 3′-PO₄ and 5′-PO₄ ends.

In the experiment shown in Fig. 5A, 1 μM RtcB was reacted with 0.1 μM 5′ ³²P-labeled 12-mer pDNAp strand; aliquots were withdrawn at 0.5, 1, 2, 5, 10, and 20 min and split into two halves that were either mock treated or digested with calf intestine alkaline phosphatase (CIP) before the samples were analyzed by urea-PAGE. In the control reaction, the pDNAp strand was converted into two electrophoretically resolved products (Fig. 5A, left). As expected, CIP treatment eliminated the radiolabeled pDNAp substrate strand by hydrolyzing the [³²P]phosphomonoester (Fig. 5A, right). CIP treatment also eliminated the majority of the radiolabel in the “lower” product (Fig. 5A, right), signifying that the predominant species was 3′ capped pDNAppG. The instructive finding was that the “upper” radiolabeled product was refractory to CIP, i.e., the labeled 5′-PO₄ was guanylylated by RtcB3 to form a doubly capped strand, GppDNAppG. A second CIP-resistant species, which migrated faster than GppDNAppG and slightly slower than pDNAppG, corresponded to a GppDNA_OH strand generated after CIP hydrolyzed the 3′-PO₄ of a 5′ capped GppDNAp RtcB3 reaction product. Thus, the lower product in the untreated control reaction actually comprised a mixture of comigrating pDNAppG and GppDNAp strands, with pDNAppG predominating. (Note that CIP conversion of 3′-PO₄ to 3′-OH is expected to slow the electrophoretic mobility of an otherwise identical DNA strand.) The kinetic profile of the RtcB3 reaction suggests that the ends are capped sequentially, with most events entailing 3′ capping prior to 5′ capping (Fig. 5A and B).

FIG 5.

5′-PO₄ capping activity of RtcB3. (A) Reaction mixtures containing 50 mM Tris-HCl (pH 8.0), 2 mM MnCl₂, 0.1 mM GTP, 0.1 μM 5′ ³²P-labeled 12-mer pDNAp substrate (depicted below the gel, with the ³²P label denoted by ●), and 1 μM RtcB3 were incubated at 37°C. Aliquots (10 μl) were withdrawn at the times specified and mixed with 8 μl of 2.5× CIP buffer (125 mM NaCl, 25 mM MgCl₂, 5 mM ZnCl₂). The samples were split into equal aliquots, which were then incubated at 37°C for 30 min either with no further additions (series at left) or after supplementation with 1 μl (10 U) calf intestine alkaline phosphatase (+CIP series at right). The products were analyzed by urea-PAGE. An autoradiograph of the gel is shown. The positions and identities of the radiolabeled pDNAp substrate strand and the singly and doubly capped strands are indicated on the left and right, respectively. (B) The product distributions from no-CIP control reactions are plotted as a function of time. Each datum is the average ± SEM of the results from three separate experiments.

5′ capping of pDNA_OH and pRNA_OH by RtcB3.

To focus exclusively on 5′ capping and to query whether 5′ capping is contingent on prior 3′ capping, we reacted RtcB3 with a 5′ ³²P-labeled 12-mer pDNA_OH strand and observed the conversion of pDNA to a single slower-migrating GppDNA product, the yield of which was optimal in Tris buffer at pH 7.0 to 7.5 (Fig. 6A). No 5′ capped product was formed when either GTP or manganese was omitted from the reaction mixture (not shown). In the experiments shown in Fig. 6B, we reacted RtcB3 with a series of 5′ ³²P-labeled 24-mer, 12-mer, and 6-mer pDNA_OH strands and tracked the kinetic profile of the conversion of pDNA to the GppDNA product. The rate and yield of 5′ capping were virtually identical for the 24-mer and 12-mer pDNA substrates; the extent of capping of the 6-mer pDNA was reduced slightly (Fig. 6B). A 12-mer pRNA_OH substrate was capped half as well as a 12-mer pDNA with an identical nucleotide sequence (Fig. 6B). We conclude that RtcB performs 5′-PO₄ polynucleotide capping without requiring a 3′-PO₄ end or its guanylylation.

FIG 6.

DNA and RNA 5′-PO₄ capping by RtcB3. (A) Reaction mixtures (10 μl) containing either 50 mM Tris-acetate (pH 4.5, 5.0, 5.5, 6.0, or 6.5) or Tris-HCl (pH 7.0, 7.5, 8.0, 8.5, or 9.0), 2 mM MnCl₂, 0.1 mM GTP, 1 pmol (0.1 μM) 5′ ³²P-labeled 12-mer pDNA_OH substrate (depicted in panel B), and 10 pmol (1 μM) RtcB3 were incubated at 37°C for 20 min. The products were analyzed by urea-PAGE. An autoradiograph of the gel is shown. The positions and identities of the radiolabeled pDNA substrate strand and the 5′ capped strand are indicated on the left. (B) Reaction mixtures containing 50 mM Tris-HCl (pH 7.5), 2 mM MnCl₂, 0.1 mM GTP, either 0.1 μM 5′ ³²P-labeled 24-mer, 12-mer, or 6-mer pDNA_OH substrates or 0.1 μM 5′ ³²P-labeled 12-mer pRNA_OH substrate (with nucleotide sequences as shown), and 1 μM RtcB3 were incubated at 30°C. Aliquots (10 μl) were withdrawn at the times specified, and reactions were quenched with formamide-EDTA. The products were analyzed by urea-PAGE and quantified by scanning the gel. The extents of 5′ capping are plotted as a function of time for each substrate. Each datum is the average ± SEM of the results from three separate experiments.

