Substrate specificity of the pseudouridine synthase RluD in Escherichia coli Margus Leppik, Lauri Peil, Kalle Kipper, Aivar Liiv and Jaanus Remme Institute of Molecular and Cell Biology, Tartu University, Tartu, Estonia Pseudouridines (Y) are the most common modifications in stable RNAs. Pseudouridine was discovered as a fifth nucleotide in yeast tRNA 50 years ago [1]. Pseudo- uridines are synthesized from uridine by pseudouridine synthases, a reaction that does not need additional cofactors or external energy sources. Pseudouridine synthases are classified into five families according to their amino acid sequence [2,3]. Despite low sequence homology of the enzymes, structural comparison of crystal structures reveals that all pseudouridine synth- ases share a core with a common fold and a conserved active site cleft [4]. Pseudouridines are found in all tRNAs and high- molecular rRNAs. 16S ribosomal RNA from Escheri- chia coli contains one pseudouridine Y516 formed by RsuA [5]. 23S rRNA from E. coli contains ten Y resi- dues, which are made by six enzymes RluA–RluF [6]. Enzymes such as RsuA and RluB isomerize only one uridine in the substrate RNA whereas others (RluC and RluD) make three pseudouridines [7–9]. RluA modifies uridine 746 in 23S rRNA and uridine 32 in some specific tRNA species [10]. RluD isomerizes uridines at positions 1911, 1915, and 1917 in stem-loop 69 (H69) of 23S rRNA [8,9]. Y1917 is found at the corresponding position of the large ribosomal subunit RNAs throughout all king- doms. It is the most conserved modification in rRNA [6]. Y1915 is also highly conserved [6]. Y1915 is methylated at N3 in several eubacteria [11]. Y1911 is also well conserved, except in archaea [6]. Y to C mutation at position 1917 has a dramatic effect on the ribosome functioning, which is explained by the universal nature of Y1917 [12]. H69 of 23S rRNA directly interacts with tRNA at the A and P site [13,14]. H69 forms the intersubunit bridge 2 with helix 44 of 16S rRNA [15,16]. Y1917 forms a reverse Hoogsteen base pair with A1912, which in turn forms A-minor interaction with base pairs C1407– G1494 of 16S rRNA [16]. Pseudouridine residues can stabilize the 3D RNA structure as revealed by Keywords 23S rRNA; helix 69; pseudouridine; ribosome assembly; RluD Correspondence J. Remme, Riia 23, 51010 Tartu, Estonia Fax: +372 42 0286 Tel: +372 73 75031 E-mail: jremme@ebc.ee (Received 24 May 2007, revised 6 Septem- ber 2007, accepted 10 September 2007) doi:10.1111/j.1742-4658.2007.06101.x Pseudouridine synthase RluD converts uridines at positions 1911, 1915, and 1917 of 23S rRNA to pseudouridines. These nucleotides are located in the functionally important helix-loop 69 of 23S rRNA. RluD is the only pseudouridine synthase that is required for normal growth in Escherichia coli. We have analyzed substrate specificity of RluD in vivo. Mutational analyses have revealed: (a) RluD isomerizes uridine in vivo only at posi- tions 1911, 1915, and 1917, regardless of the presence of uridine at other positions in the loop of helix 69 of 23S rRNA variants; (b) substitution of one U by C has no effect on the conversion of others (i.e. formation of pseudouridines at positions 1911, 1915, and 1917 are independent of each other); (c) A1916 is the only position in the loop of helix 69, where muta- tions affect the RluD specific pseudouridine formation. Pseudouridines were determined in the ribosomal particles from a ribosomal large subunit defective strain (RNA helicase DeaD – ). An absence of pseudouridines in the assembly precursor particles suggests that RluD directed isomerization of uridines occurs as a late step during the assembly of the large ribosomal subunit. Abbreviations Y, pseudouridine; ASL, anticodon stem loop; H69, stem-loop 69. FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS 5759 thermodynamic studies on the isolated helix-loop 69 [17]. Deletion of the yfiI(rluD) gene reduces the growth rate by three- to five-fold [8,9] and leads to accumula- tion of the precursor particles for the large and small subunits [18]. Ribosomes lacking RluD specific pseudouridines are less stable at low magnesium ion concentration and exhibit reduced activity during poly(U) translation in vitro [19]. These effects were attributed to the lack of pseudouridines in H69 [19]. However, in the presence of an as yet unidentified