Báo cáo khoa học: Factors influencing RNA degradation by Thermus thermophilus polynucleotide phosphorylase ppt

Factors influencing RNA degradation by Thermus thermophilus polynucleotide phosphorylase Marina V. Falaleeva, Helena V. Chetverina, Victor I. Ugarov, Elena A. Uzlova and Alexander B. Chetverin Institute of Protein Research of the Russian Academy of Sciences, Pushchino, Moscow Region, Russia Polynucleotide phosphorylase (PNPase, polyribonucleo- tide:orthophosphate nucleotidyltransferase, EC 2.7.7.8) is a 3¢fi5¢ exoribonuclease that catalyses the phospho- rolysis of oligo- and polyribonucleotides producing ribonucleoside 5¢-diphosphates (ppN), and this can be reversed by decreasing the inorganic orthophosphate (P i ) concentration and increasing the ppN concen- tration [1]: ðpNÞ n þP i $ðpNÞ nÀ1 þppN PNPase has been found in all types of bacteria [2] except mycoplasma [3]. Although PNPase-encoding sequences are absent from all archaea and yeast genomes examined to date (which do however contain homologues of the related RNase PH gene), they are present in higher eukaryotes [4–6]. Plant and animal PNPases are encoded by nuclear genes, but are mainly localized in organelles [7]. In plants, there are two PNPase species, one of which is targeted to chlorop- lasts and the other to mitochondria [4,8]. The PNPase gene (pnp) encodes a 80–100 kDa poly- peptide composed of five evolutionary conserved domains: two N-terminal core domains homologous to Escherichia coli RNase PH interspaced with an a-helical domain, and two C-terminal RNA-binding domains, KH and S1 [3–6,9,10]. RNase PH is a 3¢fi5¢ phosphorolytic exonuclease that is responsible for processing of the 3¢ ends of precursor tRNA molecules [11]: the KH (‘K homology’) domain was originally identified in the human RNA-binding K protein [12], and the S1 domain is named after the RNA-binding protein S1 of the E. coli ribosome [13]. Keywords 3¢fi5¢ exoribonuclease; oligonucleotide protection; RNA phosphorolysis; 3¢-terminal modifications; Thermus thermophilus Correspondence A. B. Chetverin, Institute of Protein Research of the Russian Academy of Sciences, Pushchino, Moscow Region, 142290 Russia Fax: +7 495 632 7871 Tel: +7 496 773 2524 E-mail: alexch@vega.protres.ru (Received 30 November 2007, revised 29 February 2008, accepted 4 March 2008) doi:10.1111/j.1742-4658.2008.06374.x At the optimal temperature (65 °C), Thermus thermophilus polynucleotide phosphorylase (Tth PNPase), produced in Escherichia coli cells and isolated to functional homogeneity, completely destroys RNAs that possess even a very stable intramolecular secondary structure, but leaves intact RNAs whose 3¢ end is protected by chemical modification or by hybridization with a complementary oligonucleotide. This allows individual RNAs to be isolated from heterogeneous populations by degrading unprotected species. If oligonucleotide is hybridized to an internal RNA segment, the Tth PNPase stalls eight nucleotides downstream of that segment. This allows any arbitrary 5¢-terminal fragment of RNA to be prepared with a precision similar to that of run-off transcription, but without the need for a res- triction site. In contrast to the high Mg 2+ requirements of mesophilic PNPases, Tth PNPase retains significant activity when the free Mg 2+ con- centration is in the micromolar range. This allows minimization of the Mg 2+ -catalysed nonenzymatic hydrolysis of RNA when phosphorolysis is performed at a high temperature. This capability of Tth PNPase for fully controlled RNA phosphorolysis could be utilized in a variety of research and practical applications. Abbreviations P i , inorganic orthophosphate; PNPase, polynucleotide phosphorylase; Tth PNPase, PNPase of Thermus thermophilus. 2214 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS The PNPase molecules of E. coli and Streptomyces an- tibioticus are composed of three identical subunits that form a doughnut-shaped structure with a central hole capable of accommodating an RNA strand [10,14], while the spinach chloroplast PNPase seems to be composed of two such ring [4,15]. Interestingly, the three-dimensional structures of the bacterial PNPase [10] and of the archaeal exosome core of RNase PH [16] are highly homologous, suggesting that the molec- ular mechanisms of action of PNPase and RNase PH may be very similar [16,17]. As the intracellular P i concentration is usually high, the primary physiological role of PNPase is thought to be the phosphorolytic degradation of RNA [2,18]. Important manifestations of this function include the control of RNA quality by eliminating defective mole- cules [19–21], regulation of the virulence and persis- tence of pathogenic bacteria [22], anticancer activity due to degradation of tumour-associated mRNAs [23,24], and RNA processing [25]. However, PNPase sometimes employs its polymerizing ability in vivo to extend the 3¢ ends of RNAs instead of using a poly(A) polymerase [26,27]. For a long time, due to its synthetic activity in the absence of template, PNPase was widely used as a tool for producing a variety of model nucleic acids to solve important biological problems, such as establishing the genetic code and studies on physicochemical properties of polyribonucleotides [28]. These applications have been largely replaced by more versatile chemical synthesis of oligori- bonucleotides with defined sequences and in vitro transcription of DNA templates for the synthesis of longer RNAs. In contrast to nucleotide polymerization, the exonu- clease activity of PNPase was rarely exploited. This is mainly due to the fact that, despite the high processivi- ty of the enzyme, phosphorolysis is not easily control- lable. The rate and extent of degradation depend on the stability of the RNA structure and vary greatly among RNA species [2]. Increasing