The capped 12-mer DNA product formed by RtcB3 was gel purified to radiochemical homogeneity (Fig. 7A, compare lanes M and 0) and subjected to digestion with either nuclease P1 alone or sequential digestions with nuclease P1 and nucleotidyl pyrophosphatase (NPPase). The intact capped 12-mer and the digestion products were analyzed by polyethyleneimine (PEI)-cellulose thin-layer chromatography (TLC) (Fig. 7B). Whereas the intact 12-mer DNA remained at the chromatographic origin, nuclease P1 treatment liberated a radiolabeled cap dinucleotide that, upon further treatment with NPPase, was converted to a more rapidly migrating ³²P-labeled species that comigrated with a cold CMP standard. These results fortify our designation of the RtcB reaction product as 5′ capped GppDNA.

FIG 7.

Decapping of GppDNA by aprataxin or nucleotidyl pyrophosphatase. (A) A ³²P-labeled 12-mer GppCAATTGCGACCC strand generated by RtcB3 was gel purified and used as a substrate for the purified recombinant fission yeast aprataxin. Reaction mixtures containing 50 mM Tris-HCl (pH 8.0), 5 mM EDTA, 40 mM NaCl, 3 nM GppDNA, and 0.6 μM aprataxin were incubated at 37°C. Aliquots (10 μl) were withdrawn at the times specified and quenched with formamide-EDTA. The products were analyzed by urea-PAGE in parallel with an aliquot of the 5′ ³²P-labeled pCAATTGCGACCC strand (lane M). An autoradiograph of the gel is shown. (B) A ³²P-labeled 12-mer GppCAATTGCGACCC strand (60 fmol) was incubated for 20 min at 50°C with 0.002 U of nuclease P1 (Nucl P1) (Sigma) in a 10-μl reaction mixture containing 50 mM sodium acetate (NaOAc) (pH 5.2) and 1 mM ZnCl₂. A duplicate sample was digested with nuclease P1 and then adjusted to 50 mM Tris-HCl (pH 8.0), 10 mM MgCl₂, and 100 mM NaCl, supplemented with 0.001 U of nucleotidyl pyrophosphatase (NPPase) (Sigma), and incubated for an additional 30 min at 37°C. The digestions were quenched by adding 2 μl of 5 M formic acid. Aliquots of the nuclease P1 and nuclease P1 plus NPPase digests, and of an undigested GppDNA control, were spotted on a polyethyleneimine (PEI)-cellulose TLC plate (Merck) alongside unlabeled marker nucleotide CMP. Ascending TLC was performed with 0.75 M LiCl as the mobile phase. An autoradiograph of the TLC plate is shown, with the position of the origin indicated on the left.

Decapping of GppDNA by aprataxin.

The DNA repair enzyme aprataxin has a DNA 3′ decapping activity whereby it converts DNAppG (synthesized by E. coli RtcB) to DNAp and GMP (12, 24). Aprataxin is a member of the histidine triad family of nucleotidyltransferases that act via a covalent enzyme-(histidinyl)-NMP intermediate. Mutations in aprataxin are the cause of the human neurological disorder ataxia oculomotor apraxia-1 (25, 26). Aprataxin was initially shown to deadenylate abortive AppDNA intermediates that can accumulate when classic DNA ligases attempt to seal DNA 5′-PO₄ ends at the sites of damage or RNA 5′-PO₄ ends embedded in DNA (27, 28). Its dual capacity to deadenylate a 5′ capped AppDNA and deguanylate a 3′ capped DNAppG indicates that aprataxin is eclectic in its substrate specificity. Crystal structures of Schizosaccharomyces pombe aprataxin with AMP or GMP in the active site revealed that AMP and GMP bind at the same position and in the same anti nucleoside conformation and that aprataxin makes more extensive nucleobase contacts with guanine than with adenine (24, 29, 30). Because RtcB3 is the first instance of an enzyme that caps a DNA 5′-PO₄ end with GMP, we were interested in testing whether aprataxin could deguanylylate the 12-mer GppDNA strand produced by RtcB3. As shown in Fig. 7A, recombinant S. pombe aprataxin quantitatively converted ³²P-labeled GppDNA to pDNA in a time-dependent fashion.