sec- ond site mutation, bacteria lacking RluD are able to grow at a similar rate as wild-type cells. This pseudo- revertant strain does not contain Y in H69 and exhib- its normal ribosome assembly [19, L. Peil & J. Remme, unpublished data]. The exact role of pseudouridines in H69 for the function of ribosomes remains unclear. Grosjean and coworkers have highlighted two important questions concerning tRNA-modifying enzymes: (a) at which stage of the complex maturation process does the modification of a given nucleoside occur and (b) how does the corresponding enzyme rec- ognize the target site within the tRNA architecture [20]? Similar questions on the rRNA-modifying enzymes remain unanswered for nearly all rRNA-mod- ifying enzymes, including RluD. RluD belongs to the RluA family [2]. Binding of RluA to one of its substrates, tRNA Phe anticodon stem-loop, induces reorganization of the RNA [21]. An ability of the RNA substrate to adopt the alternative fold with a reverse Hoogsteen base pair is used by RluA to recognize its substrate [21]. It is possible that such indirect sequence readout is used also by other members of the RluA family (e.g. RluD). Under cell- free conditions and in vitro, transcribed rRNA RluD has low substrate specificity, converting one out of 20 uridine residues in 23S rRNA and one out of eight uri- dine residues in 16S rRNA to pseudouridine [9]. We have analyzed the substrate specificity of the pseudouridine synthase RluD in vivo by using 23S rRNA variants. Single point mutations were intro- duced to the plasmid copy of the 23S rRNA gene. 23S rRNA variants were expressed in vivo and purified by affinity tag. Pseudouridines around helix-loop 69 were determined by chemical modification. Results RluD synthesizes Y only at U1911, U1915, and U1917 Helix-loop 69 of E. coli 23S rRNA contains uridine at positions 1911, 1915, and 1917, which are all converted to pseudouridines by the pseudouridine synthase RluD (Fig. 1). To test whether or not RluD is able to modify uridine at other positions of the H69, nucleotides A1912, C1914, A1916, and A1919 were substituted by uridine as single point mutations. Mutant genes were expressed in vivo and the mutant ribosomal particles were isolated as described in the Experimental proce- dures. All 23S rRNA variants were incorporated into fractions 50S and 70S. Pseudouridines were determined by chemical modification, followed by reverse trans- criptase directed primer extension. The primer exten- sion stop on the CMCT treated RNA (+ lane) indicated the presence of pseudouridine at the particu- lar position when the stop was not present in the con- trol (– lane). m 3 Y present at position 1915 [11] causes primer extension stop independent of CMCT treatment (Fig. 2). It must be noted that m 3 Y can form a Wat- son–Crick base pair in a syn conformation of glycoside bond, allowing low level readthrough by reverse trans- criptase. Therefore, it was not possible to identify pseudouridylation at position 1915. rRNA isolated from an E. coli strain lacking a gene for RluD was used as a control. Primer extension patterns of the wild-type 23S rRNA from the 50S and the 70S particles were similar: stop sites were found at positions 1911 and 1917 on the CMCT treated RNA and not on the control RNA (wild-type 70S and wild-type 50S; Fig. 2). This shows that pseudouridines were present at positions 1911 and 1917 in the wild-type 23S rRNA in the 50S subunits and in the 70S ribosomes. 23S rRNA from a strain lacking RluD did not have pseudouridines in H69 as Fig. 1. A scheme of E. coli 23S rRNA stem-loop 69 (H69). Numbers of the pseudouridine residues and m 3 Y are indicated according to standard E. coli 23S rRNA numeration. RluD specificity in vivo M. Leppik et al. 5760 FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS expected (RluD – ; Fig. 2). 