temperature results in destabilization of the RNA structure and more thor- ough and uniform degradation [29]. However, E. coli PNPase is unstable above 55 °C and is rapidly inacti- vated at 65 °C [2]. Conditions have not been estab- lished that would allow the desired RNA or the desired fragment of RNA to selectively survive the phosphorolysis reaction. The only successful example of a controlled phosphorolysis by PNPase has been the faithful removal of poly(A) tails from mRNAs. This was achieved by carrying out the reaction at 0 °C, when the rest of mRNA was structured and hence resistant to PNPase [30]. In this regard, PNPases from thermophilic organ- isms deserve special attention as they are expected to be active at high temperatures at which the second- ary and tertiary structures of RNA are melted. There have been a few reports on such PNPases. In one of them, PNPases from thermophilic bacteria Bacillus stearothermophilus and Thermus aquaticus showed highest polymerizing activities at 69 and 80 °C, respectively. However, their ability to phosp- horolyse structured RNAs was not characterized, the enzyme from T. aquaticus was not isolated, and the purified enzyme from B. stearothermophilus was reported to be composed of atypically small subunits of 51 kDa [31]. In another paper, PNPase from Thermus thermophilus (Tth PNPase), a close relative of T. aquaticus, was purified to apparent homogene- ity as tested under non-denaturing conditions, but the purified enzyme completely lacked phosphorolytic activity, presumably due to proteolytic degradation that had occurred during the isolation procedure. The purified enzyme was reported to be composed of three unequal polypeptides of 92, 73 and 35 kDa [32]. In 1997, the nucleotide sequence of a putative pnp gene from T. thermophilus was submitted to the Gen- Bank database (accession number Z84207). This open reading frame has the capacity to encode a 78.2 kDa polypeptide (a molecular mass that is similar to that of the E. coli PNPase, 77.1 kDa [33]) in which all struc- tural domains characteristic of typical PNPases can be identified (NCBI Protein Database accession number CAB06341). However, details of the expression, isola- tion and characterization of the enzyme, although noted in the NCBI Protein Database submission, have not been published. In this paper, we report cloning of the pnp gene of T. thermophilus, its expression in E. coli cells, and characterization of biochemical properties of the iso- lated enzyme. We show that the isolated Tth PNPase is capable of phosphorolysis, and that even RNAs with very stable secondary structures are readily degraded to completion. We further show that RNA can be pro- tected from phosphorolysis by either modifying its 3¢ end or annealing its 3¢-terminal sequence to a comple- mentary oligonucleotide. If a mixture of protected and unprotected RNAs is treated with Tth PNPase, then only unprotected RNA is degraded. Finally, we show that hybridization of oligonucleotide to an internal segment of RNA protects that segment, the upstream portion and a 8 nt downstream sequence of the RNA from Tth PNPase. These features could make Tth PNPase a useful tool for controlled RNA degradation in vitro. M. V. Falaleeva et al. RNA phosphorolysis by Tth PNPase FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS 2215 Results Isolation of Tth PNPase from E. coli cells The pnp gene was PCR-amplified using the T. thermo- philus chromosomal DNA as a template, cloned within a plasmid downstream of the T7 promoter, and expressed in T7 RNA polymerase-producing E. coli cells. The expression product has a higher molecular mass (78 203 Da, as calculated from the amino acid sequence deposited in the NCBI Protein Database under accession number CAB06341) than its E. coli counterpart (77 122 Da [33]), and this allowed removal of the E. coli enzyme from the Tth PNPase prepara- tion to be monitored by PAGE analysis (see supple- mentary Fig. S1). The Tth PNPase isolation procedure included heating of the cell lysate at 70 °C, resulting in denaturation and precipitation of most of the host proteins, including the E. coli PNPase; ion-exchange chromatography on a DEAE-Sepharose column; incu- bation in the presence of P i and Mg 2+ (‘autolysis’, during which any endogenous RNA was completely degraded by PNPase; phosphatase was included to eliminate any possible 3¢ phosphoryl groups on RNA that might interfere with the phosphorolysis [34]); and gel filtration through a highly porous Superose 6 column. The full-sized 78 kDa Tth PNPase polypeptide con- stituted > 50% of the final preparation, with the rest of protein comprising multiple shorter bands grouping at around 60–70 and 30–40 kDa. These minor poly- peptides co-purified with PNPase, and similar products accumulated in the E. coli cells upon the induction of T7 RNA polymerase (supplementary Fig. S1). There- fore, they are most likely fragments of the 78 kDa polypeptide, which is highly susceptible to proteolysis [32]. As the enzyme preparation did not contain con- taminating activities (as shown below), we did not attempt to further purify it. The enzyme yield (based on the content of the 78 kDa polypeptide) was 3.3 mg from 9 g of the Tth PNPase-producing E. coli cells. The ADP-polymerizing activity of the final preparation was approximately 200 unitsÆmg )1 , which is higher than reported for the PNPase isolated from T. thermo- philus cells (approximately 70 unitsÆmg )1 [32]), and, unlike that preparation, our Tth PNPase retained phosphorolytic activity (see below). Conditions optimal for exonuclease activity The isolated Tth PNPase degraded unprotected RNA, but did not degrade RNA whose 3¢ end was hybridized to a complementary oligodeoxyribonucleotide (Fig. 1). This observation indicated that the enzyme preparation does indeed