DISCUSSION

The present characterization of three RtcB paralogs from M. xanthus extends our knowledge of the bacterial branch of this enzyme family beyond the single case of E. coli RtcB that was studied previously. The instructive findings are that all RtcBs are not created equal regarding their RNA splicing activities in vitro. We found that M. xanthus RtcB1 resembles E. coli RtcB in its ability to perform intra- and intermolecular sealing of a _HORNAp substrate and capping of a DNA 3′-PO₄ end. The resemblance extends to the genetic organization of M. xanthus rtcB1 in a ⇐rtcR·rtcB–rtcA⇒ operon and the high degree of structural conservation of M. xanthus RtcB1 and E. coli RtcB (243 positions of amino acid identity plus 35 positions of side chain similarity). M. xanthus RtcB2 can splice RNA but has 5-fold-lower RNA ligase specific activity than that of RtcB1 under the assay conditions employed. M. xanthus RtcB2 is less similar to E. coli RtcB (119 positions of amino acid identity plus 54 positions of side chain similarity).

In contrast, M. xanthus RtcB3 is distinctively feeble at ligating the _HORNAp substrate, although it caps a DNA 3′-PO₄ end. It is conceivable that RtcB3 is uniquely fastidious in its RNA substrate requirement (e.g., for a specific RNA sequence, secondary structure, or three-dimensional fold). Alternatively, the ligase function of M. xanthus RtcB3 might rely on a specific partner protein. M. xanthus RtcB3 is most closely related to a separate branch of bacterial RtcBs, referred to as release factor H-coupled RtcBs, because they are encoded by genes that are physically clustered with bacterial genes encoding PrfH, a homolog of class I protein release factors that terminate translation (31). PrfH-coupled RtcBs are found in the proteomes of Salmonella spp., Pseudomonas spp., and Ralstonia spp. and have 50 to 55% amino acid identity to M. xanthus RtcB3. Certain strains of E. coli also encode a PrfH-coupled RtcB paralog, in addition to the RtcB encoded by the rtcBA operon. Indeed, M. xanthus RtcB3 is more distantly related to the E. coli RtcB encoded by the rtcBA operon (with which it shares 117 positions of amino acid identity and 50 positions of side chain similarity).

The novelty of M. xanthus RtcB3 is its capacity to cap DNA and RNA 5′-PO₄ ends to form GppDNA and GppRNA products, respectively. This reaction differs from that of canonical mRNA capping enzymes, which transfer GMP from GTP via a covalent enzyme-(lysyl-Nζ)-GMP intermediate to a 5′-diphosphate terminus to form a GpppRNA product (32, 33). Although the prevalence of GppDNA and GppRNA caps in vivo is uncharted territory, one can imagine that 5′ capping of nucleic acids by RtcB3-like enzymes might serve useful functions, e.g., protecting DNA/RNA ends from 5′ exonucleases and/or installing a modification mark that directs downstream transactions. Indeed, recent studies have shown that a subset of bacterial RNAs are capped at their 5′ ends with NAD⁺ (reviewed in reference 34), although the biochemical mechanism of NAD⁺ capping has not yet been determined.

The present study revealed RtcB3 to be versatile in capping 3′-PO₄ and 5′-PO₄ ends. As such, RtcB3 joins a growing list of enzymes that can cap both ends of a polynucleotide substrate. Covalent adenylylation of an RNA 3′-PO₄ end to form RNAppA was first described as a fleeting intermediate in the synthesis of an RNA 2′,3′-cyclic phosphate terminus by RtcA, which transfers AMP from ATP via a covalent RtcA-(histidinyl)-AMP intermediate (35, 36). Subsequent studies showed that RtcA catalyzes the highly efficient adenylylation of RNA and DNA 5′-PO₄ ends to form stable AppRNA and AppDNA end products (37). Thus, there is considerable plasticity in the direction of the RtcA pathway (2′,3′ cyclization versus 5′ A capping), with resulting uncertainty as to what the real substrates of RtcA are. This theme was amplified by the report that certain classic ATP-dependent thermophilic RNA ligases, which join 3′-OH/5′-PO₄ ends via an AppRNA intermediate, are also capable of adenylylating a DNA 3′-PO₄ end to form a stable DNAppA product (38). In effect, the thermophilic RNA ligases double as DNA 3′ capping enzymes, albeit with an A cap rather than a G cap.

The theme of action at either end extends to the decapping of nucleic acids by aprataxin. Initially described as an adenylate decapping enzyme acting on abortive AppDNA intermediates formed by ATP-dependent DNA ligases, aprataxin was subsequently shown to have 3′ decapping activity on a DNAppG end formed by E. coli RtcB (12, 24) and on DNAppA ends generated by an ATP-dependent thermophilic RNA ligase (38). Here, we extend the aprataxin repertoire by showing that it can remove a 5′ guanylate cap from GppDNA formed by M. xanthus RtcB3.