23S rRNA variants A1912U (70S), C1914U (70S and 50S), and A1919U (70S, 50S) exhibited the same CMCT ⁄ alkali induced primer extension stop site pattern as wild-type 23S rRNA (Fig. 2), indicating that pseudouridines were present at the wild-type positions (1911 and 1917) and not at mutant positions. The 23S rRNA variant A1916U derived from free 50S subunits did not exhibit pseudo- uridine specific stop sites at positions U1911, U1916, and U1917 (Fig. 2). Thus, pseudouridine was not detected in H69 of the 23S rRNA variant A1916U iso- lated from free 50S subunits. In the 70S particles, the CMCT induced stop sites were just on the border of detection limit, indicating that only traces of pseudo- uridines were present (Fig. 2). This result suggests that A1916 is an important specificity determinant for RluD. Pseudouridines were found at wild-type posi- tions (1911 and 1917) in spite of the presence of uri- dine at other positions. We conclude that RluD is specifically recruited to positions U1911 and U1917 of E. coli 23S rRNA, at least in vivo. Mutations at position A1916 affect the specificity of RluD Substitution of uridine by cytidine at position 1911 leads to the disappearance of a CMCT dependent stop signal at position 1911 as expected (Fig. 3). Similar results were obtained with the transition at posi- tions 1915 and 1917. These mutations had an effect on the formation of pseudouridine exclusively at the mutant position and did not affect either of the other two positions (Fig. 3). Y1917 was also found in the 23S rRNA variant containing the double mutation U1911C ⁄ U1915C (Fig. 3). Thus, isomerization of uri- dines 1911 and 1917 in H69 occurs autonomously of each other and is independent of modification at 1915. Additional 23S rRNA variants were analyzed regarding RluD specificity. Mutations A1912C, A1913G, C1914A, A1916C, A1918G and A1919G showed either no or very little effect on the RluD spe- cific pseudouridine formation at positions 1911 and 1917 (Fig. 4). It must be noted that the band corre- sponding to the Y1911 of the variants A1913G and A1918G is very weak due to strong stop sites at 1917 and 1915 (Fig. 4). Longer exposure revealed the pres- ence of a Y specific band at position 1911 (not shown). The transversion A1916U reduced pseudouridine for- mation to undetectable level in 50S subunits and caused strong reduction in 70S ribosomes by RluD as described above. (Fig. 2). Y residues at positions 1911 and 1917 were found in 70S ribosomes when A1916 was substituted by G, albeit at a reduced level. By contrast, Y residues at positions 1911 and 1917 did not show up in 23S rRNA extracted from free 50S Fig. 2. Primer extension analysis of the pseudouridines in the helix-loop 69 of 23S rRNA. 23S rRNA variants were expressed in vivo, 70S ribosomes and free 50S subunits containing mutant RNA were isolated. +, CMCT ⁄ alkali treatment; –, untreated RNA. Bands corresponding to the 23S rRNA positions 1911, 1915, and 1917 are indicated. DNA sequencing lanes are shown on the right. Note that CMCT induced stop site is one nucleotide below the actual modification. Fig. 3. Pseudouridine sequencing analysis of 23S rRNA variants. Effect of substitution of uridine on the RluD activity. For details, see Fig. 1 and Experimental Procedures. M. Leppik et al. RluD specificity in vivo FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS 5761 subunits (Fig. 4). The presence of pseudouridines in 70S but not in free 50S subunits can be explained by reduced rates of pseudouridine formation due to the transition of A1916 to G. The effects of the mutations in 23S rRNA on the pseudouridine formation in H69 are summarized in Table 1. It is evident that only mutations at position A1916 to G and U affect RluD activity. A weak stop site at position 1915 was detected in CMCT untreated samples of 23S rRNA variants A1916U and A1916G (only 50S fraction), suggesting a low level of N 3 methylation of uridine. RluD modification occurs during late assembly We have analyzed whether RluD forms pseudouridines during early assembly on naked 23S rRNA or alterna- tively requires the presence of r-proteins for its activ- ity. We used an E. coli strain (deaD – ), negative for the RNA helicase DeaD (CsdA), which has been shown to be deficient in ribosomal large subunit assembly [22]. 40S particles accumulating in this strain are assembly precursors of 50S subunits (L. Peil & J. Remme, unpublished results). We have analyzed 23S rRNA from 40S, 50S and 70S particles regarding Y residues in H69 of 23S rRNA. Primer extension analysis showed that, in the wild-type strain, both 70S and 50S particles contain RluD specific pseudouridines. 70S ribosomes from the deaD – strain contain Y residues at positions 1911 and 1917, indicating that RluD is active in the absence of DeaD. 23S rRNA in the 40S and 50S particles shows only traces of RluD specific Y residues (Fig. 5). 