possess the 3¢fi5¢ exonuclease activity and was free from endoribonucleases. Also, RNA remained intact in the absence of P i , suggesting that all degradation was due to phosphorolysis, rather than hydrolysis of polyribonucleotides. Although RNA was not degraded in the presence of 56 lm EDTA, which would sequester trace amounts of divalent cations, complete degradation of unprotected RNA occurred when the concentration of added Mg 2+ was equal to that of EDTA (Fig. 1). Given the value of the dissociation constant for the Mg 2+ :EDTA complex (approximately 10 )9 m [35]), this indicates that the requirement of Tth PNPase for free Mg 2+ is very low (£ 1 lm), which is much lower than is the Mg 2+ requirement of a mesophilic (Micrococcus luteus) enzyme, for which the apparent K m value for Mg 2+ is 0.8 mm [36]. In subsequent experiments described here, the free Mg 2+ concentration was main- tained at about 40 lm, which allowed enzymatic phos- phorolysis of RNA to be performed at a high temperature without significant concomitant Mg 2+ - catalysed hydrolysis. Under such conditions, the opti- mal temperature for phosphorolysis was around 65 °C, whereas, in the absence of PNPase, RNA remained substantially intact even at 85 °C (Fig. 2). Figure 3(A) shows that, at a saturating enzyme con- centration, when all RNA strands are phosphorolysed synchronously, a 109 nt 3¢-terminal fragment of a highly structured RQ135 RNA [37,38] is degraded to near completion within 2 min, suggesting a rate of approximately 50 nt per min at 65 °C. This is much faster than the rate of 3.5 nt per min that is achievable Fig. 1. Mg 2+ dependence of RNA phosphorolysis. Nondenaturing PAGE patterns for the 3¢ fragment of RQ135 RNA treated with Tth PNPase after it had been annealed in the absence or presence of the oligodeoxyribonucleotide R-38 that is complementary to the RNA 3¢ end. Each sample contained 0.056 m M EDTA, which was present in the RNA and enzyme preparations, and the indicated concentration of added MgCl 2 . RNA phosphorolysis by Tth PNPase M. V. Falaleeva et al. 2216 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS with the E. coli PNPase at its optimal temperature (37 °C) using a similarly structured tRNA-like 3¢-ter- minal domain of tobacco mosaic virus RNA [2]. How- ever, in contrast to the E. coli enzyme, which saturates RNA at a 1 : 1 m ratio, a larger amount of Tth PNPase is required for saturation. For the concentra- tion of RNA used in the experiments shown in Fig. 3 (50 nm), saturation occurs at a fourfold higher enzyme concentration; the Tth PNPase concentration was cal- culated assuming that one enzyme molecule active in phosphorolysis consists of three intact (78 kDa) subun- its [34]. The same saturating enzyme-to-RNA ratio was determined when complex formation in the pres- ence of Mg 2+ (but in the absence of P i to prevent phosphorolysis) was monitored by a shift of the 32 P-labelled RNA band during PAGE (Fig. 3B). Based on the data shown in Fig. 3, a K d value of approxi- mately 100 nm was estimated for the Tth PNPa- se:RNA interaction. Hence, the affinity of Tth PNPase for RNA is approximately 10 times lower than that of the E. coli enzyme (K d = 10–20 nm [39]). The lower affinity does not necessarily mean that larger amounts of Tth PNPase have to be used. Instead, the complete degradation of RNA can be achieved by either reduc- ing the reaction volume to concentrate the reactants, or increasing the incubation time. Selective protection of RNA by terminally hybridized oligonucleotides As determined in an RT-PCR assay, under the optimal conditions established here (temperature, Mg 2+ and EDTA concentrations, enzyme-to-RNA ratio, reaction time), Tth PNPase eliminated more than 99% of a highly structured unprotected RNA without affecting RNA whose 3¢-terminal sequence was protected by annealing with a complementary oligonucleotide (Fig. 4A). This indicated that hybridization of an oligonucleotide at the 3¢ end of an RNA could selec- tively protect it from degradation by Tth PNPase. Figure 4B demonstrates that this is indeed the case. In this experiment, a mixture of two RNA species, 5¢ and 3¢ fragments of RQ135 RNA [37,38], was annealed with an oligodeoxyribonucleotide complementary to the 3¢ end of the 3 ¢ fragment. Subsequent phosphoroly- sis with Tth PNPase resulted in complete degradation of the 5¢ fragment and yielded a virtually pure 3¢ frag- ment (except for the oligodeoxyribonucleotide itself, which can be removed by electrophoresis or DNase treatment). Fig. 2. Temperature dependence of RNA phosphorolysis. Denatur- ing PAGE patterns of 0.5 pmol of the 3¢ fragment of RQ135 RNA incubated at the indicated temperature with or without 0.5 pmol of Tth PNPase. AB Fig. 3. Interaction of Tth PNPase with RNA. (A) Denaturing PAGE patterns of 0.5 pmol of the 3¢ fragment of RQ135 RNA treated with the indicated relative amounts of Tth PNPase for the indicated time periods. (B) Gel-shift analysis by nondenaturing PAGE of the complex formed between Tth PNPase and the [a- 32 P]-labelled 3¢ fragment of RQ135 RNA upon incubation for 15 min at 65 °C in a phosphate-free phosphorolysis buffer. RNA bands were visualized by scanning the gel using the Cyclon TM storage phosphor system (Packard Instruments, Meriden, CT, USA). The concentration of RNA in the incubation mixtures was constant (50 n M), and the concentration of PNPase varied as indicated. M. V. Falaleeva et al. RNA phosphorolysis by Tth PNPase FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS 2217 Protection of RNA by 3¢-terminal modifications Figure 5 shows that 3¢-terminal modifications of RNA make it more resistant to phosphorolysis by Tth PNPase, further confirming that RNA degradation is due to the 3¢fi5¢ exonuclease activity. Oxidation of RNA with periodate, which opens the terminal ribose ring and generates a dialdehyde [40], results in selective protection of the modified RNA species (Fig. 5A,B). This protection is moderate, but it is increased upon treatment of the oxidized RNA with aniline (Fig. 5C), which eliminates the broken nucleoside and yields RNA whose terminal 3¢ hydroxyl is phosphorylated [41]. The high resistance of the 3¢-phosphorylated RNA to Tth PNPase is not due to chemical modifica- tion of the RNA body during periodate or aniline treatment, because it is fully released upon removal of the 3¢-terminal phosphate by a phosphatase (Fig. 5C). A similar increase in protection is achieved by treat- ment of the oxidized RNA with biotin hydrazide, which adds the biotin group to the RNA 3¢ end (Fig. 5D). Comparison with PNPase of E. coli Figure 6 compares some properties of Tth PNPase with those of the E. coli PNPase. While both the enzymes readily degraded poly(A), an unstructured polyribonucleotide, they behaved differently with the 3¢-terminal fragment of RQ135 RNA, which possesses a strong secondary structure [37,38]. Unlike the T. thermophilus enzyme, the E. coli PNPase only phos- phorolysed a fraction of the RNA at its optimal tem- perature, and the rest of the RNA remained apparently intact. This resembles the degradation pat- tern of another structured RNA, tRNA [29], which exists in two conformations, of which only one is sus- ceptible to E. coli PNPase attack at 37 °C [42]. How- ever, in contrast to tRNA [29,42], E. coli PNPase was unable to completely degrade the 3¢-terminal fragment of RQ135 RNA even when temperature was raised to 55 °C (supplementary Fig. S2). Another difference between these two enzymes concerns the effects of the 3¢-terminal phosphoryl group. While this group effi- ciently protected RNA from Tth PNPase, such an effect was not observed with the E. coli PNPase, prob- ably because it was masked by the inability of the latter enzyme to completely degrade the unprotected RNA (Fig. 6). Protection of RNA by internally hybridized oligonucleotides Figure 7 shows degradation patterns of three nested RNAs that were differently extended at the 3¢ end. All three RNAs were subjected to phosphorolysis by Tth PNPase after annealing in the presence or the absence of a 38 nt oligodeoxyribonucleotide that hybridized to the 3¢-terminal sequence of the shorter RNA (SmaI) and to internal sequences of the longer RNAs (EcoRI and PvuII). The oligonucleotide completely protected the shorter RNA, and partially protected the two longer RNAs, whose degradation yielded a resistant product that migrated during PAGE slightly more slowly than did A B Fig. 4. Selectivity and completeness of RNA degradation. (A) Ethidium bromide-stained nondenaturing PAGE patterns for the products of RT-PCR of RNA that survived incubation with Tth PNPase of the indicated number of molecules of CT1n1 RNA after it had been annealed in the absence or presence of the oli- godeoxyribonucleotide R-38 that is complementary to the RNA 3¢ end. PNPase was present in all samples except those marked ‘–PNPase’. (B) Nondenaturing PAGE patterns of a mixture of the 3¢ and 5¢ fragments of RQ135 RNA incubated with or without Tth PNPase after annealing in the presence of the oligodeoxyri- bonucleotide R-38 that is complementary to the 3¢ end of the 3¢ fragment. RNA phosphorolysis by Tth PNPase M. V. Falaleeva et al. 2218 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS A C D B Fig. 5. Effects of 3¢-terminal modifications. (A–C) Denaturing PAGE patterns for a mix- ture of 0.5 pmol each of the 3¢ fragment (upper band) and 5¢ fragment (lower band) of RQ135 RNA, of which either the 5¢ frag- ment (A and C, initial mixture) or the 3¢ frag- ment (B) was oxidized with sodium periodate (3¢oxi) while the other possessed the native 3¢ end (3¢OH). The mixture was either not treated (N ⁄ T) or treated with the indicated amounts (in pmol) of Tth PNPase, either directly (A–C, initial mixture) or after treatment with aniline, resulting in a conver- sion of the oxidized fragment to 3¢-phos- phorylated (C, after aniline) and then with shrimp alkaline phosphatase, removing the 3¢-phosphate (C, after phosphatase). (D) Nondenaturing PAGE patterns for 3¢ frag- ments whose 3¢ end was either modified with biotin (3¢bio) or not modified (3¢OH), and which were either treated (+) or not treated ()) with Tth PNPase, or mixed with a fivefold molar excess of streptavidin before electrophoresis (Str). Fig. 6. Comparison of Thermus thermophilus and Escherichia coli PNPases. Nondenaturing PAGE patterns for 20 ng of poly(A) and 0.5 pmol of the 3¢ fragment of RQ135 RNA whose 3¢ end was either intact (3¢OH) or esterified with phosphate (3¢P) as a result of consecutive reactions with periodate and aniline, and which was either not treated ()) or treated in the presence of 1 m M P i and 0.1 m M Mg 2+ with 4 pmol of PNPase from E. coli (E) at 37 °C or with 1 pmol of Tth PNPase (T) at 65 °C. Fig. 7. Effects of internally hybridized oligonucleotide. Nondena- turing PAGE patterns for nested transcripts of the same plasmid digested at consecutive sites with SmaI (transcript length 109 nt), EcoRI (127 nt) or PvuII (303 nt), and treated (+) or not treated ()) with Tth PNPase after annealing in the absence or presence of oligodeoxyribonucleotide R-38 complementary to positions 72–109. White arrows indicate the resistant phosphorol- ysis products. M. V. Falaleeva et al. RNA phosphorolysis by Tth PNPase FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS 2219 the hybrid of the oligonucleotide with the shorter RNA (left part of Fig. 7). Such products were not observed in the absence of the oligonucleotide (right part of Fig. 7). The fine structure of the phosphorolysis product was analysed by high-resolution (sequencing) PAGE under denaturing conditions (Fig. 8). Before