50S particles of the deaD – strain have low functional activity, probably due to incom- plete assembly (L. Peil & J. Remme, unpublished Fig. 4. Effect of single point mutations in the helix-loop 69 on the RluD directed pseudouridine synthesis. For details, see Fig. 1 and Experi- mental Procedures. Table 1. Conversion of uridine at positions 1911 and 1917 to pseudouridine on 23S rRNA variants in the 50S subunits and 70S ribosomes in vivo. Presence of a pseudouridine residue on the 23S rRNA variant is shown by +; absence of pseudouridine residue is shown by a ). ND, not determined. Mutant 70S 50S 1911 1917 1911 1917 U1911C – + – + A1912U + + ND ND A1912C + + ND ND A1913G + + + + C1914U + + + + C1914A + + + + U1915C + + + + A1916G + + – – A1916C + + + + A1916U – – – – U1917C + – + – A1918G + + + + A1919G + + ND ND A1919U + + ND ND Fig. 5. Identification of pseudouridines in the helix-loop 69 in differ- ent stages of 50S biogenesis by primer extension. Ribosomal parti- cles were isolated from wild-type and the deaD – strain. 40S particles are assembly precursors accumulating in the deaD – strain. RluD specificity in vivo M. Leppik et al. 5762 FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS results). We conclude that RluD directed isomerisarion of uridines in the helix-loop 69 of 23S rRNA occurs as a late event during assembly of the ribosomal large subunit, but before the 50S subunit enters the 70S pool. Discussion Each pseudouridine in eubacterial rRNA is formed by a single pseudouridine synthase [6]. On the other hand, some pseudouridine synthases are able to isomerize several uridines (e.g. RluC and RluD isomerize three uridines each). It was proposed that RluD recognizes all uridines in or near the loop of helix 69 and con- verts them to pseudouridines [6]. This type of regional specificity was recently found to be used by TruA which converts any uridine at positions 38–40 of sub- strate tRNA to pseudouridine [23]. We have tested whether or not RluD has similar regional specificity by mutating nucleotides at positions A1912, A1913, C1914, A1916, and A1919 of H69 to uridine. None of the mutant uridines was converted to pseudouridine in vivo. This result demonstrates that the pseudouridine synthase RluD is specific to positions 1911 and 1917 and its specificity is not of regional type. Thus, in vivo, RluD is highly specific to positions where pseudouri- dine is found in a wide range of species. By contrast, RluD was reported to exhibit low substrate specificity in vitro [9]. RluD converted 40 uridines to pseudouri- dines in in vitro transcribed 16S and 30 uridines in the 23S rRNA transcript [9]. In vitro transcribed rRNA is not correctly folded, which can be one reason for the low specificity. Substitution of U by C revealed formation of pseudouridines at positions 1911 and 1917 auto- nomous of each other and independent of Y1915 for- mation. Moreover, Y1917 was formed on the double mutant 23S rRNA (U1911C ⁄ A1915C). Mutations in H69 exhibited similar effects on the Y synthesis at positions 1911 and ⁄ or 1917, suggesting that both have the same identity determinants. The most important position for determining RluD specificity was found to be nucleotide A1916. A and C are per- missive nucleotides. Both nucleotides are found in bacterial 23S rRNA sequences. U has the strongest and G has intermediate negative effect on the RluD activity. Suzuki and colleagues have selected 20 viable sequence variants of H69 [24]. The presence of pseudouridines was determined for a subset of vari- ants. Mutations U1915A and A1916C did not affect pseudouridinylation of U1911 and U1917 [24], in agreement with the results described in the present study. Mutation A1916 to G and U had severe effects on RluD in vivo. Y1911 and Y1917 were found on the 23S rRNA variants in the 70S ribosomes, albeit at a reduced level, but not in the 50S fraction. Substitution of A1916 by U had stronger effect on the RluD com- pared to A1916G mutation. Thus, the nucleotide A1916 in 23S rRNA is an important identity element for pseudouridine synthesis at both positions 1911 and 1917. This indicates that the identity determinants are at least partially overlapping for both positions. Although RluD is highly specific to positions 1911 and 1917, this enzyme is insensitive to the base substitu- tions in the loop of helix 69. This apparent contradic- tion can be resolved assuming that A1916 is important for the initial docking of the RluD. Identity determi- nants required for the pseudouridine formation may lie within the flipped out conformation of H69 because base flipping is obligatory for Y synthesis [4]. The cocrystal structure of E. coli TruB bound to the T-stem and loop fragment of tRNA has been deter- mined [25]. This structure suggests that TruB recog- nizes the T-loop by shape and makes sequence specific contacts with a few invariant nucleotides, such as C56 [25]. Genetic and biochemical data have shown that isomerization of U55 in tRNA by E. coli TruB or by its yeast (Saccharomyces cerevisiae) ortholog PUS4 is sensitive to base substitution in the TY-loop [20,26]. A second cocrystal structure of pseudouridine synthase and its substrate was recently determined for RluA and the anticodon stem loop (ASL) of tRNA [21]. This enzyme gains specificity by inducing a conformational change in the substrate RNA. This structure involves a reverse Hoogsteen base pair (U33:A36) and base flip- ping of three bases including the substrate uridine 32. These structural elements are absent in normal tRNA. RluA appears to recognize its substrate by indirect readout of a protein induced RNA structure [21]. It is therefore interesting to note that hairpin 35 of the 23S rRNA (another substrate for RluA) has a structure in its isolated state similar to ASL of free tRNA [27]. With three flipped out bases, hairpin 35 in the ribo- some has completely different structure [16]. It has some similarity to the conformation of ASL in com- plex with RluA. It is possible that RluA supports refolding hairpin 35 during ribosome assembly. RluD belongs to the RluA family and the catalytic cores of both proteins show similar folds [4]. H69 of 23S rRNA is known to adopt different conformations in 50S and 70S ribosomes. In 70S ribosomes, Y1917 forms a reverse Hoogsteen type base pair with A1912 which makes it similar to the ASL in complex with RluA. Deletion of RluD leads to accumulation of assembly defective ribosomal subunits [18,19]. It is thus tempting M. Leppik et al. RluD specificity in vivo FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS 5763 to speculate that RluD may have a role in ribosome assembly by an RNA chaperone function in the H69 region. Ribosomal subunit assembly involves folding of rRNA and r-proteins and association of both into functional subunits. In addition, post-transcriptional modifications are made in rRNA and post-transla- tional modifications added to r-proteins during ribo- some assembly. Ribosomal 50S subunits are formed in vivo during 1–2 min after transcription of 23S rRNA, but an additional 3 min are required before the large subunits enter the functional 70S pool [28]. This additional time is probably used for making late assembly specific modifications of rRNA and r-pro- teins and for fine adjustment of ribosome structure. Analysis of the pseudouridylation pattern in ribosome assembly precursor particles of 23S rRNA has shown that the pseudouridines in H69 are formed by RluD during late assembly. RluD can function in ribosome assembly by helping to refold H69 into a functional structure. The refolded H69, containing pseudouri- dines, supports ribosome subunit association. The results indicating that Y1911 and Y1917 are present in the 70S ribosomes but not in the 50S subunits on the 23S rRNA variant A1916G are in agreement with the role of RluD in functional 50S formation. Taken together, the results suggest that rRNA modification, in particular by RluD, is an important event during late assembly of 50S particles. Experimental procedures Plasmids and strains 23S rRNA single point mutations were initially constructed by PCR mutagenesis on plasmid pXB containing the E. coli 23S rRNA gene with streptavidine binding tag [29]. Mutant 23S rRNA variants A1912U, A1912C, A1913G, C1914U, A1916G, A1916C, A1916U, A1917G, A1919C, and A19191U were fused to ptBB expression vector in which the rrnB operon is under control of the tac promoter [12]. 