phosphorolysis, each of the three RNAs was dephosphorylated and then labelled at the 5¢ end using [c- 32 P]ATP and poly- nucleotide kinase. Figure 8A shows that the resistant product had rather heterogeneous 3¢ ends and was up to eight nucleotides longer than the terminally pro- tected RNA, whose 3¢ ends were also not uniform, but this is common for run-off transcription [43]. The het- erogeneity of the phosphorolysis product was mainly due to the heterogeneity of the protecting oligonucleo- tide, and was substantially reduced when the 5¢ ends of this oligonucleotide were made uniform by digesting a longer oligonucleotide with restriction endonuclease. Phosphorolysis of the longer RNAs, hybridized with this 5¢-terminally polished oligonucleotide, yielded a product whose 3¢-terminal heterogeneity was similar to that of a product of the run-off transcription, and which was 8 nt longer than the terminally protected shorter RNA (Fig. 8B). This indicates that Tth PNPase stalls on a phosphorolysed RNA eight resi- dues from the 5¢ end of the internally hybridized oligo- nucleotide, which parallels the observation that PNPase of E. coli stalls six to nine residues from the base of a stable stem-loop structure in an RNA strand [44]. Similarly, the exosome core of a hyperthermophil- ic archaeon Sulfolobus solfataricus, consisting of RNase PH subunits, also stalled eight or nine nucleo- tides from an upstream stem, which has been attrib- uted to the inability of a double-stranded structure to enter the central hole of the doughnut-shaped enzyme structure [45]. The presence of a phosphate group at the 5¢ end of the oligonucleotide did not affect the length of the protected RNA fragment, but slightly reduced its heterogeneity (Fig. 8B; compare lanes 3 and 4 with lanes 7 and 8). Discussion This paper reports the isolation of a thermophilic PNPase whose subunits have a molecular mass typical of PNPases and which possesses phosphorolytic activ- ity. The isolated enzyme is free from other ribonucleas- es. Biochemical characterization of the isolated enzyme revealed several important features that were not observed with PNPases from other sources. First, Tth PNPase has maximal phosphorolytic activity at a temperature at which even the most stable hairpins are melted. Therefore, it degrades to completion those RNAs that mesophilic PNPases fail to phosphorolyse. Second, the requirement of Tth PNPase for free Mg 2+ ions is extremely low. This feature is beneficial, because it allows directed 3¢fi5¢ phosphorolysis to be performed at high tem- peratures without a risk of breakage of RNA strands due to Mg 2+ -catalysed hydrolysis. Moreover, it is this feature that permitted us to discover that RNA can survive the high-temperature phosphorolysis con- ditions if its 3¢ end is protected. Protection can be achieved either by hybridizing the terminal sequence with a complementary oligonucleotide (care must be taken to ensure that the hybrid is not melted at the reaction temperature) or by modifying the terminal nucleoside. The results show that 3¢-terminal modifications pro- tect RNA from Tth PNPase. Previously, a phosphoryl group in either the 3¢ or 2¢ position of the terminal ribose, as well as the 2¢,3¢-cyclic phosphate, were found to prevent phosphorolysis of short (up to 6 nt) oligo- nucleotides by PNPases from E. coli [46] and M. luteus AB Fig. 8. RNA sequence protected by internally hybridized oligonu- cleotide. High-resolution denaturing PAGE patterns for [5¢- 32 P]- labelled SmaI, EcoRI and PvuII transcripts protected from Tth PNPase by hybridization with oligodeoxyribonucleotide R-38 (cf. Fig. 7). (A) The oligonucleotide was prepared by chemical synthe- sis without further treatment. Lanes 1 and 2 show untreated SmaI and EcoRI transcripts respectively; lanes 3 and 4 show PNPase-treated EcoRI and PvuII transcripts, respectively. The untreated PvuII transcript migrated too slowly to be shown here (cf. Fig. 7). (B) Oligonucleotides were prepared from a longer one by digestion at the SmaI site to generate uniform 5¢ ends (lanes 2–4) and by further removing the 5¢-terminal phosphate (lanes 6–8). Shown are untreated SmaI transcripts (lane 1), PNPase- treated SmaI (lanes 2 and 6), EcoRI (lanes 3 and 7) and PvuII (lanes 4 and 8) transcripts, and the EcoRI transcript partially degraded in mild alkali to produce a 1 nt ladder [51] (lane 5). RNA bands were visualized by scanning the gel using the Cyclon TM storage phosphor system. RNA phosphorolysis by Tth PNPase M. V. Falaleeva et al. 2220 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS [47]. However, there were conflicting reports on phos- phorolysis of the longer (up to 20 nt) 3¢-phosphory- lated fragments that constitute most commercial RNA preparations, of which between 20% [48] and 100% [49] was found to be degradable by E. coli PNPase. This disagreement could be explained by a different average fragment length in the RNA preparations or by different content of contaminating enzymes (such as a phosphatase that would remove the 3¢-terminal phosphate from the RNA fragments) in the PNPase preparations used. The results further show that Tth PNPase is sensi- tive to 3¢-terminal modifications that cannot result in positioning of a phosphate or a similar bulky nega- tively charged group in the subsite of the PNPase active site normally occupied by the terminal nucleo- side, the mechanism that was hypothesized to be the basis for the inhibitory action of the 3¢-terminal phos- phate [50]. The fact that periodate oxidation of RNA, resulting a conversion of the 2¢,3¢-cis-glycol to dialde- hyde, also has an inhibitory effect indicates that inter- actions with the free 3¢ and ⁄ or 2¢ hydroxyls of the terminal ribose may be important for correct binding of substrate RNA to the active site of Tth PNPase, as was found for the