23S rRNA variants U1911C, C1914A, U1915C, and A1918G were recloned into the expression vector pKK 3535. The deaD – strain (deaD414), where the deaD gene was disrupted by a kanamycin resistance cassette, was gener- ously provided by Dr Kenneth E. Rudd (University of Miami, FL, USA). For our experiments, bacteriophage P1 transduction was used to transfer deaD::kan gene into E. coli strain MG1655 [30]. The E. coli strain MG1655 was used as parental strain to construct strain rluD::cat (rluD114) according to the method of Datsenko and Wanner [31]. The cat cassette from plasmid pKD3 was amplified by PCR using Pwo polymerase (Roche Diagnostics GmbH, Mannheim, Ger- many) and primers rluD114::cat(pKD3) 5¢ (5¢-GCT ACA ATA GCA CAC TAT ATT AAA CGG CAA AGC CGT AAA ACC CC G TGT AGG CTG GAG CTG CTT CG- 3¢) and rluD114::cat(pKD3) 3¢ (5¢-GAC CAG ATT AAT GTG AAA AGA AAA TCA CGC GTA CCG GAT CGT CTT G AT GGG AAT TAG CCA TGG TCC-3¢) (comple- mentary regions to the rluD-flanking regions are under- lined). The resulting 1123 bp PCR product was gel-purified using UltraClean 15 DNA Purification kit (MoBio, West Carlsbad, CA, USA). Twenty nanograms of of purified PCR product was then electroporated into MG1655 ⁄ pKD46 competent cells, previously grown in the presence of 10 mm arabinose and made competent by con- centrating ten-fold and washing five times with ice-cold 10% glycerol. Selection for the recombination event and the elimination of pKD46 plasmid was performed as described previously [31]. Colonies were tested for the rluD deletion by PCR, using primers flanking the gene, and by Southern analysis. The rluD::cat strain had a full deletion of both rluD and yfiH, together with their annotated pro- moter sequences. The inserted cat cassette had a distance of 90 nt from the b2595 gene and 177 nt from the clpB gene. Preparation of ribosomes and rRNA 23S rRNA variants A1912U, A1912C, A1913G, C1914U, A1916G, A1916C, A1916U, A1917G, A1919C, and A19191U were expressed in the E. coli strain XL1-Blue transformed with the corresponding ptBBtag plasmid. Cells were grown at 37 °Cin2· YT medium supplemented with ampicillin (100 lgÆmL )1 ). Ribosomes were isolated from cells 2 h after induction with isopropyl thio-b-d-galactoside (Fermentas, Vilnius, Lithuania) (1 mm) until an attenuance of 0.2 at D 600 was reached. Other strains were grown in 2 · YT medium until an attenuance of 0.2–0.3 at D 600 was reached (deaD414 cells were grown at 25 °C). Bacteria were collected by low-speed centrifugation, resuspended in lysis buffer and lysed by freeze-thawing as described previously [12].The lysate was diluted with an equal volume of ice-cold buffer LLP (10 mm Tris ⁄ HCl (pH 8.0), 60 mm KCl, 60 mm NH 4 Cl, 12 mm MgOAc, 6 mm b-mercaptoethanol). Two millilitres of diluted lysate were layered onto a 10–25% (w ⁄ w) sucrose gradient in buffer LLP and centrifuged for x 2 t ¼ 2.7 · 10 11 at 4 °C. Gradients were analysed with con- tinuous monitoring at 254 nm. Fractions containing ribo- somal particles were collected and stored at )70 °C. 70S ribosomes and free 50S subunits were isolated by sucrose gradient centrifugation as described previously [12]. Plasmid encoded mutant ribosomal large subunits were separated from wild-type ribosomes according to a previously described method [29]. 50S or 70S particles were incubated with streptavidine-Sepharose (GE Helthcare Biosciences RluD specificity in vivo M. Leppik et al. 5764 FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS AB, Uppsala, Sweden) in buffer [10 mm Tris ⁄ HCl (pH 8.0), 60 mm KCl, 60 mm NH 4 Cl, 1 mm MgOAc, 6 mm b-mer- captoethanol]. Plasmid encoded ribosomes were eluted in the same buffer containing 100 m m biotine. Mutant ribo- somes contained less than 10% wild-type ribosomes accord- ing to primer extension around the tag site. 23S rRNA variants U1911C, C1914A, U1915C, and A1918G were expressed in E. coli strain MC315 (Dlac, DrecA, D7 prrn) [32,33] transformed with pKK3535 deriva- tives. 70S, 50S, and 30S gradient fractions were collected and precipitated with 2.5 volumes of ice-cold ethanol. rRNA was prepared using a modified protocol [12]. In brief, ribosomes were dissolved in 200 lL of water and 1 mL of PN solution (catalog no. 19071; Qiagen, Hilden, Germany) was added. Ribosomal proteins were extracted by vigorous shaking for 20 min at room temperature. Twenty microlitres of a 50% silica suspension in water was added and RNA was bound for additional 10 min at room temperature with gentle mixing. Silica was pelleted by cen- trifugation at 3000 g for 30 s and washed twice with 70% ethanol. RNA was eluted with 50 lL of water (10 min at room temperature). Determination of pseudouridines Pseudouridines were determined according to the method of Ofengand [34]. Fifteen micrograms rRNA were dissolved in 20 lL of water, 80 lL of BEU buffer (7 m urea, 4 mm EDTA, 50 mm Bicine ⁄ NaOH pH 8.5) and 20 lLof CMCT ⁄ BEU buffer (1 m CMCT in BEU buffer) (CMCT; Sigma-Aldrich Chemie GmbH, Steinheim, Germany) were added. One hundred microlitres of BEU buffer were added to 15 lg of rRNA in 20 lL of water serving as negative control. Both samples were incubated at 37 °C for 20 min for CMCT modification of U, G and Y residues. After incubation, 38 lL4m NaOAc were added, followed by 600 lL of cold 96% ethanol. Samples were kept at )20 °C for 10 min and the RNA precipitate was collected by cen- trifugation at 6000 g and 4 °C. The supernatant was care- fully removed and RNA was washed twice with 600 lLof 70% ethanol. The precipitate was dried at 37 °C for 10 min. rRNA was dissolved in 50 lL of NPK buffer (20 mm NaHCO 3 ,30mm Na 2 CO 3 ,2mm EDTA) and the samples were incubated at 37 °C for 4 h to allow removal of CMCT from U and G residues. After incubation, rRNA was precipitated and washed as described above. The pre- cipitate was dissolved in 20 lL of water and stored at )20 °C. Pseudouridine sequencing of rRNA was carried out by primer extension using primer U1 (CAG CCT GGC CAT CAT TAC GCC) and AMV reverse transcriptase (Seikagaku Corp., Tokyo, Japan) in the presence of [a- 32 P]dCTP (Amersham Biosciences, Piscataway, NJ, USA). The resulting DNA fragments were resolved in 7% polyacrylamide-urea gel. Radioactivity was visualized by Typhoon phosphor imager (GE Healthcare, Tokyo, Japan). Acknowledgements This paper is dedicated to the memory of Professor James Ofengand from the Univeristy of Miami. We thank Dr U ¨ . Maiva ¨ li for critically reading the manu- script and Joachim Gerhold (both from Tartu Univer- sity) for correcting the English. The research was supported by Estonian Science Foundation Grants No. 5822 (JR) and No. Aivar (AL). References 1 Davis FF & Allen FW (1957) Ribonucleic acids from yeast which contain a fifth nucleotide. J Biol Chem 227, 907–915. 2 Koonin EV (1996) Pseudouridine synthases: four fami- lies of enzymes containing a putative uridine-binding motif also conserved in dUTPases and dCTP deaminas- es. Nucleic Acids Res 24, 2411–2415. 3 Kaya Y & Ofengand J (2003) A novel unanticipated type of pseudouridine synthase with homologs in bacte- ria, archaea, and eukarya. RNA 9, 711–721. 4 Hamma T & Ferre-D’Amare AR (2006) Pseudouridine synthases. Chem Biol 13, 1125–1135. 5 Wrzesinski J, Bakin A, Nurse K, Lane BG & Ofengand J (1995) Purification, cloning, and properties of the 16S RNA pseudouridine 516 synthase from Escherichia coli. Biochemistry 34, 8904–8913. 6 Ofengand J (2002) Ribosomal RNA pseudouridines and pseudouridine synthases. FEBS Lett 514, 17–25. 7 Conrad J, Sun D, Englund N & Ofengand J (1998) The rluC gene of Escherichia coli codes for a pseudouridine synthase that is solely responsible for synthesis of pseudouridine at positions 955, 2504, and 2580 in 23 S ribosomal RNA. J Biol Chem 273, 18562–18566. 8 Raychaudhuri S, Conrad J, Hall BG & Ofengand J (1998) A pseudouridine synthase required for the forma- tion of two universally conserved pseudouridines in ribosomal RNA is essential for normal growth of Escherichia coli. RNA 4, 1407–1417. 9 Huang L, Ku J, Pookanjanatavip M, Gu X, Wang D, Greene PJ & Santi DV (1998) Identification of two Escherichia coli pseudouridine synthases that show multisite specificity for 23S RNA. Biochemistry 37, 15951–15957. 10 Wrzesinski J, Nurse K, Bakin A, Lane BG & Ofengand J (1995) A dual-specificity pseudouridine synthase: an Escherichia coli synthase purified and cloned on the basis of its specificity for psi 746 in 23S RNA is also specific for psi 32 in tRNA (phe). RNA 1, 437–448. 11 Kowalak JA, Bruenger E, Hashizume T, Peltier JM, Ofengand J & McCloskey JA (1996) Structural charac- terization of U*-1915 in domain IV from Escherichia coli 23S ribosomal RNA as 3-methylpseudouridine. Nucleic Acids Res 24, 688–693. M. Leppik et al. RluD specificity in vivo FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS 5765 12 Liiv A, Karitkina D, Maivali U & Remme J (2005) Analysis of the function of E. coli 23S rRNA helix-loop 69 by mutagenesis. BMC Mol Biol 6, 18. 