S. solfataricus exosome [45]. Terminal protection allows desirable RNA species to be isolated from heterogeneous populations by degrading other (unprotected) strands. This could also be used for reducing artifactual sequence recom- bination and increasing the specificity of RT-PCR assays. The observation that biotin at the 3¢ end of an RNA strand efficiently protects it from phospho- rolysis suggests an efficient way of obtaining a 100% pure biotinylated RNA preparation. Many other Tth PNPase applications of this sort can undoubtedly be conceived. Finally, the results of this study show that any arbi- trary 5¢-terminal fragment of an RNA strand can be obtained by a controlled phosphorolysis with a preci- sion similar to that of run-off transcription, but with- out the need for a restriction site. This can be done by protecting the RNA strand with an oligonucleotide complementary to an internal RNA segment that lies 8 nt upstream of the desired 3¢ end. In summary, this paper demonstrates that the prop- erties of Tth PNPase allow fully controlled RNA deg- radation, which could be used in various research and practical applications. In view of the structural similar- ities between bacterial PNPase [10] and the S. solfatari- cus exosome [16], one may expect that the archaeal enzyme, which is also thermophilic, will possess bio- chemical properties similar to those reported here for Tth PNPase. Experimental procedures Cloning and expression of the pnp gene of T. thermophilus The pnp gene was PCR-amplified using the chromosomal DNA of T. thermophilus VK1 as a template and primers 5¢-GACGTCGACAT ATGGAAGGCACACCCAATG-3¢ [matching the start of the PNPase-coding sequence, positions 977–995 of GenBank accession number Z84207 (underlined) and introducing an NdeI site (bold)], and 5¢-TTCGAATTC T TACTTGCGCCGCCTGGG-3¢ [matching the end of the PNPase-coding sequence, positions 3101–3117 of GenBank accession number Z84207 (underlined) and introducing an EcoRI site (bold)]. The PCR product was digested at the NdeI and EcoRI sites, and ligated into plasmid vector pT7-7 [52] between these sites. The resulting plasmid, pT7- PNP, was cloned [53] and expressed in E. coli B834(DE3) ⁄ pLysS cells [54] by inducing T7 RNA polymer- ase synthesis using 1 mm isopropyl b-d-thiogalactoside for 4 h at 37 °C. Isolation of Tth PNPase Nine grams of Tth PNPase-producing E. coli cells pel- leted from 6 L of culture were suspended in 40 mL of ice-cold buffer A (20 mm Hepes ⁄ NaOH pH 8.2, 100 mm NaCl, 0.1 mm Na-EDTA, 2 mm b-mercaptoethanol, 1 mm phenylmethanesulfonyl fluoride) and disrupted in a Gaulin Micron Lab 40 homogenizer (APV Thermotech GmbH, Artem, Germany) at 1400 bar. The lysate was 10-fold diluted with buffer A and incubated for 60 min at 70 °C with stirring. After pelleting of the cell debris and precipitated proteins for 40 min at 25 900 g in rotor JA-14 of a J2-21 centrifuge (Beckman, Vienna, Austria), the lysate was applied to a 11 mL DEAE-Sepharose FF (Amersham Biosciences, Vienna, Austria) column equili- brated with buffer B (20 mm Tris ⁄ HCl pH 8.0, 1 m m Na-EDTA, 0.1 m NaCl, 2 mm b-mercaptoethanol, 5% w ⁄ w glycerol). After washing the column with buffer B, the enzyme was eluted with a linear gradient of NaCl (0.1–0.5 m) in the same buffer. Fractions with a high PNPase activity were pooled, supplemented with DNase (1 lgÆmL )1 ) and calf intestinal alkaline phosphatase (1 unitÆmL )1 ), and dialysed overnight against buffer B containing 3 mm MgCl 2 and 5 mm Na-phosphate. After an additional incubation for 30 min at 37 °C, followed by 60 min at 66 °C and then chilling on ice, the precipi- tated proteins were pelleted as above. The supernatant was concentrated to a volume of 0.25 mL using a Centr- icon-30 filter (Amicon, Beverly, MA, USA), cleared by centrifugation in a microcentrifuge at 12 100 g, and sub- jected to gel filtration through a 24 mL Superose 6 HR column (Amersham Biosciences) in buffer B. Fractions with the highest PNPase activity were pooled, dialysed M. V. Falaleeva et al. RNA phosphorolysis by Tth PNPase FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS 2221 overnight against the storage buffer (50 mm Tris ⁄ HCl pH 8.0, 100 mm NaCl, 10% w ⁄ w glycerol, 0.1% Nonidet P40 (BDH Ltd, Poole, England), 2 mm dithiothreitol), cleared by centrifugation, and stored at )70 °C after the addition of glycerol to 50% w ⁄ w. Enzyme assay Tth PNPase activity was determined colorimetrically as P i [55] released from ADP during the synthesis of poly(A) at 70 °C in a mixture containing 50 mm Tris ⁄ HCl pH 8.2, 20 mm ADP, 0.1 m NaC1, 1 mm MgCl 2 and no primer. One unit of activity was defined as the amount of enzyme that polymerized 1 lmol of ADP per hour. Protein was measured by the method described by Lowry et al. [56]. Enzyme purity was analysed by SDS–PAGE (supplementary Fig. S1). The percentage of full-sized PNPase polypeptide was estimated from Coomassie G-250-stained gel images using program imagej (http:// rsb.info.nih.gov/ij/). Isolation of E. coli PNPase Unless otherwise stated, all procedures were carried out at 4 °C. Homogenization Two hundred grams of pelleted E. coli B cells were sus- pended in 200 mL of buffer C (50 mm Tris ⁄ HCl pH 8.0, 1mm Na-EDTA, 10 mm MgCl 2 , 0.2 m KCl, 5 mm b-mer- captoethanol, 5% w ⁄ w glycerol, 1 mm phenylmethane- sulfonyl fluoride) and disrupted in homogenizer 15M-8TA (APV Gaulin Inc., Everett, MA, USA) at 600 bar. Phase partitioning After removal of cell debris (for 40 min at 30 100 g in rotor JA-14 of the Beckman J2-21 centrifuge), the lysate was subjected to the liquid polymer phase partitioning procedure previously developed for the isolation of Qb replicase, using a mixture of polyethylene glycol 6000 (Merck KGaA, Darmstadt, Germany) and Dextran T500 (Amersham Biosciences), followed by extraction of RNA-bound