13 Korostelev A, Trakhanov S, Laurberg M & Noller HF (2006) Crystal structure of a 70S ribosome-tRNA com- plex reveals functional interactions and rearrangements. Cell 126, 1065–1077. 14 Selmer M, Dunham CM, Murphy F, Vt Weixlbaumer A, Petry S, Kelley AC, Weir JR & Ramakrishnan V (2006) Structure of the 70S ribosome complexed with mRNA and tRNA. Science 313, 1935–1942. 15 Yusupov MM, Yusupova GZ, Baucom A, Lieberman K, Earnest TN, Cate JH & Noller HF (2001) Crystal structure of the ribosome at 5.5 A resolution. Science 292, 883–896. 16 Schuwirth BS, Borovinskaya MA, Hau CW, Zhang W, Vila-Sanjurjo A, Holton JM & Cate JH (2005) Struc- tures of the bacterial ribosome at 3.5 A resolution. Science 310, 827–834. 17 Sumita M, Desaulniers JP, Chang YC, Chui HM, Clos L & Chow CS (2005) Effects of nucleotide substitution and modification on the stability and structure of helix 69 from 28S rRNA. RNA 11, 1420–1429. 18 Ofengand J, Malhotra A, Remme J, Gutgsell NS, Del Campo M, Jean-Charles S, Peil L & Kaya Y (2001) Pseudouridines and pseudouridine synthases of the ribosome. Cold Spring Harb Symp Quant Biol 66, 147–159. 19 Gutgsell NS, Deutscher MP & Ofengand J (2005) The pseudouridine synthase RluD is required for normal ribosome assembly and function in Escherichia coli. RNA 11, 1141–1152. 20 Becker HF, Motorin Y, Sissler M, Florentz C & Gros- jean H (1997) Major identity determinants for enzy- matic formation of ribothymidine and pseudouridine in the T psi-loop of yeast tRNAs. J Mol Biol 274, 505– 518. 21 Hoang C, Chen J, Vizthum CA, Kandel JM, Hamilton CS, Mueller EG & Ferre-D’Amare AR (2006) Crystal structure of pseudouridine synthase RluA: indirect sequence readout through protein-induced RNA struc- ture. Mol Cell 24, 535–545. 22 Charollais J, Dreyfus M & Iost I (2004) CsdA, a cold- shock RNA helicase from Escherichia coli, is involved in the biogenesis of 50S ribosomal subunit. Nucleic Acids Res 32, 2751–2759. 23 Hur S & Stroud RM (2007) How U38, 39, and 40 of many tRNAs become the targets for pseudouridylation by TruA. Mol Cell 26, 189–203. 24 Hirabayashi N, Sato NS & Suzuki T (2006) Conserved loop sequence of helix 69 in Escherichia coli 23 S rRNA is involved in A-site tRNA binding and translational fidelity. J Biol Chem 281, 17203–17211. 25 Hoang C & Ferre-D’Amare AR (2001) Cocrystal struc- ture of a tRNA Psi55 pseudouridine synthase: nucleo- tide flipping by an RNA-modifying enzyme. Cell 107, 929–939. 26 Gu XYuM, Ivanetich KM & Santi DV (1998) Molecu- lar recognition of tRNA by tRNA pseudouridine 55 synthase. Biochemistry 37, 339–343. 27 Lebars I, Yoshizawa S, Stenholm AR, Guittet E, Dou- thwaite S & Fourmy D (2003) Structure of 23S rRNA hairpin 35 and its interaction with the tylosin-resistance methyltransferase RlmAII. EMBO J 22, 183–192. 28 Lindahl L (1975) Intermediates and time kinetics of the in vivo assembly of Escherichia coli ribosomes. J Mol Biol 92, 15–37. 29 Leonov AA, Sergiev PV, Bogdanov AA, Brimacombe R & Dontsova OA (2003) Affinity purification of ribo- somes with a lethal G2655C mutation in 23 S rRNA that affects the translocation. J Biol Chem 278, 25664– 25670. 30 Blattner FR, Plunkett G III, Bloch CA, Perna NT, Bur- land V, Riley M, Collado-Vides J, Glasner JD, Rode CK, Mayhew GF et al. (1997) The complete genome sequence of Escherichia coli K-12. Science 277, 1453– 1474. 31 Datsenko KA & Wanner BL (2000) One-step inactiva- tion of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 97, 6640–6645. 32 Asai T, Zaporojets D, Squires C & Squires CL (1999) An Escherichia coli strain with all chromosomal rRNA operons inactivated: complete exchange of rRNA genes between bacteria. Proc Natl Acad Sci USA 96, 1971– 1976. 33 Liiv A & O’Connor M (2006) Mutations in the inter- subunit bridge regions of 23 S rRNA. J Biol Chem 281, 29850–29862. 34 Ofengand J, Del Campo M & Kaya Y (2001) Mapping pseudouridines in RNA molecules. Methods 25, 365– 373. RluD specificity in vivo M. Leppik et al. 5766 FEBS Journal 274 (2007) 5759–5766 ª 2007 The Authors Journal compilation ª 2007 FEBS . analyzed the substrate specificity of the pseudouridine synthase RluD in vivo by using 23S rRNA variants. Single point mutations were intro- duced to the plasmid. of 23S rRNA. RluD is the only pseudouridine synthase that is required for normal growth in Escherichia coli. We have analyzed substrate specificity of RluD