proteins from the dextran phase using NaCl [57]. DEAE cellulose After dialysis against buffer D (10 mm Tris ⁄ HCl pH 7.5, 1mm Na-EDTA, 5 mm MgCl 2 ,5mm b-mercaptoethanol, 5% w ⁄ w glycerol), until conductivity of the extract dropped to that of buffer D100 (buffer D plus 100 mm NaCl), it was cleared by centrifugation for 40 min at 30 100 g in rotor JA-14 and applied to a 200 mL DE-52 column (Whatman, Florhom Park, NJ, USA) equilibrated in buffer D100. The column was washed with 2.5 vol- umes of D100, then with 2.5 volumes of D150 (buffer D plus 150 mm NaCl) and eluted with a gradient of NaCl (150–400 mm) in buffer D. Fractions with a high PNPase activity were pooled, protein was concentrated by precipi- tation with ammonium sulphate (40 gÆ100 mL )1 ), and the pellet was solubilized in 20 mL of buffer and dialysed overnight against buffer D100. Autolysis After clearing the dialysed preparation by centrifugation, it was supplemented with 50 lg of DNase, 50 units of calf intestinal alkaline phosphatase, 1 mm ZnCl 2 and 1 mm phenylmethanesulfonyl fluoride, and incubated for 60 min at 37 °C. Then 1 m K-phosphate (pH 8.0) was added to a concentration of 100 mm, and incubation was continued for an additional 30 min. Sepharose CL-6B The autolysed preparation was cleared by centrifugation and subjected to gel filtration through a column of Sepha- rose CL-6D, Amersham Biosciences (2.6 · 100 cm) equili- brated in buffer D100. Mono Q Fractions with high activity were pooled and chromato- graphed in several portions on a FPLC system through a Mono Q HR 5 ⁄ 5 column (Amersham Biosciences) equili- brated in buffer D100 and utilizing a 100–1000 mm NaCl gradient in buffer D. Superose 6 Active Mono Q fractions were pooled, concentrated by passing through a Centricon-30 filter (Amicon), and sub- jected to gel filtration through a 24 mL Superose 6 HR column (Amersham Biosciences). Fractions were analysed by SDS–PAGE, and those containing the least amount of impurities were pooled, and stored at )70 °C after addition of glycerol to 50% w ⁄ w. The final preparation (2.2 mg of protein) was > 90% pure (supplementary Fig. S1C, lanes 2 and 3) and was free from endonuclease activity. The specific activity of the enzyme was approximately 450 unitsÆmg )1 , with 1 unit being defined as the amount of enzyme that released 1 lmol of P i from ADP per hour at 37 °C in a mixture containing 100 mm Tris ⁄ HCl pH 8.0, 10 mm ADP, 0.5 mm Na-EDTA, 10 mm MgCl 2 and no primer. RNA phosphorolysis by Tth PNPase M. V. Falaleeva et al. 2222 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS RNA sequences CT1n1 RNA is a derivative of RQ135 )1 RNA [37] carry- ing a 43 nt insert (underlined), generated in unpublished experiments on intermolecular RNA recombination. The RNA preparation was obtained by transcription of a SmaI-digested plasmid, encoding the recombinant sequence: GGGGUUCCAACCGGAAGUUGAGGGAUGCCUAGG CAUCCCCCGUGCGUCCCUU AAAGCUUCAUUCUUC CUUUCUUUAAAAGAGAGAGAGAGAAAGCGAGGG AUUUGAGAGAUGCCUAGGCAUCUCCCGCGCGCC GGUUUCGGACCUCCAGUGCGUGUUACCGCACUG UUAGCCC. The 5¢ fragment of RQ135 RNA is a 52 nt 5¢ terminal sequence of RQ135 )1 RNA [37] carrying a 23 nt extension (underlined), obtained by transcription of the 5¢ fragment- encoding plasmid [58] digested at site BamHI: GGGG UUCCAACCGGAAGUUGAGGGAUGCCUAGGCAUC CCCCGUGCGUCCCUU CUGCAGCUCGAGUCUAGAG GAUC. The 3¢ fragment of RQ135 RNA (transcript SmaI) is a 81 nt 3¢ terminal sequence of RQ135 )1 RNA [37] carrying a 28 nt extension (underlined), obtained by transcription of the 3¢ fragment-encoding plasmid [58] digested at the SmaI site: GGCGCUGCAGCUCGAGUCUAGAGGAUCCUA CGAGGGAUUUGAGAGAUGCCUAGGCAUCUCC CG CGCGCCGGUUUCGGACCUCCAGUGCGUGUUACC GCACUGUUAGCCC. Transcript EcoRI (the 3¢ fragment of RQ135 RNA extended at the 3¢ end by a 18 nt sequence, underlined) was obtained by transcription of the 3¢ fragment-encoding plas- mid [58] digested at site EcoRI: GGCGCUGCAGCUCGA GUCUAGAGGAUCCUACGAGGGAUUUGAGAGAUG CCUAGGCAUCUCCCGCGCGCCGGUUUCGGACCU CCAGUGCGUGUUACCGCACUGUUAGCCC GGGUA CCGAGCUCGAAUU. Transcript PvuII (the 3¢ fragment of RQ135 RNA extended at the 3¢ end by a 194 nt sequence, underlined) was obtained by transcription of the 3¢ fragment-encoding plasmid [58] digested at site PvuII: GGCGCUGCAGCUC GAGUCUAGAGGAUCCUACGAGGGAUUUGAGAGA UGCCUAGGCAUCU CCCGC GCG CCGGU UUCGG AC CUCCAGUGCGUGUUACCGCACUGUUAGCCC GGG UACC GAGCUCGAA UUCGUAAUCAUGGUCAUAGC UGUUUCCUGUGUGAAAUUGUUAUCCGCUCACAA UUCCACACAACAUACGAGCCGGAAGCAUAAAGU GUAAAGCCUGGGGUGCCUAAUGAGUGAGCUAA CUCACAUUAAUUGCGUUGCGCUCACUGCCCGCUU UCCAGUCGGGAAACCUGUCGUGCCAG. RNA preparations RNA samples used in this work were prepared by run-off transcription of plasmids digested at appropriate restriction sites and purified by PAGE as described previously [58]. Where indicated, RNA was oxidized with sodium perio- date, followed by treatment with aniline and shrimp alka- line phosphatase as described previously [58]. Biotinylated RNA was prepared by incubation of periodate-oxidized RNA at 4 °C for 20 h with 5 mm biotin hydrazide (Sigma, St Louis, MO, USA) in the presence of 0.1% SDS and 0.2 m Na-acetate (pH 4.8). Unincorporated biotin hydra- zide was removed by gel filtration through a Sephadex G-25 (Amersham Biosciences) spin column. RNA:oligonucleotide hybridization Immediately before hybridization, RNA (in 0.1 mm Na-EDTA pH 8.0) and oligodeoxyribonucleotides (in 10 mm Tris ⁄ HCl pH 9.0, 0.01 mm Na-EDTA) were sepa- rately melted in a boiling bath for 2 min followed by rapid chilling on ice. Hybridization was carried out by incuba- tion, at 65 °C for 15 min, of a 5 lL mixture containing 80 mm Hepes-KOH pH 8.1, 150 mm KCl, 0.2 mm Na-EDTA, 0.5 pmol RNA and 2.5 pmol of complementary oligodeoxyribonucleotide. RNA degradation and analysis Unless otherwise stated, RNA degradation and analysis were performed using a 10 lL reaction mixture containing 50 mm Hepes-KOH pH 8.1, 75 mm KCl, 1 mm dithiothrei- tol, 1 mm Na-phosphate, 0.06 mm Na-EDTA, 0.1 mm MgCl 2 , 0.5 pmol RNA, 2 pmol Tth PNPase, and, where indicated, 2.5 pmol of a protective oligodeoxyribonucleo- tide. Reactions were carried out at 65 °C for 15 min and terminated by the addition of 10 lLof10mm Na-EDTA. Nucleic acids were isolated by phenol extraction [53] and analysed by nondenaturing PAGE (in the presence of TBE buffer: 100 mm Tris ⁄ HCl, 100 mm boric acid, 2 mm Na-EDTA) or by denaturing PAGE (in the presence of buffer TBE containing 7 m urea). Gels were stained after electrophoresis using silver [59]. Preparation of oligodeoxyribonucleotides with uniform 5¢ ends To prepare oligodeoxyribonucleotide R-38 with uniform 5¢ end, two other oligonucleotides were synthesized: a longer oligodeoxyribonucleotide R-50 (3¢-AAAGCCTGGAGGTC ACGCACAATGGCGTGACAATC GGGCCCCCTTAAG GT-5¢) and a shorter complementary oligodeoxyribonucleo- tide F-24 (5¢-CACTGTTAGCCCGGGGGAATTCCA-3¢). Upon annealing, these produced a partial duplex (last 24 nucleotides of R-50 with complete sequence of F-24), which was digested with restriction endonuclease SmaI at the site underlined in R-50. After digestion, the preparation was melted, and the 38 nt long oligodeoxyribonucleotide R-38 was isolated by PAGE. M. V. Falaleeva et al. RNA phosphorolysis by Tth PNPase FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation ª 2008 FEBS 2223 [...]... messenger RNA J Mol Biol 88, 233–245 31 Wood JN & Hutchinson DW (1976) Thermostable polynucleotide phosphorylases from Bacillus stearothermophilus and Thermus aquaticus Nucleic Acids Res 3, 219–229 32 Hishinuma F, Hirai K & Sakaguchi K (1977) Thermophilic polynucleotide phosphorylase from Thermus thermophilus Purification and properties of an altered form of enzyme which lacks phosphorolytic activity to polynucleotide. .. The Authors Journal compilation ª 2008 FEBS 2225 RNA phosphorolysis by Tth PNPase M V Falaleeva et al 44 Spickler C & Mackie GA (2000) Action of RNase II and polynucleotide phosphorylase against RNAs containing stem-loops of defined structure J Bacteriol 182, 2422–2427 45 Lorentzen E & Conti E (2005) Structural basis of 3¢ end RNA recognition and exoribonucleolytic cleavage by an exosome RNase PH core... dealing with RNA Nat Struct Mol Biol 12, 562–564 Coburn GA & Mackie GA (1999) Degradation of mRNA in Escherichia coli: an old problem with some new twists Prog Nucleic Acid Res Mol Biol 62, 55–108 Cheng ZF & Deutscher MP (2003) Quality control of ribosomal RNA mediated by polynucleotide phosphorylase and RNase R Proc Natl Acad Sci USA 100, 6388– 6393 Li Z, Reimers S, Pandit S & Deutscher MP (2002) RNA quality... Grunberg-Manago M (1999) Polynucleotide phosphorylase In Encyclopedia of Molecular Biology, Vol 3 (Creighton TE, ed.), pp 1911–1918 John Wiley and Sons, New York Thang MN, Guschlbauer W, Zachau HG & GrunbergManago M (1967) Degradation of transfer ribonucleic acid by polynucleotide phosphorylase I Mechanism of phosphorolysis and structure of tRNA J Mol Biol 26, 403–421 RNA phosphorolysis by Tth PNPase 30 Soreq... MF, Williams MG, Luisi BF, Jones GH & Carpousis AJ (2002) Running rings around RNA: a superfamily of phosphate-dependent RNases Trends Biochem Sci 27, 11–18 5 Yehudai-Resheff S, Portnoy V, Yogev S, Adir N & Schuster G (2003) Domain analysis of the chloroplast polynucleotide phosphorylase reveals discrete functions in RNA degradation, polyadenylation, and sequence homology with exosome proteins Plant... Biology’ of the Russian Academy of Sciences, and by the program ‘Living Systems’ of the Ministry of Industry, Science and Technology of the Russian Federation References 1 Grunberg-Manago M & Ochoa S (1955) Enzymic synthesis and breakdown of polynucleotides: polynucleotide phosphorylase J Am Chem Soc 77, 3165– 3166 2224 2 Littauer UZ & Soreq H (1982) Polynucleotide phosphorylase In The Enzymes, Vol XV (Boyer... of the pnp gene of Escherichia coli encoding polynucleotide phosphorylase Homology of the primary structure of the protein with the RNAbinding domain of ribosomal protein S1 J Biol Chem 262, 63–68 34 Soreq H & Littauer UZ (1977) Purification and characterization of polynucleotide phosphorylase from Escherichia coli Probe for the analysis of 3¢ sequences of RNA J Biol Chem 252, 6885–6888 35 Martell AE... Z, Reimers S, Pandit S & Deutscher MP (2002) RNA quality control: degradation of defective transfer RNA EMBO J 21, 1132–1138 Houseley J, LaCava J & Tollervey D (2006) RNA- quality control by the exosome Nat Rev Mol Cell Biol 7, 529–539 Clements MO, Eriksson S, Thompson A, Lucchini S, Hinton JC, Normark S & Rhen M (2002) Polynucleotide phosphorylase is a global regulator of virulence and persistency in... cleavage by an exosome RNase PH core Mol Cell 20, 473–481 46 Singer MF (1958) Phosphorolysis of oligoribonucleotides by polynucleotide phosphorylase J Biol Chem 232, 211–228 47 Chou JY & Singer MF (1970) A kinetic analysis of the phosphorolysis of oligonucleotides by polynucleotide phosphorylase J Biol Chem 245, 995–1004 48 Littauer UZ & Kornberg A (1957) Reversible synthesis of polyribonucleotides... basis for polynucleotide phosphorylase catalytic activity, processivity, and regulation Structure 8, 1215–1226 11 Kelly KO & Deutscher MP (1992) Characterization of Escherichia coli RNase PH J Biol Chem 267, 17153– 17158 12 Siomi H, Matunis MJ, Michael WM & Dreyfuss G (1993) The pre-mRNA binding K protein contains a novel evolutionarily conserved motif Nucleic Acids Res 21, 1193–1198 13 Bycroft M, . Factors influencing RNA degradation by Thermus thermophilus polynucleotide phosphorylase Marina V. Falaleeva, Helena. PNPase, polynucleotide phosphorylase; Tth PNPase, PNPase of Thermus thermophilus. 2214 FEBS Journal 275 (2008) 2214–2226 ª 2008 The Authors Journal compilation